Page 1

THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

KOMMentarer p책 forskning i framkant.

1


2


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

3

innehåll 5 

8  14 20 26   32

40 46 54  60 66

74 80 86 92 98 104

112 118

Det bästa av två världar - förord av Sara Rosengren The roles of advertising A matter of equity, med förord av Annika Rehn ”I’m smart enough to not want it”, med förord av Annika Rehn Too cheap to work for? med förord av Henrika Thomasson Users’ engagement with brands and companies on Facebook med förord av Jessica Morales Re-thinking brand loyalty, med förord av Peter Granström The faces of advertising Improving the in-store customer information process using mobile augmented reality, med förord av Jan Lindforss How do advergames persuade? med förord av Peter Bergendahl Contaminated hedonic experiences, med förord av Anna Hjalmarsson Advertising repetition, med förord av Anna Hjalmarsson The changing roles of celebrities in advertising med förord av Karl Wikström Perceptions of advertising Is there gender bias in the assessment of advertising creativity? med förord av Peter Bergendahl A client perspective on advertising quality, med förord av Jan Lindforss “I” lose, “Others” gain, med förord av Henrika Thomasson An experiential account of happiness in life and in ads med förord av Jessica Morales Trust me I’m an advert! med förord av Peter Granström The influence of corporate social responsibility on consumer product responses med förord av Kim Saxberg Reception of advertising ”This program contains advertising”, med förord av Kim Saxberg It matters who ”they” think you are, med förord av Karl Wikström


4


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

5

Det bästa av två världar SARA ROSENGREn, Stockholm school of economics SWEDEN

Jag älskar att se teorier omsatta i praktiska kommunikationslösningar. Min förhoppning är att den här boken ska inspirera till just sådana. Att den ska ge dig och andra nya insikter och idéer att pröva i nya sammanhang. Oddsen är goda. Boken för nämligen samman marknadsföringsforskning i världsklass med de skarpa hjärnorna och praktiska erfarenheterna hos några av landets bästa planners. De texter som ingår boken kommer från ICORIA 2012, en vetenskaplig konferens om marknadskommunikation, som arrangeras av Centrum för konsumentmarknadsföring (CCM) vid Handelshögskolan i juni 2012. Precis som den svenska kommunikationsbranschen ligger svensk forskning på området i framkant. CCM rankas som etta i Europa sett till vetenskapliga publikationer inom marknadskommunikation. Vår forskning används som kurslitteratur på utbildningar vid framstående universitet i Europa, USA och Asien. Vi är också ansvariga för spjutspetsutbildningarna MCXL och Handelshögskolans kandidatprogram i Retail Management. Jag är glad och stolt över denna skrift som tillkommit genom ett samarbete mellan CCM och Komms utbildningskommitté. CCMs ambition har alltid varit ett nära samarbete mellan akademi och praktik. Denna bok utgör en ny möjlighet för oss att förverkliga vår ambition. Jag vill nu ta tillfället i akt och tacka. Först och främst Jessica Bjurström och Petronella Panérus som förverkligat denna idé. Ett stort tack även till de planners som bidragit med sin tid och erfarenhet genom att skriva förorden till forskarnas texter. Och naturligtvis till forskarna själva. Slutligen vill jag också tacka Torsten och Ragnar Söderbergs stiftelser vars generösa stöd möjliggjort CCM:s värdskap för ICORIA.


6


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

The roles of Advertising

7


8

Förord: Consumers in NPD Allt fler företag ser värdet av att involvera sina kunder i utvecklingen av nya produkter och erbjudanden. Det hjälper till att skapa positiva samtal om varumärket men är också ett sätt att på riktigt se till att man svarar upp mot människors behov och önskemål och tar till vara deras engagemang för varumärket och deras kreativitet. Den här studien visar på vilket sätt företag kan involvera kunderna i produktutvecklingen för att få ut mest av det. Studien har jämfört hur kunders deltagande på två olika sätt i produktutvecklingen, påverkar varumärkesvärdet. Man har dels tittat på vilken typ av produkt det är frågan om – en som är i överensstämmelse med varumärket, eller en som är en helt ny typ av produkt för varumärket. Och dels när i produktutvecklingen man involverar sina kunder – i början för att komma upp med idéer till nya produkter, eller senare för att välja bland redan utvecklade förslag till nya produkter. Människors involvering i produktutvecklingen visar sig tillföra störst värde på ett väldigt väl definierat sätt. När det gäller utveckling av helt nya typer av produkter, så tillför det mest värde till varumärket att involvera människor i idé-genereringen. Men när det gäller utveckling av produkter som är i linje med varumärket, så tillför det mest värde att involvera människor först när det är dags att välja bland redan utvecklade förslag. Studien avslöjar också de bakomliggande orsakerna till de här resultaten. Det innebär att alla som arbetar med produktutveckling, men även utveckling av koncept för erbjudanden eller kommunikation, här får hjälp att aktivera sina kunder i den del av processen där de tillför varumärket mest värde.

ANNIKA Rehn


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

9

A MATTER OF EQUITY: EFFECTS OF OTHER CONSUMERS’ ENGAGEMENT IN NEW PRODUCT DEVELOPMENT Karina Töndevold Stockholm School of Economics Sweden Micael Dahlén Stockholm School of Economics Sweden Kristina Heinonen Hanken School of Economics Finland

INTRODUCTION This paper explores how consumers respond to information about other consumers’ engagement in a brand’s new product development (NPD). More specifically, we investigate how the two common forms of engagement, ideation and selection of new products, impact on brand equity. Within this framework, we hypothesize that ideation impacts more favorably for incongruent new products and selection more favorably for congruent new products. It is becoming increasingly popular to involve consumers in corporate NPD, especially since collaboration and communication is greatly facilitated by various online solutions. Some of the most known examples of consumers engaging in corporate NPD include: Dell using IdeaStorm to access consumers’ new product ideas (Bayus, 2011); Threadless inviting consumers to both design and vote for t-shirts to be sold through its online shop (Ogawa and Piller, 2006; Fuchs and Schreier, 2010); LEGO collaborating with lead users (von Hippel 1986) and using open source for the development of the Mindstorm product (Koerner, 2006); and Muji, a Japanese manufacturer of consumer goods who integrates new product concepts in one of its product lines after the concept has been deemed attractive by measure of enough binding pre-orders from customers (Ogawa and Piller, 2006). However, consumers engaged in corporate NPD are only a minority of all the consumers who will come in contact with the company’s products and communication, since the new products they are engaged in developing (and the engaged consumers themselves) are often used in marketing communications aimed at a wider group of non-engaged consumers. While a body of literature is building on consumer engagement in NPD, only one study to date has investigated how the non-engaged majority of consumers react. In that study, Fuchs and Schreier (2011) found that, overall, consumers responded favorably to other consumers’ engagement in the ideation and selection of a brand’s new products. This paper extends their findings by comparing the two forms of engagement, and testing the hypotheses that they work differently for congruent versus incongruent products depending on their varying impacts on brand equity. CONSUMER ENGAGEMENT IN NEW PRODUCT DEVELOPMENT The last few years have witnessed a growing body of theory on consumer engagement in NPD. To date, literatures have mainly formed in three areas. The first centers on consumer motivation where, for example, von Hippel studies participants’ enjoyment of the activity and their feeling of control over their own work in the Open Source movement (2005), as well as lead users developing product solutions because their needs cannot yet be


10

accommodated for by the market (1986). The second area of research regards how consumers engaged in the corporate NPD process perceive the product and brand, For example, Fuchs, Prandelli and Schreier (2010) found that consumers who take part in NPD develop a stronger product demand and feeling of ownership of the selected products. The third research area focuses on the actual products that the NPD results in, for example, Poetz and Schreier (2011) compare the quality of output of crowdsourced versus professionally, internally new products. Our study contributes to a fourth area, namely the effects of consumer engagement in terms of marketing communications and other consumers’ perceptions of the brand and the new product. In the only study to date in this area, Fuchs and Schreier (2011) investigated the two common forms of consumer NPD engagement (which they call “empowerment”): creating ideas for new product designs, and selecting among which product designs to be produced. Lumping the two together, they found that consumer engagement overall can generate higher levels of perceived customer orientation, and consequently more favorable corporate attitudes and stronger behavioral intentions. However, we expect that consumer engagement will also impact on perceptions of the brand’s equity, that is, the extent to which 1) the brand is perceived to be unique and differ from others (e.g., Keller, 1993), and 2) the value of its offer can be attributed solely to the brand (e.g., van Osselaer and Alba, 2003). More specifically, we expect consumer engagement in new product ideation to impact more on the extent to which the brand is perceived as unique and different from others, and consumer engagement in new product selection to impact more on the extent to which the value of the offer is attributed to the brand. Consequently, we hypothesize that consumer engagement in new product ideation is evaluated more favorably for incongruent versus congruent products, and that consumer engagement in new product selection is evaluated more favorably for congruent versus incongruent products. CONSUMER ENGAGEMENT AND BRAND EQUITY In it simplest form, brand equity is often defined as the impact of a brand’s name on product evaluations (e.g., Erdem and Swait, 1998). For the sake of our investigation, we suggest that the impact of the brand would be contingent on two factors. First, the associations that the brand evokes must be unique and differ from the competition; otherwise the brand will not offer any value to consumers beyond those that are generic to the product category (e.g., Keller, 1993). Second, the value must be attributed specifically to the brand as opposed to other potential factors. For example, van Osselaer and Alba (2003) showed that making competing explanations salient to consumers, such as a specific ingredient, reduced their use of the brand as a cue for their product evaluations. While a brand’s equity impacts consumer evaluations of its products, the products, and consumers’ perceptions of them, in turn, impact on brand equity in a dynamic process (e.g.,Keller, 1993). In other words, the extent to which consumers perceive a brand’s products as unique and different from the competition, feed back to the brand’s equity to either increase (if perceived as unique and different) or reduce (if not perceived as unique and different) it. Similarly, if consumers’ experience of the product is attributed specifically to the brand, it will feed back to the brand’s equity (either favorably or unfavorably), otherwise it will not. We expect that the two common forms of consumer engagement in NPD relate differently to brand equity. More specifically, we expect that new product ideation would relate to consumer perceptions of the extent to which the brand is unique and different, whereas new product selection would relate the extent to which the value of the product is attributed to the brand.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

11

Starting with new product ideation, we hypothesize that consumer engagement impacts more favorably on brand equity for incongruent than congruent products. This would be in line with brand schema theory, which suggests that consumers hold expectations for the brand’s congruent products, that is, the products that the brand is trusted to produce, but do not hold expectations for products that are incongruent with the brand (e.g., Johnson and Russo, 1984). In other words, consumer-ideated incongruent products could thus add something novel and unique to the brand and enhance brand equity, whereas consumer-generated congruent products risk reducing consumers’ perceptions of the brand as unique and different and erode brand equity (“apparently, it’s not only the brand that can come up with this type of products”): H1: Consumer engagement in new product ideation impacts more favorably on brand equity for incongruent versus congruent products. Similarly, we expect consumer engagement in new product ideation to generate more favorable evaluations for incongruent than congruent products, by way of its impact on brand equity: H2: Consumer engagement in new product ideation impacts more favorably on brand and product evaluations for incongruent versus congruent products. Turning to new product selection, we hypothesize that consumer engagement impacts more favorably on brand equity for congruent than incongruent products. We expect that consumers would be more prone to attribute the value of the new product to the brand if they knew that the brand selected (and thereby explicitly endorsed) the product, and that they would be less prone to attribute the value of the new product to the brand when informed that it was selected by other consumers (“this is not necessarily something that the brand stands for”). Therefore, we expect that consumer engagement in the new product selection impacts more favorably for congruent products, which consumers expect from and already associate with the brand, rather than for incongruent products, which consumers do not expect from or associate with the brand (and, according to brand schema theory, may therefore subtype and thereby not attribute to the brand, e.g., Sujan and Bettman, 1989): H3: Consumer engagement in new product selection impacts more favorably on brand equity for congruent versus incongruent products. Similarly, we expect consumer engagement in new product ideation to generate more favorable evaluations for congruent than incongruent products, by way of its impact on brand equity: H4: Consumer engagement in new product selection impacts more favorably on brand and product evaluations for congruent versus incongruent products. METHOD The study employed a 2 (consumer engagement in ideation vs. selection) X 2 (congruent vs. incongruent product), full factorial, between-subjects design. Stimuli and Procedure The study was conducted as a between-subjects experiment with manipulated introductory texts. The respondents were first acquainted with a brand of hamburger restaurants (called “X” to avoid confounding effects from previous experience), and then informed that the brand was about to launch a new product/service. For our manipulation of consumer


12

engagement, the new product/service was presented to be either developed by a consumer or selected by consumers. For our manipulation of product congruency, we presented a product/service that was congruent or incongruent with the brand (and hamburger restaurants offers in general). Product congruency had been pre-tested (n=51) on a three-item, ten-point scale (“unexpected/expected”, “atypical/typical”, “low/high similarity”), Mcongruent = 8.1 vs. Mincongruent, 4.2, p<.01. To increase the level of generalizability, half of the stimuli were products (e.g. a new type of hamburger) and half of the stimuli were services (e.g. being offered to pick up cinema tickets at the restaurant cashier). A nation-representative cross-section of participants were recruited via an online panel (n= 386, mean age = 39.71, range 16-64, 47% female) and were randomly assigned to one of the 8 cells in the experiment. Measures Brand equity was gauged with two three-item measures on a ten-point Likert scale (1 = do not agree, 10 = agree). Brand uniqueness included the items: “X is unique”, “X is different from competitors” and “X is special”. These questions were used to create a mean index (α = .975). Brand attribution included the items:“The product says a lot about X”, “The quality of the product is related only to X”, “The quality of the product is a result of X:s efforts” (α = .90). Both brand evaluations and product evaluations were measured by three items on the same Likert scale as described above. The items were “good”, “like” and “positive” (brand evaluations: α = .986, product evaluation: α = .976). ANALYSIS AND RESULTS First, we ran a MANOVA (multivariate analysis of variance) on all the dependent variables simultaneously. Consumer engagement had a significant main effect (F(4,386) = 2.90, p < .01), as did our effect of interest, the consumer engagement x product congruency interaction: (F(4,386) = 3.00, p < .01). Next, we conducted planned contrasts (Table 1). Supporting H1 and H2, the consumer-ideated incongruent new products rated more favourably than the congruent new products both in terms of brand equity and in terms of product and brand evaluations. In line with the argumentation leading up to H1, the main impact on brand equity was in terms of the brand’s uniqueness. Testing the argumentation leading up to H2, which suggested that the impact of consumer engagement in the new product ideation on product and brand evaluations would come by way of its impact on brand equity, we employed Sobel’s mediation test. It showed that brand uniqueness mediated the effect on both product evaluations (z = 2.23) and brand evaluations (z = 2.25). Supporting H3 and H4, the consumer-selected congruent new products rated more favourably than the incongruent new products both in terms of brand equity and in terms of product and brand evaluations. In line with the argumentation leading up to H3, the main impact on brand equity was in terms of brand attribution. Testing the argumentation leading up to H4, which suggested that the impact of consumer engagement in the new product selection on product and brand evaluations would come by way of its impact on brand equity, we employed Sobel’s mediation test. It showed that brand attribution mediated the effect on both product evaluations (z = 2.24) and brand evaluations (z = 2.26).


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

13

DISCUSSION The results of our experiment suggest that revealing information on consumer engagement in NPD will have differing effects on brand equity, product evaluations and brand evaluations depending on the congruency of the product as well as when in the process the consumers were invited to engage with the NPD. In this study, consumer-ideated incongruent new products rated more favourably than the congruent new products, both in terms of brand equity and product- and brand evaluations. Similarly, consumer-selected congruent new products rated more favourably than the incongruent new products. Brand equity mediated the effect of consumer engagement on both product and brand evaluations. More specifically, brand uniqueness mediated the effects of consumer engagement in new product ideation, and brand attribution mediated the effects in new product selection. This study extends the research on consumer engagement in NPD by adding the explanatory power of brand equity and brand schema to the marketing effects of the engagement. The results open up for more detailed knowledge of the consequences of, and how to best use, consumer engagement in NPD. It also extends the brand equity literature by updating it to a reality where no longer only the companyâ&#x20AC;&#x2122;s products and actions impact on brand equity, but consumersâ&#x20AC;&#x2122; engagement can have significant impact, too. Especially consumer brands may find the results of this study useful as it can guide them to either consider what type of brand extension (congruent/ incongruent) they intend to release, or how they prefer to engage with consumers in NPD (ideation/selection). Either could form the basis for how to best proceed in order to strengthen both product and brand evaluations while engaging consumers in NPD. As this was a first step towards a new research orientation where the effects of consumer engagement is investigated in more detail, there is much left to find out. For example, it would be interesting to find out more about the effects of consumer engagement in NPD in terms of other marketing objectives such as purchase intention and word of mouth. Future research could also explore whether the results above differ for different types of products, and for brands with different levels of equity and types of brand schemas.


14

Förord: Det är inte vad vi säger utan hur vi säger det Detta är en ofta upprepad sanning bland oss som arbetar med marknadsföring och kommunikation. Och den blir bara mer relevant i en tid då produkter och erbjudanden blir allt mer lika varandra. Den här studien visar på ett övertygande sätt att hur ett varumärke uttrycker sig, påverkar människors uppfattning, inte bara om reklamen, utan också om varumärket självt. Studien har jämfört människors uppfattning om annonser där det huvudbudskap man har velat föra fram har varit det samma, men uttryckts på olika sätt. I den ena typen av annonser mer funktionellt och underskattande av människors intelligens, och i den andra typen mer kreativt och med tilltro till människors förmågor. De mer funktionella annonserna var inte lika uppskattade som de mer kreativa. Dessutom gav de mer funktionella annonserna en sämre uppfattning om både erbjudandet och varumärket. De förmedlade i större utsträckning känslan av att varumärket inte brydde sig om sina kunder, och att varumärket missade att uppfylla sin del av överenskommelsen om att ge något tillbaka till dem som tar sig tid att uppmärksamma varumärkets reklam. Extra intressant var att studien också visade att huvudbudskapet i båda typerna av annonser uppfattades lika tydligt. De flesta annonser har till uppgift att inte bara förmedla ett budskap utan också bidra till att uppfattningen om varumärket stärks. Denna studie visar att mer kreativt formulerade budskap klarar båda uppgifterna, medan mer funktionella bara klarar den första. Studien kastar därmed hjälpfullt ljus över en fråga som ofta debatteras i samband med utvecklingen av annonser och annan reklam och kommunikation.

ANNIKA Rehn


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

15

”I’M SMART ENOUGH TO NOT WANT IT”: UNDERESTIMATING CONSUMERS’ INTELLIGENCE IN ADVERTISING SIGNALS LESS CONSUMER CARE Micael Dahlén stockholm School of Economics Sweden Amna Kirmani University of Maryland US Sara Rosengren Stockholm School of Economics Sweden Edith Smit University of Amsterdam Netherlands Helge Thorbjornsen Norweigan School of Economics Norway

INTRODUCTION This paper asks the question, can advertisements that underestimate consumers’ intelligence impact negatively on perceived consumer care? Merging the two recent findings that consumers may indeed perceive advertisements to underestimate their intelligence (Dahlén, Rosengren, and Smit, forthcoming), and that advertising may function as a signal of consumer care (Rosengren and Dahlén, 2011), we hypothesize that advertisements that underestimate consumers’ intelligence signal that a brand cares less about its consumers. Perceiving the brand to care less, we expect consumers to consequently evaluate the advertising less favorably, and to make negative inferences about the brand and its offering. The notion that advertising can underestimate consumers’ intelligence is about as old as modern advertising itself. Already in 1968, Bauer and Greyser (1968) published findings from a survey of American consumers, showing a widespread concern that advertising in general underestimates its receivers. A decade and a half later, Bartos (1981) compared repeated surveys over the years and found that consumers’ concern had steadily increased, leading the author to conclude that advertisers are putting their brands at great risk by potentially insulting their consumers. Our investigation extends the recent finding by Dahlén, Rosengren, and Smit (forthcoming) that such advertising does indeed insult consumers by testing the hypothesis that it signals that the brand cares less for its consumers. This hypothesis builds on the recent advance of the advertising signal literature, which has shown that advertising can function as a signal of a brand’s care for its consumers. Extending Rosengren and Dahlén’s (2011) finding that (creative) advertising can increase a brand’s perceived consumer care and thereby impact favorably on ad and brand evaluations, we test whether advertising could also decrease perceived consumer care and thereby lower evaluations of the ad, the brand, and its offering. In a first experiment, we test our hypotheses by exposing consumers to otherwise identical advertisements for either a detergent or a TV brand, manipulating the (same) message to more or less underestimate their intelligence. A second experiment extends the findings to an additional category (vacuum cleaners) and rules out the influence of consumers’ persuasion knowledge, that is, their sense that they are being subjected to manipulation by the advertiser, as a competing explanation.


16

ADVERTISING THAT UNDERESTIMATES CONSUMERS’ INTELLIGENCE SIGNALS LESS CARE An extensive body of literature on advertising signals shows that the way in which a message is communicated impacts on consumers’ perceptions of the advertiser. Factors ranging from advertisement size and repetition, to production quality and creativity have been manipulated, and found to impact on consumer perceptions of a brand’s ability, confidence and smartness (e.g., Ambler and Hollier, 2004; Dahlén, Rosengren and Törn, 2008; Kirmani, 1990, 1997). Most recently, Rosengren and Dahlén (2011) manipulated the creativity of advertisements and found that more creative advertising signaled that the advertiser cares more about its consumers (by suggesting that it understands that today’s fatigued and advertisingsavvy receivers require something extra). Flipsiding this argument, we expect that advertising that underestimates consumers’ intelligence signals that the brand cares less about its consumers, by suggesting that the brand has not made the effort to learn enough about its consumers’ knowledge, experience and preference to craft the message appropriately: H1: Underestimating consumers’ intelligence in an advertisement impacts negatively on perceived consumer care. We also expect that a brand’s lower perceived care for its consumers will impact negatively on evaluations of its advertising. Research suggests that consumers perceive advertising as an exchange, where they offer their time and attention, which they value highly, and expect to be appropriately rewarded; if not, they react negatively to this so-called inequitable exchange (e.g., Dahlén, Granlund and Grenros, 2009; Hunt and Kernan, 1991; Phillips and Noble, 2007). Accordingly, we hypothesize that the equitability of the advertising exchange drops together with the brand’s perceived care, impacting negatively on consumer evaluations of the advertisement: H2: By lowering perceived consumer care, underestimating consumers’ intelligence in an advertisement impacts negatively on evaluations of the advertisement. Building further on the advertising signal literature, we expect that consumers will use the signal about the sender to make inferences about the brand and its offering (Kirmani 1990; 1997). In our case, that would mean that consumers will infer from the signal that the brand cares less about its consumers that its offerings will be of less value “if they don’t care much about me, they’re not likely to understand what, or find it worthwhile making something, I really want ”. This would flipside the finding by Rosengren and Dahlén (2011) that increased perceived care impacts favorably on evaluations of the brand and its offering: H3: By lowering perceived consumer care, underestimating consumers’ intelligence in an advertisement impacts negatively on evaluations of the brand and its offering STUDY 1 & 2 The hypotheses were tested in two experimental studies using an online panel (study 1: n=161; study 2: n=62). The studies manipulated the advertising message (underestimating consumers’ intelligence vs. not underestimating consumers’ intelligence) as a between- subjects factor. Participants were exposed to otherwise completely identical advertisements for either a TV or a detergent brand (study 1) or a vacuum cleaner brand (study 2). The allocation of ads was random.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

17

Stimuli and Procedure Print advertisements introduced the participants to brand “X”. The brands were offered as an anonymized brand to the respondents to avoid potentially confounding effects of familiarity or personal experience with existing brands. Similar to Dahlén, Rosengren and Törn’s (2008) study on the signal effects of creativity, we manipulated the advertisements to be identical between our two conditions, with regard to visual elements, layout, number of words and the intended message. The only thing that differed was the choice of words. In the TV advertisements, the words were: “We have new technology which makes it lighter to put up and clearer to watch” (underestimating consumers’ intelligence), versus “Crystals. Easy on the back and easy on the eye. That is crystal clear” (not underestimating consumers’ intelligence). In the detergent advertisements, the words read: “Put it on the clothes. Rinse. Clean. And Easy” (underestimating consumers’ intelligence), versus “No waste in no time. Just aim and shoot” (not underestimating consumers’ intelligence). The vacuum cleaner ads read: “It has never been this easy. Just press the button, point, sweep. Once is all it takes” (underestimating consumers’ intelligence), versus “It takes the work out of housework. It makes no repeat motions, no engineering skills necessary” (not underestimating consumers’ intelligence). In pretests, 60 consumers, who were not included in the experiment afterwards, rated the advertisements on a four-item ten-point semantic interval: underestimates/overestimates my intelligence, requires very little/very much of me, targets someone less smart/smarter than me, seems oversimplified/too sophisiticated to me. The items were averaged into an index (α = .83), which rated significantly lower and well below the midpoint of the scale for the advertisements that were manipulated to underestimate consumers’ intelligence (M = 2.52) compared to the advertisements that were manipulated to not underestimate consumers’ intelligence, which were near the scale’s midpoint (M = 5.29), t =16.34, p < .01. Measures Perceived consumer care (H1) was measured with four items on a ten-point Likert scale. The four items were “X cares about its consumers”, “X devotes great effort to its consumers”, “X focuses on its consumers” and “X is caring”, averaged to form an index (α = .90). Ad attitude (H2) was measured with three items on a ten-point semantic differential, namely good/bad, positive/negative, appealing/unappealing (α = .91). For the test of H3 we used three three-item measures of consumers’ perceptions of the brand and its offerings, namely brand attitude (good/bad, positive/negative, satisfactory/unsatisfactory, α = .91), perceived quality (high quality, better than average, better than the competition, α = .91), and intentions, (buy, try, tell friend about, recommend, α = .80). In study 2, persuasion knowledge was gauged on a three-item ten-point semantic differential: manipulative/ not manipulative, honest/dishonest, pushy/not pushy (α = .88). In both studies we used the confound check taken from Dahlén, Rosengren and Törn (2008), in which participants were asked to tick the correct message out of six alternatives (the alternatives were the same across all cells and were designed to be plausible for both brands). There were no significant differences in message identification between the advertisements that were manipulated to underestimate consumers’ intelligence (over 96% correct in both studies) and the advertisements that were designed not to underestimate consumers’ intelligence (over 94% correct in both studies), suggesting that participants would perceive the way the message was communicated differently, but not the explicit message itself.


18

RESULTS A MANOVA (multivariate analysis of variance) on all dependent variables simultaneously in study 1 found that our manipulation (underestimating vs. not underestimating consumers’ intelligence) had significant main effect on all dependent variables, F(5,161) = 3.27, Wilks’ α = .87, p < .01. Planned contrasts supported the hypotheses (Table 1). Sobel mediation tests showed that perceived consumer care mediated the effects on ad attitude (z = 4.62) as well as evaluations of the brand and its offering (z > 4). The results of study 1 thus support our hypotheses. Including persuasion knowledge as a covariate in study 2, a MANCOVA on all the dependent variables simultaneously revealed that even when accounting for persuasion knowledge as a competing explanation our manipulated factor (underestimating vs. not underestimating consumers’ intelligence) still produced significant effects on all dependent variables, F(5,62) = 3.83 Wilks’ α = .94, p < .01. Planned contrasts are reported in Table 2. Sobel mediation tests showed that perceived consumer care mediated the effects on ad attitude (z = 3.24) and evaluations of the brand and its offering (z > 3). The results of study 2 thus support our hypotheses. DISCUSSION Taken together, the two studies suggest that underestimating consumers’ intelligence when crafting an advertising message can send a signal that the brand cares less about its consumers. This signal impacts negatively on both ad attitudes and evaluations of the brand and its offering. Our studies thus show that consumers may not only react favorably or unfavorably to an advertisement based on what it actually says or where it is said, but also to how it is said. Our investigation adds to the advertising signal literature by showing that consumers are sensitive to how much a brand actually cares about them, by showing that consumers may use the brand’s advertising as a proxy for its care, and by showing that consumers may react to negatively to inequitable exchange and infer from advertising that signals that a brand that cares less that its offering is of lower quality. The investigation also finds that consumers may indeed perceive that advertising underestimates their intelligence, and therefore cautions advertisers to go beyond the path of least resistance when crafting a message, and that consumer experience and comprehension need to be carefully taken into consideration in the planning.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

19


20

förord: En studie om hur reklam påverkar varumärken som potentiella arbetsgivare Bondesson et al har i denna studie undersökt om det finns ett positivt samband mellan reklamens kreativitet och varumärkets attraktionskraft som arbetsgivare. Eftersom ”employer branding” har växt fram som ett särskilt fokusområde inom marknadsföring, och allt oftare dyker upp som uppdrag på byrån, är resultatet av studien intressant för såväl uppdragsgivare som utformare av reklam. Kanske framför allt för de företag för vilka rekrytering är en kritisk faktor. Dagens allt högre krav på effektiv reklam – en stramare reklambudget som ska räcka till mer – ökar kraven på reklam som kan leverera på ”fler” områden samtidigt. Bondesson et al kommer fram till att det finns ett positivt samband. Ju mer kreativ reklam, desto mer attraktivt framstår varumärket som arbetsgivare. Studiens resultat har dock två förbehåll; Det första är att respondenterna är Handelsstudenter. Beroende på arbetsgivare, kanske det inte alls är Handelsstudenter – med deras antagligen relativt homogena preferenser– man vill attrahera. Ett forskarbibliotek, en advokatbyrå eller en bank vars värdegrund inte alls vill locka utåtriktade personer, kanske inte drar till sig rätt typ av rekryteringsunderlag med kreativ reklam. Det andra är att man testade studenterna på ett befintligt, känt, varumärke, där förförståelsen inte var kartlagd. Oavsett förbehållen, slår studiens resultat tydligt fast att de som arbetar med att utforma konsumentreklam ska tänka in vilka övriga grupper företaget kommunicerar med och vilken effekt reklamen har på dem. Det kan vara potentiella anställda, såväl som potentiella samarbetspartners, leverantörer etc.

henrika Thomasson


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

21

TOO CHEAP TO WORK FOR? EXPLORING THE IMPACT OF ADVERTISING ON THE BRAND AS POTENTIAL EMPOLYER Sara Rosengren Stockholm School of Economics Sweden Niklas Bondesson Lund University Sweden

PURPOSE OF THE STUDY This paper looks at the effects of advertising on potential employees’ perception of a brand as a potential employer. More specifically, we investigate how advertising influences perceptions of brand ability (cf. Dahlén, Rosengren, and Törn, 2008) and the impact this might have on attitudes and intentions towards the brand as a future employer. Advertising is a highly visible part of a company. The main target of advertising is typically (potential or existing) customers. However, due its visibility other stakeholders will be exposed to and react to such ads too. Another stakeholder group that is likely to be affected is future and current employees. Still, little research has investigated how employees react to advertising (Gilly and Wolfinbarger 1998) and the effects it might have on potential employees (Ewing et al 2002). The current study examines the effects of advertising (directed at consumers) on attractiveness of a brand as a future employer. Although ads typically focus on communicating what the company has to offer, the way it executes these messages also implicitly conveys information about what company in terms of, for instance, its core values and trustworthiness (Gilly and Wolfinbarger, 1998). For instance, recent research shows that creatively communicating a message will send a signal about brand ability and product quality (Dahlén, Rosengren, and Törn 2008, Rosengren and Dahlén 2012). In this study we test if this signal will also influence the attractiveness of a brand to future employees. In an experimental study we explore the effects of communicating the same message (low prices) in a more or less creative way on perceptions of the sender as a future employee. THEORETICAL FRAMEWORK Employer branding research applies marketing and branding theories to a matter that traditionally has been addressed mainly within the human resources (HR) field. It sets out to investige the attractiveness of the organisation as an employer, as perceived by current and potential employees. The term employer brand was allegedly coined by Ambler and Barrow (1996), who defined it as “the package of functional, economic and psychological benefits provided by employment, and identified with the employing company” (p. 187). During the last decade, employment branding has grown extensively as a field (Edwards, 2010). One reason is that it has become increasingly difficult for companies to attract and keep skilled employees (Schlager et al., 2011), which has made it important to have a unique “employee value proposition” (Moroko & Uncles, 2008).


22

Typically, employer branding research has applied theories and models developed for an external audience (customers) to be used for an internal audience (employees) as well. For instance, Wilden et al (2010) develops an employee-based brand equity model based on Erdem and Swait’s (1998) consumer-based brand equity model and Berthon et al (2005) study “how astute employers can embrace the principles and practices associated with external brand management and marketing communication, internally” (p. 152). Consequently in studying the effects of advertising on employees this research has focused on advertisments focusing on the company as an employer, such as employer-, or recruitment- advertising (Moroko & Uncles, 2008; Berthon et al., 2005). Such advertisements are distinct from traditional advertising, which focus on what the brand has to offer consumers. In the current study we investigate the effects of advertising directed at customers on perceptions of the brand as a potential employer. Thus we contribute a novel perspective on employee branding as well as an employer perspective to traditional advertising research. Hypotheses Recent research on creative advertising executions suggests that advertising creativity functions as a marketing signal. More specifically, Dahlén, Rosengren and Törn (2008) find that creative advertising executions are seen as a sign of fitness or ability on behalf of the sender. By being creative in their execution of a certain message, advertisers highlight their ability to think differently and this is seen as a sign of innovativeness and smartness. The fact that the brand is creative in its advertising is taken as a sign of a sender who is more able not only when in terms of its advertising, but also in terms of the quality of the products it produces (Dahlén et al., 2008; Kirmani and Rao, 2000). We believe that similar effects could be found with regards to the perceptions of the brand as a potential employer. Not only should creative advertising enhance perceptions of brand ability (Dahlén et al 2008; Rosengren and Dahlén 2012), but it should also enhance attitudes and intentions towards the brand as a future employer. The logic for our proposed hypothesis is that brand ability is taken as a sign of development possibilities offered by the company, and this, in turn, has a positive effect on attitudes and intentions. Thus we hypothesize: H1. Advertising creativity impacts positively on perceived brand ability H2. Advertising creativity impacts positively on perceived development possibilities offered for employees H3. Advertising creativity impacts positively on attitudes towards the brand as a future employer H4. Advertising creativity impacts positively on intentions towards the brand as a future employer METHOD The hypotheses were tested in an experimental study in which participants were first exposed to a (more or less creative) ad for a corporate retail brand followed by some questions about the ad and brand (cf. Dahlén, Rosengren and Törn, 2008) They were then given a brief introduction to the retailer as an employer (same text for all) and asked questions about the retailer as a potential employer. In the last part of the questionnaire they answered questions with regards to their intended reactions to a hypothetical job offer (same for all) from the brand. Ninety-five students enrolled in a marketing class participated in the study. Allocation to the different conditions was random. Although student


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

23

samples have several weaknesses, it is actually suitable for this research as they are soon to be employed by organizations (Berthon et al 2005). Stimuli Development To provide a strong test our hypotheses we decided to use a real brand and real ads. A large food retailer was selected as it offered interesting career prospects to the participants. The retailer we choose positions itself on based on its low prices. To manipulate creativity we selected two different ads with the same low price-message (pretested message comprehension both ads =100%, n=68), but with different level of creativity (pretested on scale 1=not creative/7=creative, 4.2 vs. 1.6, p<.01, n=68). Measures Perceived ability was measured with two items taken from Rosengren and Dahlén (2012): “X is smart” and “X is innovative” (1 = do not agree / 7 = agree completely, r = .66). Perceived development prospects was measured by asking to what extent participants agreed with the statement “[Brand] offers good development prospects for its employees” (1 = do not agree / 7 = agree completely). To asses attitude towards the brand as an employer we used a three-item, seven-point semantic differential: good/bad, positive/negative, favorable/ unfavorable (α = .91, cf. Rosengren and Dahlén, 2012). To measure intentions towards the brand as an employer, participants were presented with a possible job offer from the brand. The job was said to “be just what you are looking for in terms of responsibilities and development prospects”. They then rated the likelihood to a) formally apply for, b) going to an interview for and c) actually accept this job (scale: 1 = not at all likely / 7 = very likely). RESULTS Manipulation checks showed that perceived ad creativity was indeed higher in our more (mean = 4.53) vs. less (mean = 1.68) creative ad condition. There was no difference in message comprehension: 100% (more creative) vs. 97% (less creative) correctly identified the message as focused on low prices. This suggests that there was a difference in ad creativity but not in the message as such. Testing the hypotheses, we first ran a MANOVA (multivariate analysis of variance) on all dependent variables simultaneously. Our manipulation (more vs. less creative ad) had significant main effect on all dependent variables, F(7,86) = 6.92, Wilks’ α = .64, p < .01, eta2=.36. Next, we used planned contrasts to compare the mean values between conditions and test the hypotheses individually (Table 1). As indicated above all hypotheses are supported at p<.05. To further test our reasoning that the effects are caused by brand ability signaling more development possibilities we performed a mediation analysis. This showed that the effect of brand ability on employment attitude is, indeed, mediated by perceived development possibilities (z = 3.26, p<.01). DISCUSSION This paper set out to explore the effects of advertising on potential employees’ perception of a corporate brand as a potential employer. Our results show that advertising directed at consumers does indeed influence perceptions of potential employees. The results show that advertising influences perceptions of brand ability and potential development possibilities as well as attitude and intentions towards the brand. More specifically, by communicating a low price message in a more (less) creative way


24

the brand was perceived as more (less) smart and innovative, offering more (less) development possibilities, and attitudes and intentions towards the brand as an employer was higher (lower). We also found that the effect on employer attitude was mediated by development prospects. That advertising directed at consumers influence the perceived attractiveness of a corporate brand as a potential employer is something that advertisers should be aware of. In fact, rather than just borrowing from existing marketing communications in devising employment communications we suggest that advertisers should include this perspective in marketing decisions regarding brands and advertising. Overlooking the effects among potential employees can lead to consumer advertising that damages the employer brand, and thus, in the long run, makes it more difficult to attract skilled employees. From a traditional HR or recruitment perspective, the findings suggest that consumer advertising in general, and creative consumer advertising in particular, could be seen a recruitment tool. Thus, not only recruitment advertising should be considered when designing strategies and programs to attract employees. An argument that, on a more general level, supports those how have suggested that marketing and human resources managers should be working more together (Knox & Freeman, 2006). The present study does have a limited scope. One specific limitation pertains to the sample that consists of potential employees with limited work experience, who have been shown to be more influenced by consumer brand signals than those with more work experience (Wilden et al., 2010). Still, the fact that we used a real brand and real ads adds to the reliability of our findings. If perceptions of a brand as an employer are affected by one single exposure in an experimental setting, itâ&#x20AC;&#x2122;s not difficult to imagine the same effects after multiple exposures in a real life setting.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

25


26

FÖRORD: First there was nothing. Then there was Facebook Facebook har som vår tids stora kulturfenomen ändrat reglerna, inte bara för hur vi hanterar en stor del av vårt sociala liv men även vilka möjligheter och utmaningar företag ställs inför idag. Medan traditionell kommunikation tappar i effektivitet blir social media ett allt viktigare, om inte avgörande, verktyg inom kommunikation och marknadsföring. Vi lever idag i en värld av demokratiserad påverkan och som Brian Solis uttrycker det i sin bok ”Engage”, så är den bistra verkligheten för dagens företag ”to engage or die”. Med andra ord ställs helt andra krav på alla involverade i dagens kommunikationsutveckling, jämfört med när räckvidd och frekvens var avgörande för ett varumärkes framgång. I följande papper ger författarna oss en bättre förståelse för vilka faktorer som påverkar konsumenters interaktion med företag och varumärken på Facebook. Resultaten visar bland annat på att integritetsfrågor, användarnas inställning till Facebook, konsumentens allmänna engagemang i varumärken samt kön är viktiga faktorer. Vikten av att förstå vem man vill prata med och på vilket sätt, är alltså av yttersta vikt för en framgångsrik social mediestrategi som skapar engagerade och involverade ambassadörer. Något utvecklat ger det följande råd på vägen: - One solution does NOT fit all – hitta din egen väg - Förstå din målgrupp på djupet och leverera relevant innehåll som de vill engagera sig i och sprida vidare till sina vänner - Det är av yttersta vikt att ha en social media strategi på plats och inte behandla sina aktiviteter som kampanjinsatser - Sikta på kvalitet före kvantitet – se till att ha rätt fans på din sida - Sätt och mät relevanta sociala KPIs

jessica morales


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

27

USERS’ ENGAGEMENT WITH BRANDS AND COMPANIES ON FACEBOOK Sofie Bitter Alpen-Adria-Universität Klagenfurt Austria Sonja Grabner-Kräuter Alpen-Adria-Universität Klagenfurt Austria

PURPOSE OF THE STUDY Today’s consumers are exchanging substantial company- or brand-specific information online and it is incontestable that interactive social media influence consumer choices (Sands et al., 2011). Nowadays, it is not only the task of companies to find (potential) customers online, but to profit from user-generated content. Clemons (2009) argues that traditional advertising is losing its influence and reaching customers with traditional promotion gets considerably more difficult. Beyond that, customer engagement behavior in online social networks (OSNs) can be seen as a new potential for advertising. To take advantage of this opportunity companies need to anticipate consumers to take part, join their group(s) and actively produce content (Pagani et al., 2011). Marketers should know with whom they are dealing with in OSNs, i.e. what distinguishes engaged customers from users who do not interact with companies/brands. This study will shed light on some key variables and interactions among these variables that shape consumers’ interaction behavior with brands/ companies in online social networks. The broad theoretical framework guiding hypothesis development and selection of customer-based variables that affect users’ interactions with brands/companies is social cognitive theory. Bandura specified the “triadic reciprocal causation” of personal and environmental factors as well as behavior, which influence each other (Bandura, 1988, p. 276). The basic assumption for this study is that environmental and social perceptions (i.e. internet privacy concerns, trust in Facebook and trust in Facebook members and social norms) as well as personal perceptions (i.e. attitude towards Facebook, perceived behavioral control (or self-efficacy), brand engagement in self-concept, gender) influence the Facebook users’ interaction with companies/brands on the social networking platform. RESEARCH MODEL AND HYPOTHESES Customer engagement behavior Customer engagement has become a common buzzword in marketing research and practice (Brodie et al.). This study focuses on consumers’ interactions with companies/brands on Facebook as a specific manifestation of customer engagement behavior. Following the conceptualization proposed by van Doorn et al. (2010) and Verhoef et al. (2010) customer engagement can be “… defined as a customer’s behavioral manifestations that have a brand or firm focus, beyond purchase…” (van Doorn, et al., 2010, p. 254). For customers, OSNs offer new possibilities, turning their passive position into active involvement; they comment, review and publish information about brands online. Such online interaction between customers and companies is needed to build long-term relationships, where customers do not only acquire the product or service but will start to recommend it (Smith, 2009). However, socialization processes such as interactions in OSNs require certain antecedents to be successful. Recent research on selected issues of trust and privacy in OSNs makes it evident that these concepts are critical and crucial regarding users’ participation and disclosure of information (e.g., Chen and Sharma; Fogel and Nehmad, 2009; Krasnova et al., 2010; Shin, 2010; Stutzman et al., 2011).


28

Trust in Facebook and in its members OSNs foster socialization and therefore, the success of OSNs is largely determined by social factors such as trust (de Souza and Preece, 2004; Toral et al., 2009). Trust has to be seen as an important catalyst in facilitating social interaction and making virtual communities vibrant (J.-J. Wu et al., 2010). Similarly, in their conceptual model for customer engagement van Doorn, et al. (2010) emphasize that trust has to be considered a vital factor influencing (positive) customer engagement behavior. Trust in OSNs exists both on a micro- and macro group level. At the macro level, the social network site captures both characteristics of an organization (the network provider) and technology (the Internet serving as a transmission medium for online activities). At the micro level, online users engage and interact with each other and have also to be regarded as trust-objects. We propose that even higher levels of trust are required for users’ interaction with companies/brands in OSNs since additional information is revealed compared to those users who solely interact with “friends”. H1: Trust in Facebook / H2 Trust in Facebook members is positively related to the user’s interaction with companies/brands on Facebook. Internet Privacy Concerns Individuals with higher concerns about their privacy will be more restrictive regarding their shared information (Nov and Wattal, 2009; Young and Quan-Haase, 2009). Since OSNs are a field of marketing opportunities for businesses that engage in Internet Marketing (Lorenzo- Romero et al., 2011) and have enormous commercial potential (e.g., Dinev et al., 2009), it is essential for companies to take users’ privacy concerns seriously. Companies will only be in the position to profit from (potential) consumers in OSNs, if those users are also willing to disclose information and are not hindered by their concerns regarding privacy. H3: Internet privacy concerns negatively impact the user’s interaction with companies/brands on Facebook. Attitude, social norm and perceived behavioral control One social-cognitive theory that has proven its worth in understanding consumer behavior is the theory of planned behavior (Ajzen, 1991, 2002). In various studies, the theory of planned behavior defines the relationship of attitude, subjective norm and perceived behavioral control towards the intention to conduct a certain behavior (e.g., Hsu and Chiu, 2004; Lin, 2010; Lu et al., 2009; Lwin and Williams, 2003; Pelling and White, 2009). Attitude refers to the users’ general viewpoint and personal position towards a behavior; subjective norm or social norm to the influence of other people on the users’ behavioral actions; and perceived behavioral control to the perceived ability to perform a certain behavior or sequence of behaviors (Ajzen, 2002). Conceptually, perceived behavioral control is more or less indistinguishable from perceived self-efficacy, a key mechanism in the framework of social cognitive theory. For this study perceived behavioral control reflects the user’s belief that s/he is capable using Facebook. H4: The user’s attitude towards Facebook / H5 subjective norm / H6 perceived behavioral control is positively related to the user’s interaction with companies/brands on Facebook. Brand engagement in self-concept A brand is a crucial criterion for the engagement of customers after purchase (van Doorn, et al., 2010). In this study, the focus is not on a specific brand but on a person’s “general engagement with brands”, a concept called brand engagement in self-concept (BESC) developed by Sprott, Czellar and Spangenberg (2009). This concept assumes that


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

29

consumers have a basic disposition towards brands affecting various brand-related cognitions, behaviors, etc. and capturing the extent to which individuals assume brands are part of their self-concept. As an individual difference variable BESC arrays consumers along a continuum ranging from low (brands are not seen as important elements of self-concept) to high, where consumers identify with brands and have special bonds with them (Goldsmith et al., 2011). It can be expected that consumers with higher levels of BESC are also more willing to interact with companies/ brands in OSNs. Hence, H7: The user’s BESC positively influences interactions with companies/ brands on Facebook. We also asked whether BESC would moderate the relationships between attitude, perceived behavioral control, subjective norm and interaction behavior with companies/brands. E.g., we propose that the effect of the user’s attitude towards Facebook on the interaction behavior with companies/brands on Facebook will be stronger for individuals with a high level of BESC than for users with a low level of BESC. H8: BESC moderates the relationship of attitude / H9 perceived behavioral control / H10 subjective norm on the user’s interaction behavior with companies/brands on Facebook. Finally, we include gender as a further predictor variable in our study and also address its moderating influence. Previous research found considerable gender differences for online communication (e.g., Wang et al., 2011), the usage of online social networks (e.g., Mazman and Usluel, 2011), time spent on online social networking sites (e.g., Cha, 2010) as well as indirect effects of gender on brand communication (e.g., Acar and Polonsky, 2008). H11: Gender will influence the interaction behavior of users with companies/ brands on Facebook. H12: Gender moderates the relationship of the user’s BESC on his/her interaction behavior with companies/brands on Facebook. EMPIRICAL INVESTIGATION Data collection and sample characteristics The data used to test the proposed hypotheses were collected by means of an online questionnaire. The survey addressed solely Facebook users and thus the link to the questionnaire was only spread via the online social network “Facebook”. A total of 427 responses were received from which 358 were valid for analyses. The sample consists of 43.3% men and 56.7% women and is dominated by 20-29 year-olds (63.4%), followed by the groups of 30-39 and <20 year olds. Measurements and construct specifications All measurements and items employed for the constructs, being the independent variables in this study, were taken from literature and measured on a 7-point Likert-type scale (7=strongly agree to 1=strongly disagree). If necessary, items were adapted to the Facebook respectively the OSN context. Items are listed in table 1. Confirmatory factor analysis (CFA) was conducted to test the constructs using AMOS 18. All items loaded significantly above the cut-off value 0.6. All constructs in the analysis showed satisfying Cronbach’s alpha values (> 0.7) and confirm, together with the average variance extracted (> 0.5) and composite reliability (> 0.7), convergent reliability. Discriminant validity was tested with the Fornell-Larcker criterion; all constructs were below the


30

critical value of 1 (for details see table 2). The dependent variable, interacting with companies/brands, was measured on an 8-point frequency scale (8=more than once daily to 1=never). Participants were asked if they suggest sites from companies/brands to their friends, use the “like” button for sites from companies/brands, comment the status/links from companies/brands, join groups from the company/brand, write on the wall of companies/brands and finally accept or reject invitations from companies/brands. The mean of those six items was calculated and the new variable was called “interaction with companies/brands on Facebook”. We split the participants into two groups, those users interacting with companies/brands at least once a month, referred to as “repeated interactors” and those user never interacting with companies/brands on Facebook, referred to as “non- interactors”. Results In order to reveal the differences between the two groups of “repeated interactors” and “non-interactors” in respect of the applied constructs we conducted a t-test. All constructs show significant differences between repeated interactors and non-interactors (except for trust in Facebook being only significant at a 10 % level). See table 3. Due to the dichotomous character of the dependent variable (repeated interactors versus non- interactors), we applied binary logistic regression analyses using PASW (SPSS) 18.0 to test the established hypotheses. All variables were standardized (M=0; SD=1) before analyses. The interaction effect model (table 4) shows good model-fit values of the Cox & Snell R Square (.433) and the Nagelkerke R Square (.581). Additionally, the overall sample of correct cases classified is 82.4 % with 78.0 % of repeated interactors and 85.7 % of non-interactors correctly identified. The coefficients for attitude towards Facebook (β=1.598, p<0.01) and for brand engagement in self-concept (β=1.762, p<0.01) are highly significant and positive, accepting H4 and H7. Also the predictor gender is highly significant (β=1.683, p<0.01) confirming H11 that men are more engaged with brand-related activities on Facebook. Additionally, a clear negative and highly significant influence of Internet privacy concerns on the interaction behavior is recorded (β=-.781, p<0.01), accepting H3. The hypothesized assumptions (H1, H2, H5 and H6) regarding the constructs trust in Facebook, trust in its members, subjective norm and perceived behavioral control have to be rejected due to non- significance. Looking at the interaction terms, it can be highlighted that brand engagement in selfconcept moderates the relationships between attitude towards Facebook (β=1.073, p<0.05) as well as perceived behavioral control (β=-1.152, p<0.01) and interaction behavior. Additionally, gender moderates the relationship between the BESC construct and interaction behavior (β=-1.359, p<0.05). Hence, H8, H10 and H12 are accepted. H9 is rejected. DISCUSSION AND IMPLICATIONS This study contributes to the current research field by further investigating the complex and vast environments of online social networks in view of marketing aspects. Overall, the results demonstrate that interactors and non-interactors vary significantly in their environmental and personal perceptions. By identifying antecedents for customer engagement behavior in OSNs we found that, against our expectations, trust (in Facebook or its members) is not significantly differentiating interactors from non-interactors. However, the results confirm the importance of other determinants of customer engagement behavior on Facebook. Especially internet privacy concerns, the users’ attitude towards Facebook and brand engagement in self-concept are crucial predictors for users’ interactions with companies/brands. Beyond that, the findings suggest that there is a significant gender difference among users who repeatedly interact with companies/brands on Facebook and those who don’t. The analysis also


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

31

contributes to the current field of research by uncovering interaction effects, being of interest for theory as well as for practice. We found that the relations between user’s attitude towards Facebook, perceived behavioral control, gender and the user’s interaction behavior with companies/brands are moderated by brand engagement in self-concept. For example, a positive attitude towards Facebook has a much stronger impact on interaction behavior with companies/brands for users with high BESC levels. Therefore, the study has several implications for the role of advertising in OSNs. First of all, different types of Facebook users have to be distinguished. The results showed that there are differences among repeated interactors and non-interactors with companies/brands on Facebook. Secondly, by knowing that repeated interactors are not as much concerned about their privacy as non-interactors, it can be assumed that those users are more willing to add information and produce content. Thus, advertisers could engage in specific communication tactics, for instance discussions on new products. The active involvement of customers could be a key criterion for successful promotion in OSNs. This research also supports the argument that it is crucial for companies to find an acceptable balance between harvesting personal customer information and causing consumers’ privacy concerns (Zimmer et al., 2010); companies need to take privacy concerns seriously and should invest in trust building measures (Chellappa and Sin, 2005). Finally, advertisers should try to concentrate on users with a positive attitude towards Facebook and a high level of brand engagement, as their chances for repeatedly interacting with companies are much higher compared to other users. To sum up, the contribution of this paper lies in providing empirical evidence on the role of some key variables in explaining consumers’ interaction behavior with companies/brands on Facebook. It is important to further investigate online social network users who actively engage with companies/brands, since the relevance of online social media marketing is continuously augmenting. OSNs are places of information exchange where advertisers meet their (potential) consumer on a “friendship” basis. Thus, “…not incorporating social networks as a part of the marketing mix is not only poor customer service, but also a surefire way to lose consumers.” (Pookulangara and Koesler, 2011, p. 352). The tendency that interacting in online social environments is part of daily life, especially for younger generations, puts even higher pressure on future empirical research, also in the area of advertising. It is central to develop a profound understanding of consumers in OSNs in order to properly target this fast increasing group of customers and to adapt and restructure advertising tactics accordingly. Limitations associated with the study include inter alia the limited number of selected antecedents, the sole focus on the OSN Facebook, a sample mainly originating from Europe and the non-random nature of the convenience sampling.


32

Förord: Vart är varumärkeslojaliteten på väg med ökad användning av interaktiva media? 100 % varumärkeslojalitet är ett Nirvana för de flesta marknadsförare: konsumenter som en gång köpt varumärket köper det igen och igen och igen. Sättet att skapa den maximala varumärkeslojaliteten har enligt klassisk teoribildning och praktiskt sätt att arbeta varit en ökad och förbättrad kommunikation, speciellt om man använder direkta media mot utvalda konsumentgrupper som stöd till publika media. Men frågan är om detta framgångsrecept fortfarande är valitt i tider, när den genomsnittliga konsumenten har obegränsad tillgång till information och har tredubblat tiden han/hon spenderar på interaktiva media, framför allt Facebook, från 43 till 130 minuter om dagen? Ett allmänt accepterat antagande är att interaktiva och digitala kommunikationssystem kan användas till att öka varumärkesvärdet och bygga starkare varumärkesrelationer till en lägre kostnad. Andra ifrågasätter detta. Lätt tillgänglig produktinformation och ökad interaktion och kunskapsöverföring mellan konsumenter kan lika gärna leda till minskad varumärkeslojalitet, säger de. Om antagandet håller, blir produktoch servicevarumärken allt mer irrelevanta och onödiga, när allt fler konsumenter säger att de inte har någon varumärkespreferens. Interaktiva media skulle alltså driva på en utveckling tillbaka till de tider när allt fler produkter blir utbytbara bulkvaror? Denna studie under ledning av Don E. Schultz belyser samband mellan ökad användning av interaktiva media och varumärkeslojalitet. Hypotesen är att en ökad användning av interaktiva media korrelerar med minskad varumärkeslojalitet i dagligvaruhandeln. Studien visar på korrelation, men kan inte fastställa kausalitet, den ger intressant vägledning, och all anledning för varje varumärkesägare att säkerställa resultaten av sina satsningar på interaktiva media som Facebook. Resultaten mätt i varumärkeslojalitet och –värde kanske är helt andra än de förväntade.

Peter GRANström


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

33

RE-THINKING BRAND LOYALTY IN AN AGE OF INTERACTIVITY Don E. Schultz Medill School of Journalism, Media, and Integrated Marketing Communications Evanston, US Martin P. Block Medill School of Journalism, Media, and Integrated Marketing Communications Evanston, US

INTRODUCTION Brand loyalty is one of the most treasured of all marketing organizational results. (Kotler& Gertner, 2002) There is evidence that loyal, continuously purchasing customers cost less to maintain over time and provide ongoing income flows to the brand owner. (Aaker, 1991) Thus, the brand manager’s primary goal, whether for a product or a retailer, is to build and maintain a cadre of brand loyal customers over time. One of the primary tools used to build and maintain brand loyalty is ongoing communication between the brand and the customer. (Holland, 2001) From that, it is believed on-going brand relationships can be developed and maintained. There are numerous ways to initiate and deliver customer brand communication programs. Some of the newest and most interesting are through digital, interactive social media systems such as Facebook, Twitter and the like. (Safko & Brake, 2009) A common assumption is that these interactive and digital communication systems can be used to enhance brand value and build stronger relationships with customers at lower cost. (Raggio & Leone, 2007; Keller & Lehmann, 2009; AdAge, 2010) Common wisdom suggests interactive media are simply additional marketer-controlled communication distribution channels. (Schultz, 2004) Thus, they can be used to create ongoing, interactive contacts and conversations with customers and prospects. (Wilson, 2010; Kabani, 2010) If so, these new communication systems can significantly improve and/or develop stronger, on-going customer-brand relationships resulting in brand loyalty. (Jang, et al, 2008; Chan-Olmsted, 2006) Many industry articles/cases have focused on examples or demonstrations to validate these beliefs. (Dickey & Lewis, 2011; van Doorn, 2010) Some however, question whether easily sourced product information, along with enhanced and expanded consumer-to-consumer experience and knowledge transfer methodologies among brand users, could just as easily result in decreased customer brand loyalty. (Lee, et al, 2011; Xingyuan, et al, 2010; Schultz & Block, 2011) At this point, no definitive answer has emerged. (Ryan & Jones, 2009) RESEARCH QUESTIONS In this exploratory study, a research methodology was developed to investigate the following: ­­ Does increased use of interactive media, particularly Facebook, leads to » measurably decreased brand preference and, ultimately, brand loyalty » If decreased loyalty is observed, does it occur at the retailer or product level or both? » Does this observable impact occur immediately or over time?


34

THE DATA SET AND DATA MANIPULATIONS The data set was taken from consumer responses to the Monthly Consumer Survey (MCS) online survey and the Media Behavior and Influence studies (MBIs), both conducted by BIGinsight (www.BIGinsight. com™), Worthington, OH, in Aug. 2010. In the MSC studies, U.S. consumers report their product purchases, media usage, social activities and a host of other factors, providing an on-going tracking system used primarily by retailers and brand managers. The monthly online surveys generate consumer samples of 8,000 to 9,000 respondents. This data set, gathered in August, 2010, consisted of 8,205 individual responses. This type of data has been widely used in academic research. To determine interactions among respondents, key questions such as “do you recommend products to others”? and “do you seek advice from others on products?’ are asked. While self-reported, it can be inferred as to which customers are “product recommenders”. The other data source, the MBIs data file (Media Behaviors and Influence studies), is an online bi-annual study, also conducted by BIGinsight™ (www.BIGinsight.com™). These studies capture consumer reported media usage and influence. This data set in August, 2011 contained 23,235 individual online responses to various media usage and influence questions. Because of the panel size, a nationally projectable sample is developed each time MBIs is conducted. These two data sets (MCS and MBIs) were combined for this study, resulting in a matched panel file of 897 unique individuals. Thus, those who were “product recommenders” could be identified and their media usage and influence could be connected. While consumer response data is confidential, each response is numbered, thus allowing the anonymous comparison of responses. A CRITERION MEASURE FOR BRAND LOYALTY Developing a criterion measure for brand loyalty is one of the most challenging brand issues. (Keller & Lehmann, 2006; Schultz & Block, 2011) Sales results, either generated internally or obtained from external third parties, are commonly favored. (Chaudhari & Holbrook, 2001) Other methodologies are extrapolations of on-going consumer panel data, retail store purchase records and the like. (Churchill, 1995) All have advantages and disadvantages. For this study, longitudinal changes in the Net Promoter Scores (NPS) for retail food stores and two consumer product brands (ready-to-eat cereals and salty snacks, i.e., chips, pretzels, etc.) were used. While the NPS approach has been criticized academically, (Hayes, 2008; Keiningham, et al, 2007) the broad and ongoing use of the measure by brand owners gives it “marketplace currency”, i.e., more than 100 companies globally use the approach to measure brand and organizational loyalty. (Bain and Company, 2012) Using the reported “give or seek advice” questions in the data set, NPS scores, using the same approach as Reichheld (Reichheld, 2003)) were created for a set of food retailers and two fmcg food products. (Using a 10 point scale, with Promoters scoring the brand 9 or 10 and Detractors scoring 0 to 6, the Detractor scores are subtracted from the Promoter scores to create the modified NPS score used in the analysis.) Since there are no national food retail chains in the U.S., a set of dominant retailers was used. The fmcg product categories were selected based on their historically large media advertising spending patterns. (Schultz, 1998; Sinclair, 2007)


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

35

The NPS scores were derived from the 897 unique individuals who said they either “gave or received recommendations” from others. Those were then connected to their responses on media usage and influence. From that, favorite retailers and their favorite product brands in the two product categories were identified. Consumers reported media behaviors in 31 media forms, (Facebook usage was employed as the criterion variable for interactive media). These were analyzed over several time periods to determine increases or decreases in media usage and influence, particularly on Facebook. Results of that analysis are in Exhibit 1. Media Usage in Minutes per Day for Selected Media and Facebook The dramatic growth of Facebook usage is clearly shown in the amount of time spent with that media form. From December, 2008 to June, 2011, minutes per day increased from about 43 to over 130. Most other media forms showed little growth. HYPOTHESES TESTED Schultz and Block (2011) have used similar data to illustrate the impact of media usage and impact on product purchases. We extend their work by focusing on social media, specifically, Facebook usage. “Brand preference” (loyalty) for both retailers and brands was derived from respondent “write in” data in the combined MSC and MBIs data sets. Media usage and influence was taken directly from consumer responses. Using that data, the following hypotheses were tested. H1: The reported increased use of social media will be positively correlated with brand preference (loyalty) by respondents for retail food stores. H2: The reported increased use of social media will be positively correlated with brand preference (loyalty) among users in the ready-to-eat breakfast cereal category. H3: The reported increased use of social media will be positively correlated with brand preference (loyalty) among users in the salty snack category. As noted, favorite retail store measures came from write-in responses in the online MSC and MBIs study questionnaires.The same write-in approach was used to measure “favorite” ready-to-eat cereals and salty snacks. Since no retail or product brands appeared in the questionnaire all listings came directly from respondents. Additionally, a separate area was provided for respondents to state “No Preference” in both the retail and fmcg product categories. Thus, “No Preference” is an active choice by respondents, not simply those who did not respond with a brand. METHODOLOGY & FINDINGS Store brand preference scores were first combined with product brand preferences. The results of those analyses are shown in Exhibit 2 (Ready-to Eat Cereals) and Exhibit 3 (Salty Snacks) and favorite retailer. Ready-to Eat Cereals Store and Brand Preferences Salty Sack Store and Brand Preferences As shown, the retail store NPS score is always greater than the manufacturer product brand score, with the exception of Kashi cereal. This means consumers are more loyal to their food store brand than to their manufacturer branded products.This raises some interesting questions for both retailers and brand marketers.


36

Most interesting is that the largest single category for both retail stores and product brands is “No Preference”. Since respondents had to check this questionnaire box, this appears to be a valid measure. This illustrates that both retail store brands and manufacturer brands are losing brand preference, and, by implication, brand loyalty over time. That finding is, or should be, quite disconcerting to all brand managers, whether they are at the store or manufacturer level. Some of those concerns have been addressed in earlier papers (Schultz and Block, 2010, 2011 & 2012) so they are not repeated here. Having determined that NPS scores can differentiate between brand and store loyalty, the next question was: why has that occurred and what factors are driving the change? We believe the new interactive media forms and the greater knowledge among consumers about brands and brand value may well be creating those changes. Therefore, the amount of social media being consumed by questionnaire respondents during the time of the study was determined (see Exhibit 1) Facebook usage (consumption) increased in all measurement periods among all respondents. At the same time, the No Brand Preference category for both retailers and brands was also growing, i.e., brand preference was declining. There is a strong and positive correlation between the increased use of social media and the growth of the “No Preference” categories for both retailers and brands. Thus, we reject the three Hypotheses presented above. We did not find that increased social media usage is correlated with increased brand preference. We found just the opposite, i.e., increased social media usage is highly correlated with declines in store and brand preferences. To our knowledge, this is the first available research that shows that increased use of social media by consumers has a negative impact on both store and manufacturer brands. Using the consumer data on retail stores, manufacturer brands and the declining preference consumers seem to have for those brands, the results were tested against Facebook usage using correlation analysis. That correlation is 0.3 thus, lending credibility to our hypotheses. Clearly, we cannot state causality but, longitudinal analysis seems to add credibility to our findings. The triangulation of increasing No Preference scores for retailers and manufacturer brands and increasing use of social media (Facebook in our study) points to some type of connection. We therefore believe there is sufficient support for our findings to provide a sounding board for discussion at the upcoming ICORIA conference. QUESTIONS THIS EXPLORATORY RESEARCH MIGHT GENERATE AT THE ICORIA CONFERENCE Below, are some suggestions which might frame a discussion of our findings: 1. Are the historical brand loyalty measures of demographics and psychographics still the right variables? Are those still relevant or should researchers move to actual measurable consumer behaviors? Is brand loyalty simply an artifact of our data gathering? If brand loyalty exists, is it growing or declining? What are the contributory factors? 2. Interactivity and consumer networking are totally new areas for brand study. How do both of these impact consumer choice? Is one now loyal to the recommendation of peers or is brand loyalty still self-determined? How important are the views of others versus the claims of the brand owner in driving brand purchase continuity?


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

37

3. What’s the impact of external, mass marketing and mass communication on building consumer continuity of purchase and brand loyalty? Is that era over? 4. Have we entered an age of “consumer indifference to brands and brand loyalty” as a result of government controls and assurances? Are we plummeting into a “no brand, everything a commodity” marketplace? 5. Do consumers live in a world where unlimited amounts of consumeravailable information now make product and service brands irrelevant and unnecessary? A number of other questions can be posed. What seems obvious, based on our investigatory research is that the entire subject of brand loyalty needs to be re-thought. Given the financial commitment that marketers have in brands, the academic community may be the only impartial investigator available. Thus, we hope these findings will generate some initiatives that will start to answer these questions.


38


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

The faces of Advertising

39


40

FÖRORD: Innan möjligheten blir en nödvändighet Infotainment hör det nya seklet till. Inget snack om saken. Nu när all information finns på några knapptryckningars avstånd är inte bara smarta förpackningar av denna information möjlig, utan troligen snart att betrakta som en hygienfaktor. Att ständigt konsumera kunskap i munsbitar. Begreppet Augmented Reality har fortfarande den där speciella Si-Fi kittlingen; att lägga lager på verkligheten, som Tom Cruice i Minority Report. Men snart nog lär tekniken vara infångad i våra vardagsbeteenden. Frågan är bara hur och på vilket sätt den blir det. Och det är just dessa frågor som jag tycker denna uppsats sätter fingret på. Den stressande gissningen. Är det rätt att vara pionjär? Vem tar kostnaden? Leder den till ett ökat kundvärde som går att omsätta i ökad intäkt eller blir det ytterligare en kostnadspost i produktkalkylen för att hålla jämna steg med konkurrensen? Svaren låter vänta på sig, men helt klart lyckas författarna visa att ännu finns möjligheten att med tekniken framstå som intressant och nydanande. Och det är ju inte så illa.

jan lindforss


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

41

IMPROVING THE IN-STORE CUSTOMER INFORMATION PROCESS USING MOBILE AUGMENTED REALITY Philipp Spreer Georg-August Universität Göttingen Germany Katrin Kallweit Georg-August Universität Göttingen Germany Klaus Gutknecht University of Applied Sciences Munich Germany

INTRODUCTION The bookselling industry is facing a far-reaching shakeup resulting from a shift in sales from local stores to online shops. The most significant reason is that online retailers can offer a wide range of product information like reviews, customer ratings or multimedia content while bricks and mortar booksellers are usually limited in terms of their product information due to spatial restrictions at the PoS. Additionally, in times of mobile technologies such as smartphones or Tablet-PCs customers are more likely to switch information and transaction channels within their customer buying process. Therefore, multi-channel retailers are anxious to support these channel switches within their own distribution systems to avoid losing potential customers to online competitors. The US market research company Gartner identified AR as one of the top ten disruptive technology trends for the coming years that causes major change in industry dynamics and consumer behaviour (2008). Using such an application, the image of the real environment on a mobile device will be extended by a digital dimension, for example, to integrate additional information or multimedia content. Simply by pointing a smartphone or Tablet-PC camera at an object, the user receives augmented information on the product displayed on the device screen. AR is generally applicable for solving the problem of displaying multimedia content in local retailing. This study shows the impact of an AR application on supporting the customer information process at the PoS using the example of a book purchase. The main focus will be the provision of media content that can only be accessed by the customer online. It is also important to consider the influences on the intention to use an AR application. LITERATURE REVIEW AND HYPOTHESES The idea to augment the physical reality through a digital dimension is not a new concept, (e.g. Sutherland, 1968) but the opportunities to realize AR applications have changed dramatically in recent years due to technological advancements. Despite the immediate relevance for retailers, little is known about AR from a marketing perspective (Olsson and Salo, 2011). Most AR scientific work has focused on IT (e.g. Güven et al., 2009) with early marketing papers aimed at eliminating the technical barriers of usage, such as weak usability or functional deficits (e.g. Zhu et al., 2006; Chang and Tan, 2010; Krevelen and Poelman, 2010). Inspiring experiences through AR have now become the focal point in science (e.g. Olsson and Väänänen-Vainio-Mattila, 2011). Even though there is a lack of universally applicable findings, early explorative studies indicate the great potential of AR in a retail context. For example, Välkkynen et al. (2010) show that three-dimensional interactive


42

illustrations on a product package create a more stimulating experience than non-interactive services using a small-scale experiment in a laboratory environment. Olsson and Salo (2011) criticize this and other findings for their general lack of meticulousness and empiricism in AR marketing science with reference to a meta-analysis conducted by Dünser et al. (2008), which showed that only approximately ten percent of the publications dealing with AR included some sort of user evaluation. Using two different AR applications, they classify a typical AR user and identify strengths and weaknesses. As mainly AR developers and heavy users are interviewed, these evaluations do not necessarily correspond to the majority of customers, meaning that the study does not produce comprehensive insights. In addition, the importance of information as one of AR’s core strengths is not requested. As a consequence, we offer the following hypotheses: H1: AR-users rate the offer of information in the store better than non-users. H2: AR-users declare having found all the information they needed in the store more often. Ganapathy et al. (2011) also mainly survey heavy-users and focus on the accepted delay using the example of a wine application. They find that tolerance towards waiting while using the AR application rises with the product price. However, the intense guidance, the small sample and its structure prevent the composition of generalizable findings. Olsson et al. (2011) note that the design of an AR application must always take into account the users’ expectations and the user experience. They analyze the impact of different components of AR technology on user experience in a shopping mall. User expectations and usage requirements are found to be rather pragmatic and utilitarian but such emotional aspects as customer inspiration and surprise also seem to play a prominent role. To date, Olsson et al. (2012) have conducted the most extensive empirical study dealing with AR in retail by using the Technology Acceptance Model for Mobile Services (TAMM, see Kaasinen, 2005). The authors confirm their prior results (Olsson et al., 2011): the quantitative online survey also attests that functional and solution-oriented applications are rated more positively than enjoyment-oriented ones. However, as the authors choose an impersonal method of collecting data it cannot be assured that all the participants understood AR as a new technology. As the TAM is proved to be appropriate in the context of AR applications in retail, it should also be used in the present study. In particular, the perceived usefulness (PU), the perceived ease of use (PEU) and the perceived enjoyment (PE) are considered in more detail. Originally, the perceived usefulness refers to a person’s feeling that the use of a certain technology improves his or her work efficiency (Davis, 1989). In this study, we understand the PU as the ability to access product information via AR applications. The PEU is generally interpreted in accordance with Davis: as the user-friendliness of the application. It has been detected that the PU has the biggest effect on the acceptance of users (e.g. Adams et al., 1992; Mahmood et al., 2001; Taylor and Todd, 1995; Davis and Venkatesh, 2000). In contrast, the PE influences the acceptance much less than the PU and the PEU (Davis et al., 1992; Igbaria et al., 1994). However, there are some exceptions within this field of research (e.g. Atkinson and Kydd, 1997; Moon and Kim, 2001), where the PE and not the PU are found to have the strongest influence on user acceptance. These studies examine, for instance, the user acceptance of the World Wide Web, home technologies, games or training methods


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

43

based on games for work-related information systems (e.g. Childers et al., 2001; Heijden, 2004, Nysveen et al., 2005). As AR bears many similarities to the aforementioned technologies, we offer the following hypothesis: H3: The perceived usefulness, perceived ease of use and perceived enjoyment deter- mine the (re-)usage of AR applications. STUDY The previous studies did not take place in a real retail environment and were thus faced with the challenge of representing the application potential of AR to the subjects (e.g. Olsson et al., 2011; Olsson and Salo, 2011; Olsson et al., 2012). A context for the evaluation and expectations towards AR has been created by different scenarios (Olsson et al., 2009) or specific use cases (e.g. furniture purchase at Olsson et al., 2012). The data collection in this experimental field study delivers evidence from real use cases. For this purpose a separate application based on the AR browser “junaio” has been developed by Metaio. In this browser, inspiring and informative content such as animated book covers, book titles, blurbs, interviews with the author, user comments and links to the online shop for selected bestsellers are stored. A smartphone or Tablet-PC with “junaio” browser installed makes it possible to scan books so that multimedia content can be retrieved. To investigate the effect of such an AR application, in January 2012 randomly selected visitors to a highly frequented bookstore were interviewed about their information behaviour and subjective evaluation of the AR application. After a brief introduction, about half of the subjects tested the AR application independently and answered another set of questions to evaluate the AR application independent of the control group. They evaluated AR in terms of perceived usefulness, ease of use, perceived enjoyment as well as the intention to reuse AR. RESULTS AND DISCUSSION For the present study, we interviewed 100 visitors in a bookstore. To ensure high data quality, four questionnaires had to be excluded, leaving a total of N = 96 data records for further analyses (treatment group = 46, control group = 50). The sample was an approximate representation of the retailers’ customer demographics, with seventy-two percent of the interviewees being female and twenty-eight percent male. As far as the age of the respondents was concerned, there was a big spread: most of them were aged between twenty and twenty-nine (26 %), thirty and thirty-nine or fifty and fifty-nine years old (20.8 % each). Forty percent of the participants owned a smart device (smartphone and/or Tablet-PC), which is slightly above the average in Germany (Nielsen, 2011). One of the major challenges for booksellers is to offer adequate information and notably multimedia content at the point of sale. AR has the potential to reduce these perceived deficits in the presentation: AR users evaluated the information provided at the PoS significantly better than store visitors who did not have access to augmented content (level of significance: p < .05). This is emphasized by the fact that AR users stated that they had found the relevant information more often at the PoS (p < .05). The acceptance of new technologies is usually measured with the Technology Acceptance Model (TAM, Davis, 1986). Since the participants of the treatment group already used the application during the experiment, the intention to reuse the AR application serves as a dependent variable instead of acceptance. The multiple regression analysis shows a highly significant model and explanation power of all three acceptance components in relation to the intention to reuse of 82.8 percent (corr. R²). If the independent variables are considered separately, it becomes clear that


44

the perceived usefulness (p < .001) and perceived enjoyment (p < .001) significantly influence the intention to reuse the application. In contrast, the influence of the perceived ease of use is not statistically significant. Furthermore, the intention to reuse is higher for men (p < .01) and younger users under thirty years of age (p < .05). The analysis of the premises of the regression model did not show any strong interrelations that could reduce the quality of the outcome. The strong influence of the usefulness of an AR application has already been postulated. In contrast, the impact of the perceived enjoyment on the intention to reuse the application is rather surprising. At this point, some interesting group contrasts can be observed: the perceived enjoyment is stronger for men by trend (p < .1). The same applies to young users (fourteen to twenty-nine years old) (p < .01). This analysis is the first experimental study in a real AR trading environment to examine the issue of acceptance and effectiveness. The subjects were randomly selected and used a professionally designed AR application. However, the self-selection bias cannot be excluded completely as we dispensed with a sampling frame. There are some further limitations that must be considered in order to generalize the results: for example, the sample represents the study partner’s structure of the target group but does not represent the German population. In addition, there are restrictions resulting from the sample size. Furthermore, price-related variables could not be collected due to the specific legal framework in Germany (fixed book price). A replication of the study in other sectors or the analysis of the mere exposure effect as a potential explanation for the importance of the enjoyment component may yield further insights. Various studies show that the explanatory content of TAM increases when modern variables are considered (e.g. Davis and Venkatesh, 2000; Bhattacherjee and Sanford, 2006; Sun und Zhang, 2006). It is also assumed that variables such as knowledge or experience of AR and the innovativeness of a person can have moderating effects in this research context (Barki and Hartwick, 1994). In addition, science and practice have an equally strong interest in the importance of specific features of AR applications. Based on the findings of this study, further research on channel interfaces to analyse the effectiveness of AR in a catalogue, an online shop or as an out-of- home application is necessary. Regarding the retail industry, the use of AR provides an opportunity to close the information gap at the point of sale and have a positive impact on customer satisfaction. Retailers are now facing the challenge of designing useful applications that provide strong functional benefits with enjoyment-oriented elements. Particularly information that is known from online shopping such as “other articles by the author” or “customers who bought X also bought Y” was mentioned by the subjects. A direct influence of the PEU on the intention to re-use the AR application could not be established. This may be down to the fact that using AR is familiar due to the customer’s internet and smartphone experience, as Venkatesh (2000) assumes that individuals rate the simplicity of use according to specific anchors. This may help retailers to implement an AR application because numerous companies already operate their own online shops and have the appropriate content. One fundamental decision to be made is whether stationary Tablet-PCs at the PoS shall be provided by the retailer or if customers will use their own smartphones for AR. To increase usage, additional sources of information such as printed flyers or specially trained salespeople should be used.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

45


46

Förord: Hur används reklamspel i marknadsföring Inget företag eller reklambyrå vågar negligera digitala medier i sin marknadsföring. Många seminarier och case-studier söker svaret på hur skillnaden mellan ”köpta” och ”förtjänade” mediakanaler påverkar konsumentens attityd till budskap och varumärke. I detta gränsland har reklamspel vuxit fram. Spel som tagits fram på uppdrag av en annonsör för att nå ut till attraktiva målgrupper, underhålla och indirekt bryta ner deras motstånd mot kommersiella budskap. I denna studie har författarna sökt kategorisera reklamspel utifrån hur väl de passar varumärkets personlighet, förstärker varumärkes erinran, budskapsförståelse och påverkar attityder hos konsumenterna. Att döma av resultatet nöjer sig de flesta företag med att hoppas på att spelen är tillräckligt underhållande. Kraven på att spelen skall ha en tydlig koppling till varumärke och budskap är lågt ställda. Om det räcker att i en digital värld låta spelmakarna få fria tyglar och hoppas att underhållningsvärdet av spelet skall färga av sig på uppfattningen av varumärket är en öppen fråga. Men i den analoga värden skulle nog de flesta uppdragsgivare svara annorlunda.

Peter bergendahl


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

47

HOW DO ADVERGAMES PERSUADE? A CONTENT ANALYSIS OF SELECTED PERSUASION MECHANISMS Martin K.J. Waiguny Auckland University of Technology New Zealand Ralf Terlutter Alpen-Adria-Universitaet Klagenfurt, Austria Johanna Roettl Alpen-Adria-Universitaet Klagenfurt, Austria Maria Groechenig Alpen-Adria-Universitaet Klagenfurt, Austria

INTRODUCTION AND PREVIOUS RESEARCH Advergames have become a common medium for brand advertising. Initially, advergames were provided on websites to encourage the customers to visit and revisit websites (Santos et al., 2007). However, ongoing research and development suggests that advergames are also a powerful tool to convey brand related messages. They are seen to have a rhetorical potential to inform consumers about brand features, persuade them about the brand, and to narrate the brand (Bogost, 2007; Ritterfeld and Weber, 2006). Applying content analysis, our paper investigates to what extent advergames make use of some of these mechanisms of rhetorical and narrative persuasion. Research on Advergaming practices Most of the recent content analyses of advergames are embedded in content analyses of food and toy manufacturers’ websites and merely analyze the presence or absence of advergames on the website. The studies clearly indicate that advergames are already a common form of communication to children in the USA (Culp et al., 2010; Weber et al., 2006; Moore and Rideout, 2007; Alvy and Calvert, 2008) and UK (Dahl et al., 2009). However, these studies do not elaborate on the persuasion mechanisms applied in advergames. More comprehensive insight into advergames, and evidence for persuasion tactics in advergames being applied to children, is provided by the studies of Moore and Rideout (2007), Lee and Youn (2008) as well as Lee et al. (2009). The majority of websites targeting children offer advergames and, in over 80% of these advergames include the brand and reinforce children to revisit and replay the game (Moore and Rideout 2007). After analyzing the websites of leading enterprises in advertising expenditures in the USA, Lee and Youn (2008) concluded that most of the advergames were in the food industry, that game-product congruence is quite low and that generating fun is the main aim of the games. In a second study, Lee et al. (2009) analyze 251 food advergames for kids, demonstrating that the logo is the most common identifier. They find that additional placements are present and often the additional placements are incorporated as active elements in the gameplay. Although the latter studies already answer some questions about different forms of brand integration and viral behavior, there is still a lack of systematic research on different persuasive mechanisms in advergames. Therefore our paper reports an initial study for assessing the rhetoric persuasive potential of advergames. Procedural Rhetoric of Advergames “Instead of telling the player what the point is, the game lets him or her experience it.” (Bogost 2007, 35). This indicates, in terms of using advergames for persuasion and learning, advergames offer possibilities to influence beliefs about a brand. Narratives, simulation, the rule-system as well as audiovisual and textual cues can be combined in an advergame


48

(Ritterfeld and Weber, 2006; Bogost, 2007; Theodorou and Sirmakessis, 2009). Games, and especially advergames, are usually played voluntarily, free of charge and without pressure. All of these aspects indicate that advergames probably reduce counter-arguing and lead to more active processing compared to TV-commercials (Waiguny and Terlutter, 2011). Like games in general, advergames usually follow a rule system and have a defined goal for success in the game. They can make players learn through these defined rules (Caillois, 2001; Huizinga, 1998). In addition, because games usually reward players with a high level of entertainment, they exert a strong motivation to repeat the game, and repeated game play increases learning effects (Caillois, 2001). Finally, the player is able to gain new skills, learn new behaviors and gain new identities (Sutton-Smith, 2001). Basically, all these mechanisms apply to advergames, too (Thomson, 2010). The brand is the key aspect of the advergame, and given that players should learn about the brand and the brand feature, the integration of the brand into the advergame has to follow several rules. As suggested by Smith and Just (2009), it is necessary to first integrate the brand into the gameplay. Second, the message must be autonomous, i.e., the gameplay must be able to convey the message about the brand, or, if this is not possible (e.g. because the message about the brand is too complex) the advergame should encourage the gamer to visit other resources for information. Third, a fit between the game and the brand â&#x20AC;&#x201C; the congruence - should be present. Interaction with and demonstration of the brand has to be part of the tasks and rules of the game (Smith and Just, 2009) and should be rewarded in the game (Bang et al., 2006). Finally, the advergame has to be designed in a way that players spend enough time on it without losing interest and intend to replay (Khaled et al., 2007). FRAMEWORK FOR ANALYZING THE PERSUASIVE MECHANISMS IN ADVERGAMES (1) Brand Integration: Most important to the integration of the brand is that it is placed in areas where the player allocates his or her attention (Khaled et al., 2007; Lee et al., 2009). Hence, it is not only about position and frequency, but also about level of integration in terms of how to interact with the brand in the game (Theodorou and Sirmakessis, 2009). Regarding the level of integration, Chen and Ringel (Chen and Ringel, 2001) differentiate between associative, illustrative and demonstrative integration. Associative integration means that only the logo or product is present and it is assumed that positive associations between brand and the favorable gaming experience occur. Illustrative integration means that the player learns about the brand by cues included in the game play (e.g. showing pictures of the usage of the brand, or showing different flavours of the product at different levels). Demonstrative integration means that the player even experiences the advantages of the product. Illustrative and demonstrative integration are not only related to conditioning, as they are part of observational or experiential learning too. Thus we propose the following question for the content analysis of the advergames: RQ1a: How often and how prominently does the advertised brand or product appear in advergames? RQ1b: What levels of brand integration are applied? (2) Message Autonomy: Autonomy is understood as how independent of further information the brand message is conveyed to and processed by the gamer. The question is whether the narrative (i.e., the plot) of the game includes a claim and a premise (Smith and Just, 2009), whether the game tells a story. How does the brand contribute to the gameâ&#x20AC;&#x2122;s story or narrative (Theodorou and Sirmakessis, 2009)? Even simple games can include one core message or not: For instance, a version of Tetris with a brand logo along the sides does not tell a story. Whereas a game where you fight as


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

49

the company vs. your competitor does tell a story. E.g.; in PEPSI-Space Invaders the player acts as Coca Cola and has to defend the world against the invaders namely Pepsi (Bogost, 2007). Other examples are e.g.; the Rama memory game, which shows pictures of the brand, and happy, healthy children. Thus the message is autonomous, as delivered in the narratives of the game. On the other hand, the Twix game Pong (although the product is well integrated) does not have an autonomous message as neither advantages of the product nor attributes are integrated. It is not possible, as yet, to claim that the more autonomous a game-message relation is, the higher its persuasiveness. In fact, the converse is possible, if the game encourages gamers to buy products or find product information outside the game (Smith and Just, 2009). We analyze the extent to which advergames contain a clear and brand related message which could be processed without additional knowledge and if the game leads to information desire about the brand. Thus we investigate: RQ2: To what extent do advergames contain a clear message which could be processed without additional knowledge? (3) Overlap of Advergame and Intended Brand Messages: Congruence or thematic connection between the featured brand and the advergame is desirable (Nelson and Waiguny, 2012). Theoretical and empirical evidence implies that incongruence increases brand recall, but that, in the case of incongruence, the receiver gets more suspicious about the placement (Balasubramanian et al., 2006; Bhatnagar et al., 2004; Van Reijmersdal, 2009); thus incongruence might negatively influence brand attitudes. In terms of gaming, congruity is necessary for the genre-advertising match, e.g., integrating Coke into a fantasy game might cause suspicion. This was observed for a range of different game genres (fantasy, WoW, racing, ego-shooter games ect.) (Yaping et al., 2010). Studies on advergames reveal also the positive influence of congruence on the attitude towards the advergame (Gross, 2010) as well as towards the brand (Wise et al. 2008). One inherent feature of a game is the possibility to integrate the message in the game’s rule system (Caillois, 2001). A persuasive effect is obtained if there is a high overlap between the goal and the story of the game and the goal of the message and the promoted brand (Bogost, 2007). A good example for this rule-message overlap is the energy saving game called “Power House.” The player is part of a daily soap-like story, and winning conditions are set so that you have to fulfill several energy saving tasks in the house. The more energy you save, the more points you earn (Bang et al. 2006). Such mechanisms for advergames can be observed: e.g., an anti-advergame against the negative business practices of McDonalds that the player has to perform during the play of the game (e.g., destroying woods, lobbying, growing soy, etc.). Thus, the game as a whole conveys the message that McDonalds’ business is doing destructive things in the world (Ahn, 2008). A more positive example is the game “Groovy Glider” by Pringles that conveys the different flavors of the brand by making the player collect the ingredients in order to gain points. Rewarding interaction is also important, through the use of operant conditioning mechanisms which engage the user to repeat the intended interaction with the brand. Thus we can assume that interaction (collecting the product, using the product, etc.) combined with rewards positively influences attitude (Bang et al. 2006). This is based on conceptual and perceptual fluency effects out of the interaction with the branded product or the brand logo within the game and the tasks and the resulting procedural learning effects through priming, conditioning (Hang and Auty, 2008; Nelson and Waiguny, 2012) and operant conditioning (Bang et al. 2006). For example, if a player collects a chocolate bar then gets faster, this fosters stored associations in the brain that the chocolate bar is good for your energy balance. If repeated often, as specified the game design, this message is processed more quickly, and finally positively influences attitudes. Furthermore as the


50

behavior is rewarded the player also generates new associates out of experiences with the brand (e.g. that the the cookie is crunchy when bitten into). This leads to following research questions: RQ3a: To what extent is congruence between game and brand/product, as well as fit between the goal of the game and the goals of the message, found in advergames? RQ3b: Are the tasks in the advergames related to interaction with branded elements. Further, are these interactions rewarded by the rules of the game? (4) Engaging Gamers to Repeat Play: Advergames in most cases are designed as casual games. Thus they are rewarding and deliver quick fun (Kuittinen et al., 2007). However for persuasion purposes the longer the exposure time the more persuasive mechanisms (such as conditioning and learning effects) are likely to work (Waiguny and Terlutter, 2011). Thus advergames should be designed so that the gamer wishes to repeat play. Having several difficulty levels or different environments are examples of gaming features that encourage repeat play (Moore and Rideout, 2007; Santos, Gonzalo and Gisbert, 2007; Lazzaro, 2008). The most persuasive effects therefore may occur in open ended games that are not boring, which is typical for casual games (Lazzaro, 2008). Thus we also investigate in the following research question if the game has a definite end, and how time consuming finishing the game is, by recording the time to finish it. For all games we also code how likely it is that players will continue playing, and to assess if the game might get boring quickly. RQ4: How time consuming and engagaging are advergames? How many advergames have a definite game end? Overall Persuasiveness of Advergames: The greater the number of the above-described strategies and elements are present in an advergame, the more persuasive the game is likely to be. Therefore we propose the last research question to examine how many different aspects of persuasion are used in current advergames by calculating an overall index. Furthermore we investigate whether different combinations of features exist in advergames, and which forms of advergames utilize different approaches to promoting a brand. Therefore we use the observed data to conduct a clustering of the games. RQ5: To what extent do advergames utilize persuasive mechanisms? Do there exist different forms of advergames in terms of their levels of persuasiveness? STUDY Method and Procedure To answer the research questions, content analysis of a sample of advergames is undertaken. To gather a comprehensive sample, we visited international and German advergaming web sites like (www.advergames. com, www.adventuregamesite.de, de.gamepost.com/ tags/Advergames) where we drew a random sample of 59 advergames in total. The small sample size was chosen as we wished code all levels of the game; this required repeat plays of each game. Table 1 presents a profile of the advergames analyzed. Two researchers independently played all of the 59 advergames. Both researchers were told in the first observation round just to play and try to finish the game as in real life. Afterwards they coded the time spent in the first round and if the game has an end and how appealing the experience was. In subsequent rounds they coded brand appearances, integration and the fit between message and game. An average overall inter-coder reliability of 86.1% was achieved (all variables had inter-coder


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

51

reliabilities of at least 80% with the exception of intention to continue playing with 64%, and autonomous message with 74%). In cases of disagreement, the team of researchers and coders revisited the games, discussed the coding, and thereafter reconciled the codes. Results RQ1 addresses the proximity and the forms of brand and/or product appearance in games. To answer RQ1a, we evaluated the positions where either the brand logo or the product was visible. As Table 2 shows, several different positions are used, and often more than one single placement position is chosen. However, in about a quarter of the games (25.4%) no brand is included, and for 35.6% no product integration is found at all. 16.9% of the games is totally unbranded. For the remaining 49 advergames, we calculated a placement index (ranging from 1 to 14 as we observed 14 different positions). One game uses 12 out of 14 positions. 5.1% use 9-11 different positions. 25.5% use between 4-8 different positions. Most of the advergames use either three (16.9%), two (20.3%) or only one (13.6%) brand/product position. RQ1b addresses the level of integration. While it is assumed that basic positioning of a brand/product in the game represents minimum associative, more advanced forms of integration, such as showing benefits of using the product or providing information about concrete product attributes, is still not common in most advergames. 27.1% of the investigated games present the benefits of the product within the game. For only 15.3%, concrete product attributes as part of the active and interactive elements of the game. RQ2 is related to the extent of message autonomy. About half of the games (49.2%) include a message in the game, which can be processed without any further information. Prior knowledge about the brand is only needed in one game (1.7%). Assuming that companies might engage customers to visit additional sources (e.g. buying the product, getting codes from packages or other websites, or visiting additional resources during or after finishing the game), we searched for these possibilities. Our data shows in none of the games are such additional resources addressed. However, for 8.5% of the games the story engages the player to further investigate the brand in more detail, as additional information on the brand is displayed at the end of the game. RQ3: The third persuasive element is the â&#x20AC;&#x153;fitâ&#x20AC;? of the game with the brand and intended message. As summarized in Table 3, a majority of the games could be used to promote any product. To collect the different advertising messages, we visited the websites of the suppliers of the advergames and collected their basic messages from other advertising campaigns to match them with the messages conveyed by the game. In a second step, we analyzed the advergames as to whether there was a defined goal and tasks related to this goal. In sum, for 81.4% of the games, a clear gaminggoal was defined; however only 49.2% of the advergames are designed to convey a message with its rules system or storyline, and only 28.8% of the gamesâ&#x20AC;&#x2122; messages overlap with general communication messages of the company. As Table 3 also shows, the integration of the brand-message in the tasks is also quite low. Collecting the product is the most common form of integration (22.0%). More advanced interaction possibilities and rewards are found in only 8.5% of the games. Gaining real-product-related extra powers, which might be used to convey a message (e.g., if you drink an energy drink, you gain power), is only found in one game. RQ4 addresses how time-consuming advergames are. On average the two coders spent 3 minutes and 20 seconds to finish a game, which is usual for casual games (min 0:23, max 11:47). 79.7% of the games had a


52

definite end however our coders indicated for 38.8% of games that the gameplay engages the player to continue. Eight games are open ended but only three are so easy that a player might get bored. Four games are too hard to play to completion. For seven (11.9%) of the cases, the player is rewarded for succeeding after finishing the game. Most common are the option to take part in an additional lottery game (three times), downloads like wallpapers, mobile ringtones, screensavers (in two cases), and in one game, a gift card worth about 20 USD was issued. Five of these seven games announce the rewards prior to the start of the game. Different difficulty levels or changes in the environment or story are observable for 52.5% of the games. In RQ5 we examine the overall persuasive power of advergames. Therefore we first calculate an index based on the four dimensions of our framework. For each element we allocate points as explained in the appendix. In sum, an advergame could be rated between 0 (low) and 16 (high). Figure 1 depicts the distribution of persuasive power in the advergames analyzed, indicating that a majority of the advergames are rated between 6 and 10; only a minority of advergames are more advanced in terms of persuasion, receiving ratings between 10-14. We conclude that advergames are still viewed by the marketer as merely an additional fun feature on websites rather than a persuasive tool. However to gain a deeper insight we investigated if there are different combinations of persuasive elements. A two step cluster analysis with an acceptable cohesion and separation measure (0.27) identifies a four-cluster solution. The summary of the cluster analysis is depicted in Figure 2. 35.6% of the advergames are “Persuasion Pioneers” which are characterized by high levels of integration, and fit between the brand message and the gameplay. 30.5% of the advergames are “Fun Transferers” that demonstrate high integration levels but lack strong interaction with the brand, thus a simple affect transfer onto the brand is assumed. 15.3% of the games have a “Recognition Focus” with high scores for brand appearance and interaction, but a weak match of the game story with the message. Finally 18.6% are “Me Too” advergames. These are weak or even non branded games with no special persuasion purposes. DISCUSSION AND IMPLICATIONS Our study contributes to existing literature in two ways. First we show a comprehensive analytical framework for analyzing advergames, based on prior empirical findings as well as theoretical insights in game design. Second, based on this framework, we analyzed 59 advergames. Our results show that, in terms of persuasive power, initial attempts to apply persuasion principles are made. In particular our cluster analysis shows about 35% of the games are early adoptors of persuasion principles but there is still some gap on the use of interaction and especially more advanced interaction to convey messages. However, the majority of the advergames either focus on transferring the games fun to the brand, or just achieving brand recognition. Both clusters use many placement positions, and to some extent also interaction, but not for telling a message. The Me Too advergames simply attempt to “provide games on the website”. To sum up, according to adoption theory, we are still in an early phase; there are attempts to build persuasive advergames, however not all persuasive features are considered by marketers and advergames’ designers. Limitations of this study are the small sample size which limits the generalizability of the results. Thus, future research might consider coding a larger number of games to analyse whether there are differences among industries, target groups or regions. Furthermore, this is an exploratory study, which is aimed to give an insight, and propose a framework to analyze advergames. To investigate if these features are successful requries additional research to measuring attitude changes among consumers.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

53


54

FÖRORD: När produktens utveckling går fortare än marknadsföringens Hur tar man tillvara på en publik som har ett brinnande intresse för film? Behöver man som annonsör vara försiktigare i en miljö där högintresseprodukter finns? Idag är intresset för film större än någonsin och dit konsumenterna går, följer också annonsörerna. Film har för en stor del av populationen blivit den nya boken. Under det senaste decenniet så har filmens utveckling varit dramatisk med hjälp av ny teknik och våra bioupplevelser blir än mer fokuserade på att vara just en upplevelse. Antalet filmer som släpps från filmbolagen har också ökat lavinartat och målgrupperna för filmerna blir mindre och mer nischade. Varje år ser vi nya rekordsiffror för filmbudgetar och våra samtalsämnen i fikarum kretsar ofta kring nya skådisar, regissörer och äventyrliga effekter i just biofilmer. Det innebär ju också att våra förväntningar på de filmer vi väljer att se på bio ökar i samma takt. Men marknadsföring i och omkring film har inte utvecklats i konsumentens ögon utan den har bara ökat i kvantitet. Även om reklamen kommer i olika former som reklamfilm eller produktplaceringar så uppfattas det ändå som reklam. Och för konsumenten skapar det en känsla av en aldrig sinande flod av budskap vilket denna rapport bevisar att det ger en sämre bioupplevelse för besökarna. Och en sämre upplevelse vill ju ingen ge besökarna, men då måste marknadsföringen utvecklas. Här finns verkligen ett ”first mover advantage” att hämta för de annonsörer som vill och vågar!

ANNA hjalmarsson


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

55

Contaminated hedonic experiences: Advertising exposure can negatively impact movie experiences Cristel Antonia Russell American University US Dale W. Russell Uniformed Services University US Jean-marc Lehu Université Paris 1 Panthéon-Sorbonne France

INTRODUCTION Consumers throughout the world are exposed to an unending array of advertisements; in the United States (US) alone, the average American urbanite is exposed to 5,000 marketing communications daily (Story, 2007). While the advent of new technologies, consumers are increasingly able to skip through advertising (e.g., DVRs) or to elect not to be exposed to advertising messages (e.g., spam filters), there are still ‘captive audience’ environments in which exposure to advertising cannot be avoided (e.g., in movie theaters and sponsored sports events; Speck and Elliott, 1997; Truong and Simmons, 2010; Wilbur, 2008). This research focuses on the impact that exposure to advertising in a movie theater has on consumers’ processing and opinions of product placements in the featured movie. The past decade has bestowed upon consumers the ability to easily view movies when and where they desire (e.g., on-demand via mobile phones). As such, cinemas have been forced to strategize to both keep moviegoers frequenting theaters and also finding new means of generating income. One method of generating income has come in the form of screening made for television commercials prior to the commencement of the feature movie. This advertising approach has garnered little attention from academic researchers (Overpeck, 2010). Screening advertisements before movie previews and the feature movie is not a new concept, as European cinemas began screening advertisements in 1894. Although the practice did not become an established business strategy until the mid-1970s in the US (Austin, 1986; Overpeck, 2010) pre-movie advertising has now moved to the national theatric screening of television-like commercials for major brands. For example, the National Cinema Media Company alone has brought such cinema advertising, including 3D spots, to over 17,000 movie screens throughout the US (NCM, 2011). According to the National Association of Theatre Owners, pre-movie advertising has become a key source of revenue for cinemas, having increased from approximately $185 million in 2002 to $658 million in 2010, with projections to continue exponential growth (Cinema Advertising Council, 2011). Advertisers continually strive to garner a sliver of consumers’ ever divided attention to convey their messages; however, it is a common occurrence for consumers to voice irritation with such advertising efforts for interfering with their hedonic pursuits (Edwards et al., 2002). Movie industry analysts have voiced some concern that pre-movie advertisements have contributed to a decline in cinema attendance because they interrupt the very hedonic experience moviegoers seek (Overpeck, 2010). Renowned movie critic Roger Ebert (2003, p. 828) echoed such a sentiment in response to pre-movie advertisements:


56

“I… resent being made into a captive victim. I… cannot understand why advertisers would want to attract hostility toward their products by deliberately offending potential customers.” There is indeed ample evidence that an advertisement irritation threshold exists, whereupon being breached consumers become irritated and respond negatively to marketing communications efforts (Danaher and Rossiter, 2009). Such irritation can manifest into a negative psychological state that interferes with one’s cognitive processes; additionally, such negativity can extend beyond the offending message and be conveyed to as associated object (e.g., the product being advertised) or even the entire experience, thereby psychologically contaminating it (Aaker and Bruzzone, 1985; Edwards et al., 2002; Morales and Fitzsimons, 2007). HYPOTHESES It is proposed that exposure to advertising messages reduces the enjoyment of a theater experience and reduces attitudes toward the movie itself. H1: Exposure to commercials before a movie reduces viewers’ enjoyment of the overall experience (H1a) and of the movie itself (H1b). Furthermore, exposure to advertising before the movie should decrease viewers’ receptiveness to product placement, as captured in the following hypothesis. H2: Exposure to commercials before a movie that contains product placements reduces viewers’ attitudes toward the practice of product placement. Moreover, we propose that it is these lower attitudes toward product placements within the movie that lead to the lower enjoyment levels. H3: Attitudes toward product placements within the movie mediate the effect of exposure to pre-movie commercials on enjoyment of the overall theater experience (H3a) and enjoyment of the movie specifically (H3b). However, because it creates interference with the processing of product placements in the movie, exposure to pre-movie advertising should be detrimental to the actual recall of placements, as captured by H4. H4: Exposure to commercials before a movie that contains product placements reduces the recall of product placements in the movie. METHODOLOGY This research was conducted in collaboration with an independently owned single-screen movie theater. The theater accommodates an audience of 150. A quasi-experiment was conducted with the theater operator agreeing to create the following viewing conditions: 1) screen only the movie; 2) screen the movie previews and the movie; and 3) screen both the commercials and the movie previews before the movie. The data were collected over a three-day period (Friday, Saturday, and Sunday) on two subsequent weekends. These high audience days were selected to maximize audience attendance. The movies shown were two newly released films: Red and Due Date, each contained multiple product placements. Upon exiting the theater, moviegoers were invited to complete a two-page paper survey regarding their theater experience. Participants were given a survey specific to their experimental condition. Measures The two-page survey began by asking how many minutes before the movie started they were seated in the theater. The next section focused on measures of overall experience enjoyment and movie enjoyment (see Appendix). Then, the same set of six items were collected regarding their opinions of commercial advertisements before the movie, movie previews before the movie and product placements in the movie (defined as ‘real products that were shown in the movie


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

57

itself”). Every survey included the product placement measure, since everyone was exposed to a movie, but the sets of measures regarding the commercial advertisements and movie previews were only included in the conditions that included those. On the back page, the first three questions focused on recall and participants were asked to “list all the movie trailers (if applicable) / commercial advertisements (if applicable) / product placements” they recalled seeing or hearing. These were checked for accuracy and total recall scores were computed for each category. Finally, demographic information was collected a well as a measure of average movie consumption: participants indicated approximately how many movies a month they watch in a theater, on television, on a computer, and on a mobile device. Sample A total of 563 surveys were collected and distributed per the following conditions: movie alone (Due Date 79 / Red 89); previews and movie (Due Date 88 / Red 93); commercials, previews, and movie (Due Date 108 / Red 106). The average age of participants was 22.4 (σ = 5.54) and included 378 males and 185 females. FINDINGS The hypotheses were tested according to a three cell (condition: movie only vs. previews + movie versus commercials + previews + movie) betweensubjects MANCOVA, with age, gender, and average monthly movie consumption as covariates. We also controlled for movie-specific effects by including the particular movie watched (Red versus Due Date) as a covariate in all the analyses. In support of H1, the analysis reveals a significant main effect of condition on enjoyment levels (F (4, 1102) = 29.37, p < .05). The movie covariate was significant (F (2, 551) = 160.05, p < .05), with participants indicating higher enjoyment levels of the experience and the movie for Due Date than Red, which is consistent with critics’ reviews of the quality difference between the two movies. None of the other covariates were significant. Subsequent univariate analyses confirm that this main effect occurs for both overall enjoyment (F (2, 551 = 40.26, p < .05) and movie enjoyment (F (2, 551) = 33.29, p < .05; see Table 1 for means). Planned contrasts reveal that participants in the commercials + previews + movie condition enjoyed the overall experience and the actual movie less than both of the other two conditions. Attitudes toward product placements were also analyzed based on the same between-subjects ANCOVA. Again, the experimental condition is significant (F (2, 551) = 119.90, p < .05) as well as the movie covariate (F (1, 551) = 214.30, p < .05). Exposure to commercials significantly reduces consumers’ attitudes toward product placements compared to the two conditions where consumers did not see pre-movie commercials, in line with H2. To test the proposed mediation process (H3), two dummy variables were created to capture the three experimental conditions: exposure to previews (0/1) and exposure to commercials (0/1). These variables were used in a series of hierarchical regressions to test the x -> y -> z proposed relationships (x = experimental condition, y = attitude toward product placements, z = enjoyment levels) per the standard Baron and Kenny (1986) procedures, but with the emphasis on the x -> y and y -> z paths that Zhao et al. (2010) emphasize. In the first step, attitude toward product placements (the proposed mediator) is regressed on the condition dummies (i.e., we estimated the x -> y path). Next, the dependent measures for enjoyment levels are regressed on the condition dummies (now estimating the x -> z


58

paths) in a first step. Then, in a second step, the mediator, attitude toward product placements, is entered into the regression and its incremental contribution to R2 as well as the beta coefficients of the condition dummies are examined for significance (y -> z and x* -> z paths). The regression results appear in Table 2. Exposure to commercials has a significant and negative effect on attitude toward product placements. Full mediation is supported for the effect of ad exposure on overall enjoyment as well as movie enjoyment, in support of H3a and H3b: adding attitude toward product placements to the regressions provides a significant increment in R2 and the beta for experimental condition ceases to be significant. Finally, these results appear to operate independently from actual recall of product placements in the movie. Indeed, an ANCOVA comparing placements recalled by experimental condition uncovers a main effect of condition (F (2, 551) = 27.72, p < .05): in line with H4, recall of product placement is lower after exposure to commercials. The movie covariate is also significant for, logically, recall is lower for the movie with fewer placements (Red). CONCLUSION Limited research exists regarding the factors that influence contemporary moviegoersâ&#x20AC;&#x2122; hedonic experiences (Ladhari, 2007). This research provides empirical evidence that advertising messages can interfere with and lower enjoyment levels of hedonic pursuits. While past work has tried to identify specific factors that make one advertisement more or less irritating than another (Aaker and Bruzzone, 1985), the current research shows that simply the presence or absence of advertisements during a hedonic experience is enough to impact evaluations of the experience. Indeed, this research supports the concerns of movie industry analysts who fear pre-movie commercials have contributed to the decline in theater attendance; moviegoers do in fact enjoy movies less when they are preceded by commercials. Because advertising exposure in this context was not volitional, future work might also examine how feelings of reduced control may impact consumer reactions to pre-movie commercials or other cases where consumers are unable to escape from or block out advertising messages. One of the most significant implications of the current findings is that, if in-theater commercials lower consumersâ&#x20AC;&#x2122; enjoyment of the movie-viewing experience, perhaps theaters should reconsider their practice of screening them. According to the National Association of Theatre Owners, pre-movie advertising has become a key source of revenue for cinemas, having increased from approximately $185 million in 2002 to $658 million in 2010, with projections to continue exponential growth (Cinema Advertising Council, 2011). However, by focusing on in-theater commercials as a primary source of revenue, it may be that cinemas are trading off long-term profits for short-term ones. If consumers enjoy their movie experiences less with pre-movie commercials, across time they may shift their movie-viewing experiences from the movie theater to in-home or online, where they can fast-forward through the previews and commercials for a more enjoyable experience. Assuming that consumers gravitate to experiences that provide them the most utility, theaters and other firms that provide hedonic experiences need to examine how advertising contamination may lower consumer enjoyment, and consider other potential sourc sources of revenue that do not detract from the hedonic experiences they are trying to provide.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

59


60

FÖRORD: Högre frekvensoptimering ställer nya krav på slitstark reklam Många annonsörer letar fortfarande efter den heliga graalen för hur många exponeringar det behövs för att charma en konsument. TV-reklam är huvudmedia för många, stora globala annonsörer och har genom historien visat sig vara väldig effektiv för att påverka försäljning och varumärke. Men hur många gånger som ett budskap ska repeteras i TV är fortfarande idag ofta baserat på den traditionella uppfattningen att 3 exponeringar gör jobbet. Men är det verkligen så fortfarande? Balansen mellan tjat och påminnelse är något som många annonsörer gärna skulle vilja veta mer om. Reklamfilmer som får rulla för tungt och för länge i TV väcker till slut anstöt hos konsumenterna och skapar en negativ uppfattning om varumärket istället för en positiv känsla. Ett ökat reklambrus har skapat ett filter som blir allt svårare att tränga igenom och i framtiden så kommer det finnas högre krav på bättre reklam och ett mer sofistikerat sätt att exponera den i olika mediekanaler. När antalet TV kanaler ökar och vi flyttar vårt TV tittande till digitala plattformar så påverkar det antalet exponeringar som tv-köparna bör optimera sina kampanjer efter. I den här rapporten hittar du rykande färsk forskning om hur annonsörerna bör optimera sina TV kampanjer i framtiden vilket också påverkar både annonsören och byråerna som gör reklamen. Det är tydligt att branschen behöver utveckla slitstark reklam som håller för så många exponeringar 6 gånger utan att uppfattas som tjatig.

ANNA hjalmarsson


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

61

ADVERTISING REPETITION – ARE THREE EXPOSURES ENOUGH? susanne Schmidt Technische Universität Dortmund Germany Martin Eisend European University Viadrina Germany

INTRODUCTION For decades, advertising researchers and practitioners have been searching for the precise number of exposures that would maximize consumers’ responses. The majority of advertisers recognize the third exposure as the magical number which tells managers how much to spend on advertising. Many practitioners still follow Krugman’s (1972) threeexposure hypothesis and rely on three exposures to be the optimum level of repetition (Shimp, 2008). The optimum level is also subject to extensive discussion amongst scholars (Tellis, 1997; Vuokko 1997). This paper analyzes the optimum repetition level by referring to meta-analytic data of prior scholarly work as well as to a large field study dataset. Although a generalized solution suffers from high variability in the individual studies and campaigns, this paper provides strong evidence that an optimum level is not reached before the sixth exposure. THEORETICAL BACKGROUND Research about advertising repetition effects is mainly based on the underlying mechanisms of the two-factor theory proposed by Berlyne (1970) and extended by Stang (1975) as well as Cacioppo and Petty (1979). The two-factor theory states that two opposing factors influence the effect of repetition on attitude. The two theories are positive habituation and negative redundancy. Within the first exposures to an unknown stimulus the respondent becomes famil- iar with that stimulus and positive thoughts develop. This familiarity evokes a good feeling, that again, positively influences attitude formation (Berlyne, 1970; Cacioppo and Petty, 1979). The familiarity effect is comparable to the one accounting for the mereexposure effect found by Zajonc (1968). In addition to the mere-exposure effect, the two-factor theory suggests that a second factor, boredom, develops with increasing repetition, leading to negative thoughts. This factor opposes the positive effect of familiarity and, thus, positive thoughts. Therefore, the two contrary developments lead to a positive repetition effect, but at a decreas- ing rate, reaching a maximum attitude at an exposure level at which familiarity and, thus, positive thoughts are at their highest. Past this exposure level, the repetition effect becomes nega- tive due to boredom and increasing negative thoughts (Berlyne, 1970; Cacioppo and Petty, 1979). The net effect of these two opposing factors causes an inverted u-curve: Within the first exposures, attitude becomes more positive until boredom outweighs familiarity, so that attitude becomes more negative with every additional exposure. Since many studies supported the inverted u-curve as proposed by the two-factor theory (e.g., Stang, 1975), it became the prominent course of effect for advertising repetition. However, the exposure rate at which attitudes are at their optimum level is subject to continuous scientific debate (Tellis, 2009). Krugman (1972) suggested that the optimum repetition level for the effectiveness of advertising stimuli is reached within three exposures. According to Krugman, every single one of these exposures has an important function.


62

The first exposure is needed to understand the stimulus. At the second exposure the stimulus is processed and evaluated. Only the third exposure is actually considered as repetition. At the third exposure the respondent does not process the stimulus anymore, but remembers it from the previous two exposures. The assump- tion of three exposures as being the optimum exposure rate is supported by several studies, some of which reach considerable consent amongst scholars (e.g., Schumann et al., 1990). The literature review conducted by Pechmann and Stewart (1988) supports this assumption, too, and suggests that attitudes peak after the third exposure. Practitioners adopted the empirical findings and many media agencies today use the three-exposure approach as a rule of thumb for media planning (Shimp, 2008). However, within the last decade an increasing number of empirical studies showed that advertising repetition can increase attitude even after the third exposure level: McQuarrie and Mick (2009) find that positive attitude can increase until the sixth exposure and Nordhielm (2002) find an increased positive attitude even at the 25th exposures. Such evidence gives reason to reinvestigate the optimum exposure level. METHOD We examine the optimum level of advertising repetition with two large datasets: meta-analytic data from previous scholarly studies and realworld data from a field study. This approach allows for validation of both approaches typically applied in advertising repetition research. First Study: Meta-Analytic Data Data and sample The meta-analytic sample comprises prior empirical research studies published in academic journals and investigate the effect of advertising repetition on attitudes. In order to find all relevant studies, we conduct a systematic literature search in three steps. First, we search all relevant databases with keywords, such as ad repetition, message repetition, ad exposure, ad frequency, or wearin and wearout. Second, we examine the reference list of each article that we found within the keyword search in order to find additional articles relevant for the topic. Finally, we conduct an issue by issue search of international research journals that are most likely to publish articles on advertising repetition, such as International Journal of Advertising, Journal of Advertising, Journal of Advertising Research, Journal of Current Issues & Research in Advertising or Journal of Interactive Advertising. The literature search provides a dataset of 50 articles on repetition effects of advertising stimuli on attitude, with 14 articles investigating the repetition effect on attitude towards the ad. The average year of publication within the dataset for attitude towards the ad is 1994. Measures Number of exposures is the independent variable. It is a continuous variable, indicating the number of times the respondents were exposed to the advertising stimulus before their attitude was measured. Attitude towards the ad is the dependent variable. The majority of studies investigate advertising repetition effect on this particular attitude variable. Procedure We code standardized measures for attitude towards the ad on each reported exposure level. Relating to the underlying mechanisms of the two-factor theory, we test an inverted u-shaped course against a linear course of effect by regression analysis. We run two models. The first model contains the variable number of exposures only. The second model extends the first one by adding a quadratic term of the variable number of exposures for displaying a non-linear effect. The change of R2 and the change of the F-value as well as the significance level of the change in F


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

63

provide information on a possible non-linear effect. Finally, we graph the course of effect over all means for attitude towards the ad in order to identify the optimum level of advertising exposure. Second Study: Real-World Data Data and sample This study is a secondary analysis that is based on the dataset provided by SevenOne Media GmbH, a major German media research agency. SevenOne Media regularly reports insights into advertising effects in their annual AdTrend reports. An AdTrend report of 2010 provides the dataset for the study at hand. The survey period for the present dataset was 1997 to 2009. About 300 German speaking citizens between 14 and 64 years of age were randomly contact- ed and surveyed by telephone on their liking of a certain ad campaign. SevenOne Media then combined the surveyed data with TV-usage data of the German research institute GfK SE Fernsehforschung and with advertising switching data of the Nielsen Media Research GmbH. SevenOne Media used this combined information to calculate the corresponding number of advertising exposures. Out of more than 150 different product categories, this dataset contains attitude measures for 60 different brands. A total of 268 campaigns are included in our da- taset. Measures Number of exposures is the independent variable. It indicates the number of exposures that the respondents had within four weeks before the interview. This variable is coded according to the one in the first study. Attitude towards the ad (campaign) is the dependent variable. Procedure We test the inverted u-curve for advertising repetition against a linear course of effect by using regression analysis as described above. We graph the corresponding course of effect and compare the results to the ones obtained by the first study. RESULTS First Study: Meta-Analytic Data The results of the regression analysis are shown in table 1. R2 changes from .001 in model 1 to .052 in model 2. Also, the change in F-value increases from .106 in model 1 to 4.753 in model 2. This change is supported by the significance level of both models. The second model, which includes the quadratic term, is significant (p < .04), while the first model with only the linear effect is not (p > .74). Thus, a non-linear course of effect for advertising repetition can be assumed. Figure 1 displays the course of the advertising repetition effect. The repetition effect on attitude takes the predicted inverted u-shaped course. The maximum effect of advertising repetition is reached at approximately six exposures. The inverted u-curve is constant with the course of effect proposed by the two-factor theory. However, six exposures as the level where maximum attitude is obtained exceeds the strongly hold assumption of three exposures to be the optimum level. Second Study: Real-World Data The results of the regression analysis are displayed in table 2. R2 changes from .001 in model 1 to .049 in model 2. Also, the change in F-value increases from .001 in model 1 to 4.530 in model 2. This change is supported by the significance level of both models. Only the second model, that includes the quadratic term, is significant (p < .04), whereas model 1 with only the linear term does not show significance (p > .99).


64

Thus, a non-linear course of effect for advertising repetition can be assumed. Figure 2 displays the course of effect for advertising repetition. It shows the inverted u-shaped course of effect which is in line with the course of effect identified by the first study. Maximum attitude is reached after the sixth exposure. The analysis of real-world data validates the inverted u-shaped course of effect found in study 1. The results from both studies show that the optimum exposure level is higher than the long hold assumed level of three exposures. DISCUSSION The two studies presented in this paper examined the repetition effects of advertising stimuli on attitude towards the ad. We focused on the course of repetition effects and the optimum number of exposures. We found an inverted u-shaped course of effect over all attitude means provided by the studies. The most positive attitude is reached at the sixth/seventh exposure as suggested by meta-analytic data of scholarly researchers and the real world-data. The findings reveal new insights into the number of exposures that should be applied in order to reach the consumer in the most optimal way. The analyses show that the most positive attitude is not reached before the sixth exposure. Therefore, three exposures as proposed by Krugman (1972) and supported by the review of Pechmann and Stewart (1988), and continu- ously applied by practitioners (Shimp, 2008) are not enough to reach consumers in the most effective way. The major limitation of the present investigation is the highly generalized findings that neglect the underlying heterogeneity of single study and campaign results: the optimum level of exposure varies between these studies and between campaigns. A next step in our investigation is to include moderating variables in the analysis. For instance, Tellis (1997) suggests that familiarity, message complexity, and novelty are important moderators of frequency effects. Pechmann and Stewart (1988) suggest factors, such as presentation mode or advertising length as well as internal factors, such as a consumerâ&#x20AC;&#x2122;s involvement towards the advertised stimulus prior to exposure. Overall, the present findings provide a contribution to the debate about the magical number of three exposures that seems to be a myth and not the truth nowadays.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

65


66

FÖRORD: Kan man sälja utan att bli en sellout? ”Sellout.” I en tid där personliga varumärken blir allt viktigare, och kända ansikten används för att sälja allt fler produkter, så kan ett enkelt ord vara brännmärkande. Att man är falsk. Köpt. Icke-autentisk. Rädslan för den brännmärkningen delas av både varumärkesägare och kändisar. Varumärken är rädda för att värdet av deras partnerskap med kändisar ska undergrävas om det upplevs som ”fejk”. Och kändisar är rädda för att omintetgöra sitt eget varumärke, och att undergräva det förtroende människor i allmänhet (och deras fans i synnerhet) känner för dem. Så frågan är: hur kan en kändis sälja en produkt utan att sälja sin själ och sitt goda rykte på kuppen? I den följande artikeln av Jasmina Ilicic och Cynthia M. Webster tas ett intressant första steg mot att börja reda ut denna kniviga fråga, tack vare dessa lärdomar: 1) Att om en kändis inte matchar varumärket, blir attityden till både reklam och varumärke betydligt sämre. Grunden för ett bra samarbete är ren kemi. Det är först om kändisen och varumärket upplevs matcha varandra, som varumärket vinner på det. (Och rent spekulativt, borde en sådan god matchning också innebära en mindre risk för kändisens personliga varumärke.) 2) Att en kändis kan ha positiv effekt på attityden till en reklam utan att i sig erbjuda relevant information om varumärket. Det vill säga, en kändis kan skapa värde genom att vara sig själv, utan att behöva direkt sälja eller informera om produkten. Deras medverkan är bidrag nog, så länge relevant information om varumärket redovisas på något annat sätt i reklamen. I vanlig ordning är dessa lärdomar inte definitiva eller enkelt applicerbara på en komplex verklighet. ”More research is needed”, som det heter. Men erfarenheten och magkänslan säger att artikelförfattarna är något viktigt på spåret. Att kändis-partnerskap fungerar bäst när personen får vara sig själv, och inte styrs till att övertydligt sälja eller rekommendera produkten. Att en kändis ska inte försöka vara en säljare, utan istället göra det de är bäst på: att vara sig själva, att vara det de blev kända för. Så rådet till kändisar och till de varumärken som vill samarbeta med dem är enkelt: Don’t be a sellout. Be yourself.

karl wikström


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

67

THE CHANGING ROLE OF CELEBRITIES IN ADVERTISING: CELEBRITY ENDORSERS AS IRRELEVANT INFORMATION IN ADVERTISEMENTS Jasmina Ilicic The University of Adelaide Australia Cynthia M. Webster Macquarie University Australia

INTRODUCTION Using celebrities as endorsers attracts consumer attention, helping brands stand out from the clutter and increasing brand awareness (Friedman and Friedman, 1979; O’Mahony and Meenaghan, 1997/1998). Celebrities, however, are not always employed as direct spokespeople, where they provide relevant information about an endorsed brand or demonstrate the product endorsed. Celebrities are at times strategically featured providing unrelated brand information by not mentioning or providing any information about the endorsed brand. We see a shift in advertising executions with more and more advertisements today highlighting the celebrities, emphasizing information on the celebrity and providing little or no relevant endorsed brand information. In recent advertisements for American Express (AMEX), the celebrities dominate. The one minute television commercial featuring Kate Winslet takes 55 seconds going through her starring film roles and leaves the last five seconds to AMEX. The print advertisement is similar with a glossy image of Kate Winslet occupying half of the space with much of the remaining devoted to information about her memories, ambitions and achievements. Only at the end of the advertisement, on one line does she state that the card she uses is AMEX. These advertisements include a clear endorsement for the brand, but also contain personal information about the celebrity, information relevant to the celebrity, but arguably irrelevant and not useful to consumers in making their judgments about the brand. Brand managers attempt to place a celebrity with a brand, whereby the image of the celebrity is seen to be consistent with that of the endorsed brand (Kamins, 1990: Kamins and Gupta, 1994; Till and Busler, 1998, 2000; Till, Stanley and Priluck, 2008) providing consumers with seemingly consistent information about the brand. For example, Katy Perry’s fun, youthful and eccentric image matches that of the Proactiv brand, providing consumers with consistent endorsed brand information. Research in celebrity endorsement suggests that consistent celebrity information, referred to as a match-up, results in a positive effect on consumer attitudes towards the products and brands with which they appear and the advertisements in which they feature, with a mismatch weakening or diluting consumer brand judgments (Kamins, 1990; Kamins and Gupta, 1994; Till and Busler, 1998, 2000; Till et al., 2008). LITERATURE REVIEW Match-up is said to occur when “highly relevant characteristics of the


68

spokesperson are consistent with highly relevant attributes of the brandâ&#x20AC;? (Misra and Beatty, 1990: 160). As such, the effectiveness of a spokesperson is determined by the perceived fit between the endorserâ&#x20AC;&#x2122;s brand attribute associations and the endorsed brandâ&#x20AC;&#x2122;s attribute associations (e.g. Kamins, 1990: Kamins and Gupta, 1994; Till and Busler, 1998; Till et al., 2008). Here the concept of match-up consists of two distinct concepts: congruence and relevance. Judgments of relevance depend on whether a stimulus communicates issue-pertinent information and judgments of congruence involve perceptions of suitability (Miniard, Bhatla, Lord, Dickson and Unnava, 1991). The literature to date on match-up focuses primarily on the celebrity and brand possessing characteristics that complement each other, reinforcing relevant brand information and the suitability of the endorsement. However, a celebrity may also convey relevant information when they mention the endorsed brand and information about the brand within an advertisement. In other words, a celebrity not only provides consumers with relevant brand information when they convey characteristics pertinent to the brand but also when they provide information relevant to the endorsed brand by describing brand benefits, demonstrating product functionality and explicitly endorsing the brand. For example, the open, friendly and fun characteristics of comedian, Jerry Seinfeld, are relevant to the honest, family friendly positioning of the financial institution Greater Building Society. Plus, Seinfeld may also be perceived as providing consumers with relevant endorsed brand information when he features in an advertisement for the Greater and he clearly endorses the brand by mentioning details about the brand. Alternatively, Seinfeld may be seen as incongruent and irrelevant when he provides little or no information about the Greater brand in advertisements, by not mentioning the Greater or any information about the brand. Despite extensive research on the effectiveness of celebrity endorsers, little research specifically examines the executional aspects of advertisements that feature celebrities. Most of the advertising research focuses on the effectiveness of celebrities as direct spokespeople, who provide relevant brand information through their consistent brand image and by clearly endorsing the brand. We maintain that celebrities provide irrelevant brand information through their mismatched (incongruent and irrelevant) image attributes and also when they feature in advertisements emphasizing their talents and providing no clear endorsement of the brand. Our aim here is to further the research on celebrity endorsements by examining the effects of irrelevant information, provided by a celebrity endorser, on consumer evaluations of endorsed brands. We examine the role of celebrity and endorsed brand match-up on consumer attitudes towards the advertisement as well as attitudes towards the brand. As such, the following hypotheses are put forward: H1: Consumers report a) the most positive attitude towards the advertisement when exposed to relevant information only, b) weaker ad attitude when exposed to relevant plus irrelevant information and c) the least positive ad attitude when exposed to irrelevant information only. H2: Consumers report a) the most positive attitude towards the brand when exposed to relevant information only, b) weaker brand attitude when exposed to relevant plus irrelevant information and c) the least positive brand attitude when exposed to irrelevant information only. H3: Consumers who perceive a mismatch between the brand and celebrity report less positive judgments relating to a) attitude towards the advertisement and b) brand attitude, than consumers who perceive a match.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

69

METHOD A research company was used to recruit subjects from regional areas within Australia to evaluate their attitude for the Greater Building Society, a financial institution. A total of 350 subjects were assigned to one of three experimental conditions: 1) irrelevant information only featuring the celebrity Jerry Seinfeld with no mention of the Greater Building Society brand nor any brand benefits, 2) relevant information only featuring a Greater customer, Jerry Seinfeld is not featured and 3) the combined relevant plus irrelevant information featuring a combination of the celebrity Jerry Seinfeld and the Greater brand customer. Of the 350 subjects recruited, 48.9% were male and 51.1% were female. Subjects were predominantly between 35-49 and 50-64 years of age (25.1% and 28.6%, respectively). One hundred and twenty-two subjects were assigned to the irrelevant condition, 114 to the relevant condition and 114 to the combined relevant plus irrelevant information condition. A median split was used to categorize subjects’ perceptions of match-up between the celebrity and brand. Measures Subjects evaluated their attitude towards the advertisement, attitude towards the brand and the degree to which they believe the image of Jerry Seinfeld matches that of the Greater Building Society’s image. Attitude towards the Advertisement was measured using Mitchell and Olson’s (1981) 5-point semantic differential scales which included four items bad/ good, dislike/like, not irritating/irritating and uninteresting/interesting (cronbach alpha= .945). Attitude towards the Brand was measured using Mitchell and Olson’s (1981) 5-point semantic differential scales bad/good, dislike/like, not irritating/irritating and uninteresting/interesting (cronbach alpha= .932). A match-up (both congruence and relevance) between the image of the brand and the image of the celebrity was measured using Till and Busler’s (2000) three item 5-point scale consisting of does not belong with/belongs with, does not go together/goes together, does not fit together/fits together (cronbach alpha= .984). Procedure Subjects were directed to view one of three advertisements. In the irrelevant information condition, subjects viewed Jerry Seinfeld in an advertisement for the Greater making no reference to the Greater brand. Instead, Jerry crawls along a desert. In the relevant information condition, subjects saw an advertisement that featured a Greater customer enthusiastically discussing her positive experience with the Greater brand highlighting her satisfaction and clearly stating a brand benefit of high interest on savings accounts. Jerry Seinfeld is not featured. In the combined relevant plus irrelevant information condition, subjects viewed the entire advertisement that combined the Greater customer discussing the brand and Jerry Seinfeld crawling along the desert. The advertisements for each condition featured the Greater Building Society’s brand logo at the end, with subjects exposed to the advertisement for approximately 20 seconds each. Once subjects viewed their allocated advertisement, they were then asked to evaluate the advertisement they just viewed. Firstly, subjects were asked their attitude towards the advertisement, then their attitude towards the brand and finally, the degree to which they believe the image of Jerry Seinfeld matches that of the Greater Building Society’s image. RESULTS The univariate statistics for attitude towards the advertisement and attitude towards the brand for the three conditions are shown in Table 1. The statistics are also broken down by perceptions of match-up. Interestingly, attitude towards the ad is slightly higher in the irrelevant and relevant combination condition than in the relevant only condition. Attitude towards


70

the brand is highest in the relevant condition with the irrelevant condition receiving the lowest score. In each condition consistently higher scores occur when subjects see Jerry Seinfeld as a good match with the Greater Building Society. Table 2 shows the two-way ANOVA results for both of the dependent variables, attitude towards the ad and attitude towards the brand. Results for attitude towards the advertisement indicate significant main effects for advertisement condition and for match-up. A significant interaction effect also exists, as highlighted in Figure 1. Post hoc comparisons in Table 3 show subjects report significant differences between all three conditions. For those who perceive a match, their ad attitude is significantly weaker in the irrelevant condition compared to the other two ad conditions. For those who perceive a mismatch, their ad attitude is significantly more positive in the relevant ad condition compared to the irrelevant and combination relevant plus irrelevant conditions. Subjects who perceive a mismatch between the celebrity and brand report attitudes towards the advertisement as hypothesized. These subjects report the most positive attitudes for the relevant condition, followed by significantly lower attitudes in the combination condition and the least favorable attitudes reported for the irrelevant condition, where the celebrity does not mention the brand nor provide any brand-related information, providing support for H1a, H1b and H1c. However, when subjects perceive a match-up between both brands they report strong positive attitude towards the advertisement in both the relevant advertisement condition and the combination relevant plus irrelevant condition, when the brand and celebrity feature together, providing no support for H2a and H2b. Subjects who perceive a match, however, do report the weakest ad attitude in the irrelevant condition featuring the celebrity only, providing support for H2c. Results for attitude towards the brand do show subjects report more positive attitudes after viewing the relevant advertisement condition and less positive attitudes for the combination with the least positive for the irrelevant condition, but the differences are not significant. There is a significant main effect for match-up, but no interaction effect between advertisement condition and match-up. Figure 2 graphically shows those subjects who perceive a match-up between the brand and celebrity report significantly higher brand attitude in all experimental conditions, than those subjects who perceive a mismatch. Subjects report the strongest brand attitude in the relevant condition, featuring no celebrity and weakest in the irrelevant condition, featuring the celebrity only, showing support for H3. DISCUSSION Results of this study identify that when celebrities act as indirect spokespeople in advertisements, where they do not provide information about the endorsed brand, consumer brand attitudes and advertisement attitudes are affected. While previous research in the celebrity endorsement literature centres on the concept of match-up of relevant characteristics between a celebrity and brand (Kamins, 1990: Kamins and Gupta, 1994; Till and Busler, 1998, 2000; Till et al., 2008), findings from this study suggest that the irrelevancy of celebrity information presented in an advertisement is an important construct that can also influence consumer judgments. Several findings from this study are consistent with the celebrity endorsement literature whereby consumer evaluations are less positive when they perceive a mismatch between the celebrity and brand (Kamins, 1990: Kamins and Gupta, 1994; Till and Busler, 1998). The results of this study show that consumers who perceive a high degree of fit have the strongest attitude towards the advertisement and brand attitude than those who perceive a mismatch between the celebrity and brand.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

71

Consumer attitude towards the advertisement is significantly affected by irrelevant brand information provided by a celebrity within an advertisement. Findings from this study show consumers who perceive a match-up between the celebrity and brand report enhanced advertisement attitudes when the celebrity and brand feature together in the combination irrelevant and relevant condition with ad attitude just as strong as when the brand features alone in the relevant information condition. On the other hand, consumers who perceive a mismatch report less positive attitudes towards the advertisement in the combined condition when the celebrity features. As such, this study provides additional insights into the concept of match-up as consumer beliefs in the celebrity providing consistent information is strengthened when they perceive a match-up and view an advertisement that features a combination of the celebrity and brand, both relevant and irrelevant information. With the use of celebrities providing irrelevant brand information in advertisements becoming an increasingly popular advertising execution tactic, it is particularly important to investigate the reasons for this shift in marketing communications. Results of this study identify that placing both a celebrity and brand together in an advertisement can have a negative impact on the brand. Research is yet to explore the reciprocal effect of endorsement relationships on consumer judgments of the celebrity. With consumers becoming more skeptical of the motives of celebrities in endorsements (Bailey, 2007), celebrities are discovering ways in which to avoid perceptions of ‘selling out’ and maintain an image of authenticity. When celebrities feature in advertisements and provide only relevant information regarding their own celebrity brand, they are in fact, preserving consumer perceptions of the “real” them-their core identity. Future research should explore consumer perceptions of celebrity authenticity in an endorsement context. Future research should also explore celebrities that provide moderately relevant information, making an implicit endorsement for the brand. For example, Jerry Seinfeld in recent advertisements for the Greater cuts down branches in a jungle whilst stating that he is making room for new branches. Here, Seinfeld is making an implicit endorsement for the brand by referring to aspects of the brand’s product. Since we explore the irrelevancy of information presented by a celebrity endorser in television advertisements, future research should be extended to investigate whether these effects are transferable to other mediums, such as print. Print advertisements may feature celebrities who provide irrelevant brand information and only relevant celebrity information, such as many fragrance advertisements that feature the celebrity only, providing no verbal endorsement of the brand nor featured holding the branded product. For example, the Gucci by Gucci fragrance ads present James Franco’s face only with the bottle presented at the bottom of the advertisement. Although James Franco may provide relevant brand information by conveying a sexy and modern image, congruent with the Gucci brand, he provides irrelevant information by not explicitly advocating the brand or holding the brand in the advertisement.


72


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

perceptions of Advertising

73


74

Förord: Påverkar kön bedömningen av kreativitet i reklamvärlden? Reklam är inte ett mysterium men att göra reklam som säljer är svårt och fordrar idéer som både väcker nyfikenhet och särskiljer varumärket på ett relevant sätt. Alla som jobbar med marknadsföring förstår att det kräver insikter om både konsumenterna och varumärket. Och för att lyckas krävs professionalism och ett förtroendefullt samarbete mellan kund och reklambyrå. Ändå lider reklambranschen fortfarande av fördomar som i värsta fall försvårar samarbetet mellan byrå och uppdragsgivare till allas förlust. Författarna till den här undersökningen har försökt sticka hål på en av dem – att kvinnliga kreatörers reklam omedvetet har svårare att hävda sig i byråbranschen. Och att det är därför branschen fortfarande är mansdominerad. Att döma av denna studie är det inte sant, åtminstone bland framtidens kreatörer som förefaller befriade från fördomar och kapabla att strikt professionellt bedöma reklam utifrån sakliga kriterier som originalitet, relevans och tydlighet. Inte heller finner författarna det styrkt att det finns ett glastak för kvinnliga kreatörer att göra karriär i byråvärlden.

Peter bergendahl


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

75

IS THERE GENDER BIAS IN THE ASSESSMENT OF ADVERTISING CREATIVITY? David Roca Universitat Autònoma de Barcelona, Spain Daniel Tena Universitat Autònoma de Barcelona, Spain Patrícia Làzaro Universitat Autònoma de Barcelona, Spain

INTRODUCTION Gender discrimination in creative departments has been studied using surveys, content analyses, and qualitative methods (Broyles, S. & Grow, J. 2008). Previous research depicts that there exists interest in variables related with assessing advertising creativity (Koslow et al. 2003a, Sasser & Koslow 2008), practitioner profiles and customers reviews (West et al. 2008) and models working for non-western advertising (Kim et al. 2010), etc. On the other hand, some articles present extensive research regarding creative departments and relationships among professionals (Gelade 1997, Hackley 2003, Nixon 2003, Chong 2006, Hackley & Kover 2007). However, little attention has been given to gender influences on ads evaluation. This research is the first step towards studying gender discrimination during selection of creative ads using experimental methodology. The main objective is to study if there is an unconscious discrimination towards advertising ideas by females. Previous research regarding discrimination in the selection processes of advertising is presented by Sego (1999) who stated: “The effects of sex and ethnicity on evaluations of advertising job candidates: do stereotypes predict discrimination?” Novel contribution of this article is the use ofthat stimuli tests are based in creative ads, which focuses onsimulating a student’s creative portfolio ratherand not in a general qualifications (grade point average at university grades, saleselling experience and recommendation letter for an entrylevel position in the accounts department). Literature perusal suggests that female students dominate advertising programs (Fullerton et al. 2006, Mallia 2008, Weisberg and Robbs 1997) and account departments (Klein 2000, Pueyo 2010). The number of male and female students is equal in portfolio schools (Grow, Roca & Broyles 2010). However, presence of women is limited in creative departments of various western countries – Spain, Sweden, United Kingdom, and USA (Alvenson, Kelly 2000, Martín 2007, Pueyo 2010, Weisberg and Robbs, 1997, Grow, J.M. & Broyles, S.J. 2011), which is linked to different factors. Creative department is considered as laddish creative environment, or what has been defined as men’s club (Alvesson, M. 1998, Fitzsimmons, E. 2008, Nixon, S. & Crewe, B. 2004, Broyles, S. & Grow, J. 2008, Etayo, C. & Del Río, J. 2008, Mallia 2009, Gregory 2009). One of the reasons is male networking outside the agency (Gregory 2009, Ibarra 1992) which leads to male dominance in judgment committees (Grow, J.M. & Broyles, S.J. 2011) which is about 80% in Spain (Roca, et al. 2011). Women lack peer recognition since they do not participate in such networks (Bosman, J. 2005, Cuneo, A. & Petrecca, L. 1997). In contrast, they appear as pink ghetto, especially when isolated to work for female products such as those, related with beauty, cleaning and children (Roca&Pueyo 2011). Some other variables explain that absence of females from creative departments is related to directive knowledge (Martín, Hernández & Beléndez 2009), lower salaries in the advertising industry for women (Martín 2007), lack of creative leadership positions (Gonzalez Andrío 2005, Pool 2001) and, especially, motherhood penalty (Grow, J.M. &


76

Broyles, S.J. 2011, Mallia 2009), which is not exclusive to advertising industry (eg. Correll, S., Benard, S. & Paik, I. 2007). Some research about gender norms associated with creative departments of advertising agency with respect to the perception of students has been developed (Windels et al. 2010). Results revealed that majority of advertising students (80.2%) perceived this field to be male- dominant. However, “gender perceptions did not influence decision-making when choosing a creative director for a student-run agency.” So, the question is, if gender influences while choosing best creative ads? Do judges experience “unconscious” discrimination and are, therefore, gender biased? LITERATURE REVIEW Prejudice against women can be generally examined with two kinds of studies. Firstly, assessing attitudes toward some groups by getting conscious answers (e.g. women) through researches where respondents are aware of examined attitudes towards gender. Secondly, conducting studies with less information than that occuring in real sets where respondents are not aware that gender biases are being examined (unconscious answers). The former gives a general description of overall public evaluation regarding social groups, since, the answers may be politically acceptable (conscious answers). The latter methodology uses unobtrusive measures and usually employs student samples rather frequently engaging professional samples. These studies evaluate gender-bias with respect to articles, essays, job applications, resumes, paintings, etc. The review examines some studies conducted on the second group (unconscious answers). Surprisingly, only two are related with advertising (Sego 1999 and Windels et al. 2010). One of the previous surveys (Hartman 1988) states that creative directors do not discriminate between men and women (conscious answers). Golberg (1968) was the first to experiment the concept of “unconscious” discrimination by changing the author’s names. He states; “Are women prejudiced against women?”. Results of his research are known as “Goldberg paradigm.” His experiment assessed students’ perceptions regarding six academic articles from various professional backgrounds written supposedly by men as well as women. The results indicated that identical articles written by men were better evaluated, although “bias against women’s work was significant only for the two articles in traditionally masculine fields” (gender congruity). Women assessed women worst than men, even for feminine fields. Swim et al. (1989) undertook metaanalysis of 123 “real world” settings, using students. They reported hardly any gender-bias related with the articles. Only, a slight effect was observed when the work being judged consisted of masculine domain. Eagly et al. (1992) drew conclusions from meta-analysis of studies regarding gender effects on evaluation of leaders, similar to those of Swim et al (1989). However, 87 of the 114 papers reviewed used student as subjects, and consequently the extent to which the findings can be extrapolated to the “real-world” is questionable (Buttner and Rosen 1989). It was found that female leaders were evaluated slightly less favorable as compared to males, particularly for stereotypic masculine styles. Earlier, Olian, Schwab, and Haberfeld (1988) have examined job applications and found considerable tendency for men to fare better than women (d = 0.41), but almost 80% of studies were based on ‘male-type’ positions. The first investigation on gender bias in advertising was conducted much later than Goldenberg. Sego (1999) researched whether advertising professionals have negative beliefs when hiring women and black people. Her goal was to “initiate empirical study of discrimination in advertising employment practices by assessing stereotypical beliefs of surrogate subjects, and creating, in a laboratory setting, a situation requiring those


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

77

subjects to evaluate the qualifications of candidates for an advertising position” (p.66). She focused on candidates who had applied for an entrylevel position in account services. Results showed that sexist stereotypes did not modify the evaluations for advertising candidates; moreover, interaction between feminist beliefs and candidate sex were not significant. Male candidates were evaluated little better than females. Recently, Windels et al. (2010) examined students’ perceptions regarding gender norms associated with creative departments of advertising agency. Experiment constituted evaluation of resumes from students for the position of creative director of a student-run agency setup in the Department of Advertising. They discovered that majority students studying advertising perceived creative advertising as male-dominant (80.2%). This perception was prevalent regardless of gender and backgrounds. However, it was discovered that gender perceptions did not influence decision-making when choosing a creative director for a studentrun agency. The experiment was built on fictitious text-based resume, including professional experience and education, similar to that conducted by Sego (1999). They did not check the creative portfolio, a tool, fundamental among professionals to access creative departments or for changing agency. Not much research regarding gender-bias in hiring for advertising has been conducted. HYPOTHESES Creativity departments are traditionally a masculine domain. “Maledominated jobs, whichare thought to require characteristics stereotypically ascribed to men.” (Eagly & Mladinic 1994). Is this true for creative departments? This study explores whether author name changes the perception of creativity level in print ads. A real portfolio was simulated based on this assumption. Three hypotheses were established to measure gender-bias related with evaluation of a creative portfolio. H1 Future advertising professional men prefer to choose ads created by males. H2 Future advertising professional women prefer to choose ads created by both males and females. H3 Gender authorship of the ad is a deciding factor when choosing the best creative ads. METHODS Respondent Sample and Procedure A questionnaire link and a letter to support the research were sent to 62 faculties teaching advertising in Spain. Additional details were provided through telephonic follow ups, in case the professors required the support. The teaching staff informed their students about the experiment. A draw for an I-pad was used to stimulate their involvement. Total of 770 students studying advertising at 33 Spanish universities (public and private) participated in the study. Following the advice from other researchers, larger sample was used for “statistically significant gender effect” (Fay & Williams 1993, Eagly & Mladinic 1994). Questionnaire for experimental manipulation were available online for 3 weeks (October 2012). First part of the questionnaire had instructions for the students, which took approximately 10 minutes. Slightly less than ¾ of the respondents were female (n=555) who were present at the university. According to our sample data, 44.6% male students and 39.46% female students wanted to work as creative professionals. Ads sample and manipulation Creativity professors selected 36 print ads, created by students of upperdivision advertising class. The second stage consisted of online test, where 60 students evaluated the ads as per instructions to rate according to the scale based on (good, average, bad) following the creativity


78

indicators of Koslow, Sasser & Riordan (2003). The model presented by these authors comprised of strategy, originality, and artistricity, which were converted into clarity, originality and attractiveness, for better understanding by the students. Finally, 9 ads were selected: 3 for each creativity rating category i.e. good, average, and bad. Names and surnames: male, female and neutral (ex. Mario, Maria, M.) were added to the final selection. Total of 27 combinations were obtained comprising of 9 masculine names and surnames, 9 female names and surnames and 9 initials and surnames. The number of names was selected randomly in order to avoid Kasof’s bias (1993) which states that female names are usually less attractive than those of males, when minimal information is shown in gender-bias researches. Participants were instructed to “imagine themselves being a creative director of an advertising agency and rate the ads” according to the instructions. Questionnaire was divided into four parts. The first part contained the presentation and general instructions. The second part consisted of questions in order to know the profile of the respondents. Third section constituted of instructions to fill the form that had nine different ads.. The final section assessed the sexist beliefs using open and closed ended questions (dichotomous-questions and responses) and closed questions (eg. men, women or indifferent). At the end of the survey, respondents were thanked and informed about the draw for the I-pad. FINDINGS Data was analyzed using SAS® v. 9.5 by statistics office at authors’ university (SEA). Statistical significance was p ≤ 0.05. We used chi-square (x2) to test whether the respondent population varied by gender. We used 3x3x3 contingency analysis to primarily test whether designation of names such as; male, female, or initials influenced quality of ratings i.e. good, average or bad). H1 Future advertising professional men prefer to choose ads created by males. Hypothesis 1 is not supported by the results. Future male professionals preferences were ads created both by men and women with no significant differences (x2= 1.4383, df=2, p=0.4872). Gender is not a key factor in preferences of men when choosing creativity quality in ads. H2 Future advertising professional women prefer to choose ads created by both males and females. Hypothesis 2 is supported by the results. Future female professionals preferred ads created both by men and women with no significant differences (x2= 0.8285, df=2, p=0.6608). Again, gender is not a key factor in preferences of men when choosing creativity quality in ads. Results from first two hypotheses lead us to conclude that respondents do not discriminate or are not influenced by creative’s gender when assessing ads. The gender of author/judge is a non-determinant in this case. H3 Gender authorship of the ad is a deciding factor when choosing the best creative ads. Hypothesis 3 is not supported by the results. Gender authorship is not essential when assessing ads created by future advertising professionals (x2= 2.59, df=2, p=0.2737). The data obtained depicts that “un-conscious discrimination” does not prevail in future ad professionals at the university. Stimulus and authorship gender does not affect the evaluation. Even, if gender discrimination exists, it must be found at advertising agencies rather at earlier stages such as university as pointed out by Windels et al. 2010. It was observed that women evaluated ads slightly better than men.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

79

DISCUSSION The purpose of this study was to evaluate gender-bias when assessing print ads. Large sample consisting of advertising students was used for this research. Although creative departments are highly masculinized, but it was found that there was no gender-bias (due to authorship or judgement) with respect to the evaluation of the ads. Thus, gender congruence hypothesis/theory that states women face discrimination in male occupations (Eagly, Makhijani, and Klonsky 1992) was not confirmed. These results are congruent with earlier research regarding advertising assessment of gender-bias (Sego 1999 and Widels et al. 2010) and are very similar to the results of the researches in other fields (Swim et al. 1989, Eagly et al. 1992 and Olian, Schwab, and Haberfeld 1988). Widels et al. (2010) found that “boys club culture does not start in the classroom, since students who were aware of gender proportions at ad agencies were not perpetuating this bias in the university environment.” So gender-bias should be found in the workplace rather in the educational setting. Ruble (1982) expounds after reviewing research on gender stereotypes that, ‘males and maleness become preferred with increasing age’ (p. 225). Pheterson (1969) cited by Ward (1979) suggested, “individuals may require a certain amount of familiarity or competence in a field before they are prone to render negative appraisals of stimulus objects.” Further research must be carried out at real sites, with real creative personnels. This would help understand if absence of women in creative departments could be related to gender-bias in idea selection. LIMITATION AND FUTURE RESEARCH Porfolio is essential for young students to access advertising agencies. Major limitation of this study was to use students as respondents, but earlier research used this kind of sample to study gender bias when hiring through fictitious ads (eg. Bosak, J. & Sczesny, S. 2008), or for the assessment of paintings (Ward 1979), employment applications (Halon & Cole 1975), research articles (Paludi&Bauer 1983, Levenson et al. 1975) or even loans applications (Fay&Williams 1993). Despite this, we believe it gave us information about the absence of gender discrimination in schools for future advertising professionals. We look forward to replicate this research with a sample of creative professionals working in advertising agencies to observe if gender discrimination exists when selecting ideas in a well-known male environment such as creative departments (Grow & Broyles 2011). This, data could be extrapolated to the “real-world” setting. Further research will allow us to answer the statement of Canary and Hause (1993): “Are there sex differences in communication? We believe there are sex differences in communication, but they are eluding us. Perhaps a definitive answer to the question of sex differences in communication will arrive within the next fifty years.” Time to answer this question has come for the creative departments in advertising.


80

FÖRORD: Receptet på ett lyckat förhållande ”Det är inte ens fel att två träter. De är inte ens fel.” Så lyder svärfaderns visdomsord till bröllopsparet och så kan man tänka att författarnas hypotes till denna uppsats skulle kunna låta i andra ord; svaret på den klassiska frågan huruvida kundens inblandning i den kreativa processen är bra eller inte. En rimlig hypotes som troligen de flesta med lite erfarenhet kan skriva under på. Bakom varje framgångsrik kampanj finns oftast en engagerad uppdragsgivare. Men det är sedan som det blir svårt. Går detta att statistiskt bevisa? Framförallt när ett av mätetalen är kreativitet. Möjligen sov jag mig igenom mina högskolekurser i statistik, men jag förhåller mig skeptisk till författarnas modellbyggen. Kanske kan man bara förlita sig på det sunda förnuftet och exemplens makt? Och då kanske en antologi över goda exempel bättre hade bevisat hypotesen?

jan Lindforss


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

81

A CLIENT PERSPECTIVE ON ADVERTISING QUALITY Richard J. Calderwood University of Waikato New Zealand Scott Koslow Macquarie University Australia Sheila L. Sasser Eastern Michigan University US

INTRODUCTION An area of active debate concerns who—clients or agencies—are most responsible for advertising quality. For example, marketers and agencies frequently attempt to pass the blame to the other as to who is more responsible for advertising quality. One perspective suggested that the blame for the lack of advertising quality should be placed on the client (Koslow, Sasser and Riordan 2006). However, another perspective places the blame on agencies saying they lack the needed skills. (Cumeo 2003; Elliott 2003; Sanders 2003 White 2003). Clients may sometimes feel they have something to add to the campaign development process, yet feel they are thwarted by agencies who wish to retain control of the process. It is unfortunate that most of the empirically oriented research has focused on data collection from agencies. To add balance to the debate, brand managers are asked about what factors they consider impact advertising quality. This research focuses on four key factors, client involvement, relationship stability, competency and expertise and flexibility. This is tested on a sample of 162 campaigns from brand management clients in Australia and New Zealand. Results show a number of complex interactions and support the role of clients in the advertising development process. CLIENT INVOLVEMENT Overall, agencies need access to insightful information about the client’s business and even Koslow, Sasser and Riordan (2006) acknowledge that access to the client’s top management is a key resource provided providing the agency with strategic insights. Clients are in control of three elements of the campaign development process. Firstly they are set the direction for the campaign through the client brief (Duncan, 2002; Wells, Burnett and Moriarty, 2003). Secondly they are control the allocations of resources to campaign development. Thirdly they are also able to control the evaluation of campaigns (Brown, 1999). To develop advertising that is of strategic value, a key element of the campaign is going to be client involvement. Although these are rational reasons for clients to be involved, there are also emotionally based reasons. One way to consider the emotional issues clients may have in approving work is from a priming perspective. What client involvement provides is a means by which managers can become familiar with edgy work that will draw consumer attention well before the managers have to make a final decision on it. A well-known effect in priming is that when one is presented with novel stimuli, they are often not liked. However, with repeated viewings, they become more liked, even if they are not remembered (Marsh et al, 1999; Smith et al, 1993; Ward, 1994). Thus, with an involved client, nothing will come as a surprise and clients are more likely to like and hence approve original advertising. And if the original work is well informed by a client interaction, then it will be creative as well. H1a: The extent of client involvement in the campaign development process is likely to impact upon the levels of originality evident in campaigns produced.


82

H1b: The extent of client involvement in the campaign development process is likely to impact upon the levels of creativity evident in campaigns produced It is likely that if the client is involved that together the client and agency are going to produce campaigns which are higher quality and more effective. Client being involved means they are able to share a brief which will mean that the campaign is more likely to be on strategy. Client and agency can make decisions about which media to use and why it is likely to be more effective. They can evaluate a campaign according to a pre-developed set of criteria. If these evaluative criteria are carefully chosen, the client and agency will be better able to judge whether a campaign will be effective and then produce more effective work. H1c: The extent of client involvement in the campaign development process is likely to impact on the levels of campaign effectiveness Another concern is that if clients are not involved they are not likely to share relevant information with the agency through an effective creative brief. Thus the agency may be unsure of what strategy to adopt and can only take cues from previous campaigns for similar products. If on the other hand the client is heavily involved they can provide a range of useful information to the agency which can improve the likelihood of the campaign being on strategy. Whether the client and agency are in s stable relationship can also influence whether a campaign which is on strategy is produced. In stable relationships the client and agency are able to share information which is important to developing effective strategy. This information may come from research or from experience with previous campaigns. H1d: Clients and agencies who are neither involved nor in stable relationships are less likely to produce campaigns which are on strategy RELATIONSHIP STABILITY When the client and agency are in a stable relationship it is likely that more original campaigns will be produced as the agency does not feel threatened. That is, when agency employees are not in a stable relationship, they feel more apprehensive about delivering original work that may offend the client (Sasser and Koslow 2008). H2a: When the client and agency are in a stable relationship they are more likely to produce an original campaign In situations where the client and agency are in a stable relationship and they both have high levels of competency and expertise, they are better able to develop campaigns which stretch the creative envelope in terms of originality. Being in a stable relationship, they both understand how their counterparts in the other organisation operate. They are likely also to know their strengths and weaknesses. The high levels of competency and expertise is also able to ensure that creative is not produced which is not on strategy or overly original for the particular product category or service. H2b: When the client and agency are in a stable relationship and they both have high levels of competency and expertise they are more likely to produce creative campaigns Given that successful campaigns are often longer running we would also expect that campaigns that are produced when the client and agency are in a stable relationship, would be more effective. We might expect that if the client and agency have been together for a considerable period of time they would know better what was likely to be an effective campaign strategy for their product or service category. They will also have developed a sound knowledge of the target audience for the campaign and the


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

83

audiences processing behaviours. The client would be likely to show higher levels of confidence in their agency to produce more effective campaigns and not need to supervise their activities as closely once the strategy has been decided. H2c: Relationship stability is likely to lead to the development of campaigns which are more effective. AGENCY FLEXIBILITY The extent to which an agency shows a flexible approach towards their clients in the campaign development process is also an important contributor to the development of quality advertising campaigns. Agencies that are more prepared to work with the client and not to become overly inflexible are more likely to produce campaigns which are more strategic. This may because they have learnt from the client or have revisited the brief and made changes to improve the strategy being used. It may also be that they have made changes based on previous campaign experiences for similar product or service categories. H3a: Agencies that are most flexible in campaign development are likely to produce campaigns which are on strategy Those agencies that are either totally inflexible or very flexible in their approach to campaign development are likely to produce work which exhibits the highest levels of originality. The totally inflexible agency will usually believe that they know what level of originality is best for a particular campaign based on past experience. They will tend to be less willing to make changes. In comparison those who are more flexible may be may be able to learn from the client or they may simply act as a â&#x20AC;&#x153;doormatâ&#x20AC;? for the client. These changes are likely to bring about more original campaigns. Those with moderate levels of flexibility are likely to produce the least original campaigns as they may not really be sure of what the overall strategy for the campaign is. H3b: Agencies with either high or low levels of flexibility are likely to produce the most original campaigns Those agencies which are flexible are also likely to produce the most effective campaigns. This will come from the agencyâ&#x20AC;&#x2122;s willingness to adopt a flexible approach to all aspects of campaign development. It is unlikely that agency flexibility will contribute to the creativity of campaigns as whilst a campaign may have an appropriate strategy the levels of originality may not be such that we could always say that the campaign was both original and on strategy. Flexibility may be indicated by a willingness of the agency to makes changes to creative when research suggests that the campaign may not have been on strategy. Such changes may contribute to more effective campaigns. H3c: Agencies who adopt a flexible approach to campaign development are likely to produce the most effective campaigns COMPETENCY AND EXPERTISE Clients are important, but there are things agencies have to bring to a relationship and the most important of these is competency and expertise. When there are high levels of competency and expertise on behalf of the client and agency the campaigns produced are likely to be based on much better strategy and be more original. This is because the strategy developed by the client will be more easily understood by the agency. The fact that the agency has high levels of competency and expertise may also mean that the client is more willing to accept more original work as they believe that the agency knows what they are doing and they are preparing campaigns that exhibit high levels of strategy.


84

H4a: Clients and agencies that have high levels of competency and expertise are likely to produce campaigns which are on strategy. Clients and agencies with high levels of competency and expertise are likely to produce more original campaigns as they have personnel with the skills to prepare original work. Those agencies with the highest levels of competency and expertise are also likely to attract the most skilled creatives to work with them. Clients are also likely to be more enabling towards agencies with high levels of competency and expertise because they believe that they have the skills to develop original creative which is in line with their desired strategy. H4b: Clients and agencies that have high levels of competency and expertise are likely to produce campaigns which are more original. It is suggested that those who have high or low levels of competency and expertise are likely to produce less effective campaigns than do those with moderate levels of competency and expertise. This may be due to the fact that those who have low levels of competency and expertise are not as well placed to judge whether an effective campaign is likely to be produced and those with higher levels of competency and expertise may have become complacent and pay less attention to detail in campaign development. Those with moderate levels of competency and expertise are likely to still have an enthusiasm for campaign development as they develop their careers. H4c: Clients and agencies that have either high or low levels of competency and expertise are likely to produce less effective campaigns than those who have moderate levels of competency and expertise. METHOD This research used questionnaires in which brand managers reported on up to three of their own campaigns. The sample was 60 brand managers in Australian and New Zealand, reporting on 162 campaigns. This represented a 40% response rate. Although the number is small, the number of brand managers in Australia and New Zealand is small as well. The researchers asked that only those with experience in 3 or more campaigns fill out the questionnaires, and this low response rate largely reflects the reality that advertising experience does not run deep in Australasian brand management. The study used dependent variables: strategy (4 items ~=.866), originality (5 items, ~ =.919, effectiveness (4 items, ~T.858) and creativity which was constructed using the scales for strategy and originality. Thus, the first, second and fourth measures drew from Koslow, Sasser and Riordan (2003). Also measured were four independent constructs: relationship stability (4 items, ~=.879), client involvement (3 items, ~=.656), competency and expertise (5 items, ~=.827), and agency flexibility (3 items, ~=.707). Full item descriptions are available from the authors. A separate factor analysis was performed on the three dependant variables and another one on the independent variables. The factor analysis for the dependent variables explained 79.2% and for the independent variables, 67%. The correct items loaded on the correct factor with no loading less than .6. For cross loadings, no item was more than .4. To form constructs, raw item scores were first summed, then centred and scaled. Each of the four dependant variables were predicted using each of the four independent variables. In initial runs of the data, all two way interactions and squared terms were include in the models, then non-significant terms were backwards deleted to determine the final models.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

85

FINDINGS The regression outputs for the four models are presented in Tables 1-4. Interactions and other complex effects were plotted in Figures 1-7. To show the slopes, the high level was defined as one standard deviation above the mean and the low was one standard deviation below. Table 1 shows positive significant parameters for both competency and expertise and agency flexibility and support H4a and H3a, respectively. Figure 1 also shows the interaction of relationship stability and client involvement. When involvement is low, stability has a strong effect, supporting H1d. Table 2 shows a positive effect of client involvement on originality supporting H1a. Figure 2 identifies the U-shaped relationship of agency flexibility on originality, which upholds H3b. Additionally the interaction between relationship stability and competency and expertise is plotted in Figure 3. When competency and expertise is low, stability has a stronger effect. This provides qualified support for both H2a and H4b. Table 3 demonstrates how when client involvement increases, so does creativity, which bolsters H1b. Figure 4 illustrates when competency and expertise is low, relationship stability has a stronger effect on creativity, confirming H2b. Figure 5 explores an unexpected exponential effect of involvement on creativity, adding detail to H1c. Finally, Table 4 illustrates the positive effect of relationship stability and agency flexibility on effectiveness, providing backup for H2c and H3c. The expected inverted U-shaped relationship of competency and expertise on effectiveness is also substantiates H4c. CONCLUSIONS Overall, strong support is found for the role of clients in the creative process. In each of the four models predicting strategy, originality, creativity and effectiveness. Clients are needed to supply needed information that agencies need to do good work. While agencies might want to treat clients as outsiders in the process, clients tell us that they need to be involved. Maybe it may not be needed with every single campaign if there is a strong relationship, but somehow the knowledge the client has needs to be incorporated.


86

förord: En studie om de bästa motivationsfaktorerna för att få människor att konsumera grönt I debatten om vilken typ av argument som bäst motiverar oss som konsumenter, har vi genom åren fått höra en rad olika teser. Klimatforskarna verksamma vid Stockholm Resilience Center testar för närvarande hypotesen att vi hellre tar en väl kalkylerad förlust än chansar på en odefinierat stor risk. Likaså har vi fått höra att miljöpropagandisternas hotbilder för jordens utveckling lätt avfärdas med att ”det där påverkar ändå inte mig”. Närodlad och ekologisk mat hade svårt att hitta en bredare marknad så länge de marknadsförde sig genom vårt kollektiva dåliga samvete, men marknaden tog fart när argumentet om en bättre egenhälsa dök upp. Personligen har jag trott att vi alltid är mer öppna för argument som handlar om den egna vinningen/förlusten än om kollektiva diton. Just därför är Alexandra Langers studie så intressant. Hon menar att i en vinnarkontext är att visa på kollektiva fördelar starkare argument än att visa på egenfördelarna. Och att i en förlustkontext är det precis tvärtom, där är egennackdelarna/-fördelarna starkare argument än de kollektiva fördelarna/förlusterna. En viss reservation måste göras gällande urvalet studenter materialet testades på; hur representativa de kan anses vara för konsumenter i allmänhet? Men om de kan anses vara representativa, är det alltså så att argument om att duscha mindre på grund av sjunkande grundvattennivåer måste ligga nära den egna individen (ditt vatten tar snart slut!), medan man kan argumentera för att köpa ekomjölk för att rädda kulturlandskapen Vad som är allra mest effektivt, verkar vara att använda egennackdelarna/ fördelarna i en förlustkontext. Vad jag själv kan vinna när alla andra förlorar, är det som biter bäst, menar Langer et al. Och den logiken är det lätt att skriva under på, den känner vi igen också från lyckoforskningen och statusforskning – våra vinster blir större och viktigare mätt mot andras förluster. Att vi alla förlorar i det långa loppet verkar vi människor vara alltför kortsiktiga för att kunna ta in. Och det är oerhört viktigt för oss kommunikatörer att inse när vi ska förklara de klimatmässiga utmaningar som ligger framför oss.

henrika Thomasson


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

87

“I” lose, “OTHERS” gain – MESSAGE FRAMING AND BENEFICIAL APPEALS IN ADS PROMOTING GREEN CONSUMPTION Alexandra Langer European University Viadrina Germany

1. Green Consumption Current consumption lifestyles put the environment in danger (United Nation 1992). Enhancing green consumption, i.e. a consumption lifestyle that reduces damage to our natural surroundings, is therefore of great importance. The ecological impact of consumption has also become an important factor in consumer choice. More and more consumers are willing to change their consumption style and claim to pay attention to ecological claims and labels (IfH 2010) and are willing to pay a higher price for ecological friendly products (GfK North America 2008). Successful advertisement of green products is therefore crucial for marketers, consumers, and society as a whole. Despite the pressing importance of altering our consumption behavior, prior research has come up with mixed results on what message type best encourages this consumer behavior. The current research makes important contributions, both theoretically and substantively, to the literature. In investigating the effectiveness of different framing types when combined with a self-benefit (vs. other-benefit) appeal, this research aims at answering the question under which circumstances gain-framed messages are more effective than loss-framed messages, by examining the moderating role of beneficial appeal (self- vs. other-benefit). This is an important question because marketers often use both appeal types in advertisement for green products. A small-scale pilot study was conducted with a random sample of 25 online ads for green brands (Interbrand 2011). The results of the pilot study showed that both gain- and loss- framed messages were commonly used (12 used gain-framed messages, 9 used loss-framed messages and 4 used a combination of both). The question of how to effectively pair message frame and beneficial appeal is hence of practical importance. The current research contributes to several research streams. First of all, it helps to better understand the effectiveness of green advertisement. Specifically, it extends prior research on message framing (Tversky & Kahneman 1981). Message framing is a commonly used framework to predict consumer behavior in general (Block & Keller 1995; Maheswaran & Meyers-Levy 1990; Shiv et al. 1997, 2004) and ecological behavior in particular (White et al. 2011). So far research has come up with mixed results on whether loss frames (Tversky & Kahneman 1981; Meyerowitz & Chaiken 1987) or gain frames (Maheswaran & Meyers-Levy 1990) are more effective in altering consumer behavior. The study extends prior work on beneficial appeals in advertisement. Some studies have considered self- vs. other-benefit appeals and demo-


88

graphic moderators (e.g. gender) (Brunel & Nelson 2000; Nelson et al. 2006). This study demonstrates that variables that can be manipulated can influence the impact of beneficial appeals on consumer behavior. The results also further work on marketing appeals for charity donation (White & Peloza 2009), which has been shown to be significantly influenced by benefit appeal type. 2. Message Framing Message framing is a commonly used framework to predict consumer behavior. It involves the comparison of gain-framed messages (i.e. messages highlighting the positive consequenc- es of engaging in a particular behavior) and loss-framed messages (i.e. messages highlighting the negative consequences of not engaging in a particular behavior) (Block & Keller 1995; Maheswaran & Meyers-Levy 1990; Shiv et al. 1997, 2004). With regard to message effec- tiveness, prior research has come up with mixed results on whether loss frames (Tversky & Kahneman 1981; Meyerowitz & Chaiken 1987) or gain frames (Maheswaran & Meyers-Levy 1990) are more favorable. Moreover, studies have suggested that loss frames are more persuasive than gain frames in situations of high involvement (Maheswaran & Meyers-Levy 1990) or when concrete mind-sets are activated (White et al. 2011). Many of these studies have focused on health related issues or other behaviors directly relevant for the consumer’s own personal well-being. However, when it comes to green products, the benefit of consuming a green product might be more ambiguous: on the one hand there might be a benefit for the consumer him-/herself (e.g. the product is healthier or lasts longer) and on the other hand there might be a more general benefit for society as a whole (e.g. resources are saved or pollution is reduced). Recent frameworks of message framing thus do not adequately translate to decisions for ecological friendly products. This study intents to provide insight into the question of whether a gain or a loss frame is more effective in advertisement appeals for ecological friendly products, by examining the moderating role of whether the a self- or otherbenefit appeal is applied. 3. Beneficial Appeals There are two different ways of emphasizing the benefit of an ecological friendly product: Firstly, one can emphasize the benefit for the environment or society as a whole. Secondly, one can emphasize the benefit for the consumer him-/herself. The first one can be seen as a rather altruistic appeal (highlighting the benefit for others) and the second one can be seen as a rather egoistic appeal (i.e. highlighting the benefit for the consumer). Following (Fisher et al. 2008) the first ones are referred to as ‘other-benefit’ appeals and the second ones as ‘self- benefit’ appeals. Although it might be possible to use a combination of both, following prior research, in this study the two are separated to examine more closely which type of appeal is more effective. Prior research has shown that whether self- or other-benefit appeals are more effective depends on the decision situation. When the decision is made in public (vs. in pri- vate), other-benefit appeals (vs. self-benefit appeals) are more effective (White & Peloza 2009). Other studies have found that promotion-focused messages are more effective when an independent self-view is activated. In contrast, prevention-focused messages are more effective when an interdependent self-view is activated (Aaker & Lee 2001). Following the results from research on product involvement and message framing, we anticipate a higher involvement (vs. lower involvement) with messages using a selfbenefit appeal (vs. other-benefit appeal). Therefore gain-framed messages are expected to be more efficacious when paired with other-benefit appeals, whereas loss-framed messages are more efficacious when paired with a self-benefit appeal.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

89

Loss- (vs. gain-) framed messages are more effective when presented with a self- benefit (vs. other-benefit) appeal. 4. Method In order to test the prediction, an experimental study with a 2 (gain versus loss frame) x 2 (self-benefit versus other-benefit appeal) design was conducted. 82 students participated in the study (59.8 % female, age mean = 23.3 years) and were randomly assigned to one of the four conditions. As study object washing detergent was used and a fictitious brand (GreenCleanTM) was created to reduce potential variance created by the use of a real brand. A pretest confirmed that none of the study participants claimed to know the brand shown in the study material. Washing detergent is a commonly used product and study participants were expected to be familiar with the product category. Study participants were shown a fictitious advertisement, consisting of a picture and an advertising appeal, for GreenClean washing detergent along with a questionnaire. They were told to review the advertisement as if they were viewing it in a magazine and then respond to a series of questions. Each advertisement appeal reads as follows: “With the new GreenClean you can get the same sized load clean using only half the detergent, water, and energy!” The two independent variables, message framing and beneficial appeal, were manipulated by altering the message that followed. Half of the participants were in the gain-framed message condition and, following White et al. (2011), read: “Think about what will be gained if you use this ecological friendly detergent”. The other half of the participants were in the loss-framed message condition and read: “Think about what will be lost if you don’t use this ecological friendly detergent”. Beneficial appeal was manipulated following White & Peloza (2009). The next line of the message read either “Only half as much detergent and water is needed per load, so you can save money and also reduce your energy bill” (self-benefit condition) or “Only half as much detergent is needed per load, so ecological resources like water and energy are saved for everyone” (other-benefit condition). After reading the advertisement, participants rated their attitude towards the advertisement and towards the brand, their motivation to process the message, persuasion effect of the message and their intention to buy the product on 7-point scales each. As a check for the message framing and beneficial appeal manipulations additional questions were included. Participants rated two questions asking to what extend the advertisement focused on what would be gained if people used this product or lost, if people did not use the product on 7-point scales each (White et al. 2011). Manipulations for the beneficial appeal were tested following White & Pelazo (2009). In particular, participants evaluated the other- and self-benefit appeals (“To what degree is this an altruistic appeal [i.e. focused on helping others]?”, “To what degree is this appeal associated with looking out for the interest of others?”, “To what degree is this an egoistic appeal [i.e. focused on helping oneself]?”, “To what degree is this appeal associated with looking out for one’s own interests?”) on 7-point scales. An index of perceived self-/other-benefit was created by reverse scoring the first two items and averaging the items into one scale. The questionnaire ended with demographic measures. After filling out the questionnaires participants were debriefed and thanked. 5. Results Manipulation Checks. Manipulations for the type of beneficial appeal proved to be successful. Participants evaluated the self-benefit appeals as significantly more focused on self-benefits than the other-benefit appeal (Ms = 3.44 and 4.32; t(80) = -4.742, p < 0.001). The manipulation of message frame also proved to be successful. Participants in the gain-


90

framed message manipulation evaluated the gain frame manipulation check item as significantly higher than participants in the loss-framed message manipulation (Ms = 2.29 and 3.27; t(80) = 3.587, p < 0.001), whereas the opposite was the case for the loss frame manipulation check item (Ms = 3.25 and 4.32; t(80) = -3.103, p < 0.05). The impact of beneficial appeal and message framing on the dependent variables was explored, measuring both attitudes and behavioral intention. Results show a consistent strong interaction effect between beneficial appeal and message framing. A 2 (message frame: gain vs. loss) x 2 (beneficial appeal: self vs. other) ANOVA on attitude towards the advertisement revealed the expected interaction effect (F(1,78) = 7.56, p < 0.05). As anticipated, when confronted with a gain frame, participants reported more positive attitude towards the advertisement when it was paired with an other-benefit appeal (M = 4.47) than with a self-benefit appeal (M = 3.50). When confronted with a loss frame, participants reported more positive attitude towards the advertisement when it was paired with a selfbenefit appeal (M = 4.02) than with an other-benefit appeal (M = 3.55). A 2 (message frame: gain vs. loss) x 2 (beneficial appeal: self vs. other) ANOVA on attitude towards the brand also revealed the expected interaction effect (F(1,78) = 6.41, p < 0.05). As anticipated, when confronted with a gain frame, participants reported more positive attitude towards the brand when it was paired with an other-benefit appeal (M = 4.42) than with a self-benefit appeal (M = 3.46). When confronted with a loss frame, participants reported more positive attitude towards the brand when it was paired with a self-benefit appeal (M = 3.74) than with an other-benefit appeal (M = 3.48). When confronted with a gain frame, participants reported more motivation to process the message when it was paired with an other-benefit appeal (M = 4.58) than with a self-benefit appeal (M = 3.92). When confronted with a loss frame, participants reported more motivation to process the message when it was paired with a self-benefit appeal (M = 4.77) than with an other-benefit appeal (M = 4.15). The 2 (message frame: gain vs. loss) x 2 (beneficial appeal: self vs. other) ANOVA on motivation to process the message showed the expected interaction effect (F(1,78) = 4.53, p < 0.05). The persuasion effect of the message was influenced as hypothesized. A 2 (message frame: gain vs. loss) x 2 (beneficial appeal: self vs. other) ANOVA on persuasion effect of the message showed the expected interaction effect (F(1,78) = 6.68, p < 0.05). As anticipated, when confronted with a gain frame, participants reported more perceived effectiveness of the mes- sage when it was paired with an other-benefit appeal (M = 4.75) than with a self-benefit appeal (M = 4.02). When confronted with a loss frame, participants reported more perceived effectiveness of the message when it was paired with a self-benefit appeal (M = 4.82) than with an otherbenefit appeal (M = 3.86). Buying intention also proved to be significantly influenced by the interaction of message framing and beneficial appeal. The 2 (message frame: gain vs. loss) x 2 (beneficial appeal: self vs. other) ANOVA revealed the expected interaction effect (F(1,78) = 8.70, p < 0.01). As anticipated, when confronted with a gain frame, participants reported more intention to buy the product when it was paired with an other-benefit appeal (M = 5.08) than with a self- benefit appeal (M = 3.38). When confronted with a loss frame, participants reported more intention to buy the product when it was paired with a self-benefit appeal (M = 4.84) than with an other-benefit appeal (M = 4.16). 6. Contribution


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

91

The findings of this research are a contribution to the understanding of advertising effective-ness for green products. The research demonstrates that gain-framed messages are more efficacious when paired with otherbenefit appeals, whereas loss-framed messages are more efficacious when paired with a self-benefit appeal. This is an important, substantive finding which extends the current literature on message framing and advertising effectiveness. The findings stimulate further research as they raise further questions regarding other possible moderators and the underlying mechanism of the study results. The results of the study have practical implications for marketers of green products by providing important opportunities to increase the efficacy of green marketing campaigns.


92

FÖRORD: KAN VI KONSUMERA OSS TILL LYCKA? Kön ringlar nedför gatan så långt ögat når. Nu är det snart dags. Armbågar vässas, pulsen ökar, förväntan i blick. Startskottet går och en hord av människor strömmar framåt med samma mål – att lägga vantarna på en iPad 2. De lyckliga promenerar ut på gatan med ett leende på läpparna och den vita kartongen i handen. Euphoooriaa! Eller? Blir vi verkligen lyckligare av prylar? Vinner den som har mest saker när man dör? Definieras lycka av ägande? På följande sidor visar författarna att så inte behöver vara fallet. Det är nämligen inte ägandet av fysiska ting som gör oss lyckliga, det är själva upplevelsen som konsumtionen resulterar i som leder till lycka. Resultaten från de två studierna som ligger till grund för pappret visar ett tydligt samband mellan upplevelser och lycka, snarare än egendom och lycka. Vad kan detta bero på? En förklaring som påvisas i annan forskning kan vara att positiva upplevelser tenderar att leva länge i våra minnen och få oss att må bra medan man förr eller senare tröttnar på sin en gång så högt värderade pryl. För oss som arbetar med kommunikation så pekar resultaten i en tydlig riktning – vi måste komma underfund med hur de varumärken vi jobbar med kan leverera och kommunicera upplevelser som väcker intresse och adderar värde till konsumenten. Vi måste på djupet förstå konsumenten och vad hon behöver och vill ha. Vi måste reflektera över vad vi verkligen säljer och anpassa budskapet därefter. Vi måste leverera en överlägsen varumärkesupplevelse i varje kontakt med konsumenten. Och vi måste göra det konsekvent över tid. De varumärken som förstår vikten av detta kommer inte bara skapa en differentierande fördel i sin kategori med lojala kunder som följd. De kommer även ha en väldigt lycklig CFO.

jessica morales


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

93

AN EXPERIENTIAL ACCOUNT OF HAPPINESS IN LIFE AND IN ADS J. Joško Brakus Leeds University UK Bernd H. Schmitt Columbia University US Lia Zarantonello IESEG School of Management France

INTRODUCTION Happiness is a theme that is often communicated in ads. For example, a recent ad and event campaign for Coca Cola carries the tagline “Open Happiness” (Coca Cola, 2009). Similar examples include BMW’s “Joy” and Budweiser’s “Good Life” ad campaigns. Lewis and Hill (1998) conducted a content analysis of food ads on British television and concluded that a majority depicted consumers being joyful and happy in real consumption situations. In general, marketers have been increasingly trying to appeal to consumers’ pursuit of happiness assuming that consumers want to be happy and that such communication strategy, therefore, would be persuasive (Aaker et al., 1986). The present paper does not address the communication effectiveness of happiness as ad content. Instead, we focus on the relationship between experience and happiness, two areas of increasing importance in marketing and advertising as well as in psychology (Brakus et al., 2009; Carter and Gilovich, 2010; Peterson et al., 2005). In this theory-building research, we first explore how daily experiences that are evoked during ordinary everyday consumption activities—such as eating and preparing food, shopping, entertaining oneself—contribute to consumers’ happiness. Then we look at how ads can communicate specific experiences to consumers—sensory, affective, bodily and intellectual (Brakus et al., 2009)—and explore to what extent the communicated consumption experiences affect consumers’ perceived happiness when consumers are passive recipients of ad messages rather than active participants in consumption. We conducted two studies to study the relationship between experience and happiness, first in an “active” consumption context (i.e., participants actively participated in consumption activities that evoked specific experiences) and then in a “passive” consumption context (i.e., participants were exposed to print ads that communicated specific experiences). Both studies demonstrate that intellectual experiences contribute to happiness as meaning, that sensory and behavioral (bodily) experiences contribute to happiness as pleasure, and that behavioral (bodily) and intellectual experiences contribute to happiness as engagement. Theoretically, our results indicate that happiness-as-meaning, happiness-as-pleasure, and happiness-as- engagement are not only chronic individual predispositions (Peterson et al., 2005), but also distinct happiness states that individuals go through as a consequence of the specific experiences evoked by consumption episodes. LITERATURE REVIEW Experience is a multidimensional construct. Based on works in philosophy (Dewey, 1922; 925) and cognitive science (Pinker, 1997), Brakus et al. (2009) identified four dimensions of experience: sensory, affective, intellectual, and behavioral. Sensory experience refers to the stimulation of the five senses. Affective experience includes moods and emotions. Intel-


94

lectual experience can include analytical and imaginative thinking. Behavioral experience refers to motor actions and the body’s interactions with the environment. In psychology, there has been strong interest in happiness during the last decade or so (e.g., Seligman and Csikszentmihalyi, 2000). Positive psychologists have distinguished two distinct chronic approaches toward achieving happiness: pleasure (Kahneman et al., 1999) and meaning (Waterman, 1993). The hedonic approach focuses on pleasure and stresses that happiness results from experiencing sensorily and affectively pleasurable moments or episodes. The eudaimonic approach stresses that happiness results from engaging in meaningful activities. That is, whereas the hedonic route concerns the small, pleasurable elements in life, the eudaimonic route to happiness focuses on search for lasting meaning. Recently, Peterson et al. (2005) have added another orientation to happiness and life satisfaction—happiness through engagement. Engagement is similar to a state of flow. Time passes fast. The self seems lost in the activity (Csikszentmihalyi, 1990). Conceptually, the experience construct is closely tied to the concept of happiness. Both constructs are concerned with elements that transcend everyday life. More important, the experience dimensions (sensory, affective, intellectual, and behavioral) seem to map closely with the happiness dimensions (pleasure, meaning, and engagement). Experience seems to be an antecedent of happiness. That is, sensory-affective as well as behavioral (bodily) experiences may contribute to “pleasure happiness”; intellectual experiences may contribute to “meaning happiness”; and behavioral experiences may contribute to “engagement happiness.” Thus, overall, we predict that experience is an important contributor to happiness. Moreover, we will empirically explore the relation between consumption activities and happiness —for example, does eating or shopping make people happier?—but do not offer specific hypotheses. Overall, we predict that experience is an important contributor to happinessas-state in everyday life (i.e., happiness-as-meaning, -as-pleasure, and -as-engagement). We also expect that when consumers are directly engaged in a consumption that evokes specific type of experience, their judgments of the state of happiness that they are going through at that moment will be largely independent of their chronic orientation to happiness (Peterson et al., 2005). In other words, we think that salient experiences, evoked during consumption episodes, are stronger predictors of happiness-in-the-moment then the type of an individual’s chronic predisposition to happiness. Here is the summary of our predictions: H1: We predict that in an active consumption context (when consumers directly participate in consumption): a. the evoked sensory, affective and behavioral experiences contribute to consumers’ happiness-as-pleasure; b. the evoked behavioral experience contribute to consumers’ happinessas-engagement; c. the evoked intellectual and behavioral experiences contribute to consumers’ happiness-as-meaning. H2: The predictions in H1 are independent of consumers’ chronic orientation to happiness.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

95

We do not offer explicit hypotheses related to the relationship between the experiences communicated in ads and the perceived happiness (in a passive consumption context), but we will explore that relationship. We want to see to what extent the relationship between experiences and happiness is consistent across the two contexts. Next, we present the two studies in which we test our predictions. STUDY 1 Method To understand the relationship between everyday consumption experiences and the perceived happiness and life quality, we conducted a diary study with consumers. Such method permits the examination of reported events in their natural setting (e.g., Bolger et al., 2003), which was especially important considering our objective of investigating consumers’ “in-the- moment” happiness states in everyday life. Respondents had to complete the diary every day, for one week. We asked the respondents the same set of questions every day. First, we asked them to report the “level of activity” by stating how much they engaged in each of the following activities during the day: “eating or preparing food”, “entertaining yourself”, “engaging in physical activities”, “grooming and dressing”, and “shopping” (from 1 = “very little” to 7 = “very much”). Second, we asked how much each of the five activities stimulated a specific experience, using a similar scale (“level of experience”). Third, we asked consumers to provide open-ended descriptions of how each of the activities that they did stimulated a specific experience. Lastly, the respondents had to complete a nine-item scale on happiness, adapted from Peterson et al. (2005), that captured the three dimensions of happiness-as-state and a four-item scale on life quality adapted from Diener et al. (1985). (Note that the our objective was not develop a new scale, but the scale that we adapted from Peterson et al. (2005) to capture the happiness states satisfies the usual reliability criteria: the three dimensions load on three separate factors and the Cronbach’s alphas for each of the three dimensions are > .7.) Finally, before they started keeping the diary, respondents also responded to Peterson et al. original orientationto-happiness scale to assess their chronic orientation to happiness. To make sure that the respondents addressed all the types of experiences, we randomly divided the sample into four groups and asked the respondents in each group to focus either on sensory or affective or intellectual or behavioral (bodily) experiences. Therefore, we prepared four types of diaries that we distributed between subjects. The diaries were given to international consumers (n = 163) living in Paris. Results and Discussion First, we examined the relations between consumption activities and experiences by estimating a series of regressions. Since we instructed each respondent to focus only on one type of experience, we estimated these regressions within each diary/experience type (see above). The results are presented in Table 1. As Table 1 shows, all consumption activities can evoke specific experiences. If we consider specific types of experience, however, some differences emerge judging by the relative magnitude of the standardized OLS coefficients (refer to the coefficients in bold in Table 1). Specifically, “eating and preparing food” evokes especially sensory experience; “entertaining yourself” evokes most strongly sensory and behavioral (bodily) experiences; “engaging in physical activities” evokes affective, intellectual (i.e. imaginative), and behavioral (bodily) experiences; “grooming and dressing” result in affective and intellectual (imaginative) experiences; finally, “shopping” evokes mostly affective experiences.


96

Next, we investigated the relations between the specific types of experience and the dimensions of happiness. The “level of experience” was the independent variable this time and the dependent variables were the dimensions of happiness. Results are reported in Table 2. As Table 2 indicates, pleasure is a consequence of the evoked affective and behavioral (bodily) experiences. Sensory and intellectual (imaginative) experiences have a positive and significant effect as well, but weaker. Engagement results especially from behavioral (bodily) experiences. Finally, consumers seem to find behavioral experiences and, not surprisingly, intellectual (imaginative) experiences especially meaningful. These results support H1. Respondent’s chronic orientation to happiness does not affect or moderate these results. Therefore, and consistent with H2, experiences evoked by specific consumption episodes are more predictive of individual’s actual happiness state than is individual’s chronic predisposition for specific types of happiness. Most important, qualitative diary data corroborate all of these findings (and are skipped here for brevity, but are available from authors). STUDY 2 In Study 2 we extended the results of the diary study to explore whether there is a difference in the relationship between the communicated experience and the resulting perceived happiness when consumers are not actively engaged in consumption (as they were in the diary study). Specifically, Study 2 investigates how advertisements that employ experience evoking imagery contributed to the perceived happiness. Method We designed 18 mock-up print ads. These mock-up ads promoted six fictitious brands in various product and service categories: yoga, cooking, learning, health and beauty spas, vacationing, and personal music players. We did not specifically control for the communicated type of experience, but we had chosen those six categories because we expected that the specific imagery, in combination with the specific product or service category, would communicate the four types of experiences. Therefore, we asked the respondents to rate these ads on the experiential dimensions—sensory, affective, intellectual, and behavioral—using the items adapted from the brand experience scale (Brakus et al., 2009). The respondents also rated the ads on the three happiness dimensions again using the scale adapted from Peterson et al. (2005) (e.g., “I am rarely distracted by what is going on around me when I engage in the activity shown in the ad” for engagement, “I find the activity shown in the ad to be pleasurable” for pleasure, and “Being engaged in the activity shown in the ad reminds me that my life has a lasting meaning” for meaning). Our respondents (n = 82) were adult consumers (employees of a leading global investment bank situated in London). We randomly divided the 18 ads into six three-ad groups. Each respondent rated the ads in one group only (three ads in total) to minimize the mental burden on the respondents and possible boredom. Results and Discussion We estimated three OLS models to examine how specific experiences affect the perceived happiness, but this time the respondents did not actively participate in experience; the experience was communicated to them through ads. In each regression, the dependent variable was one of the happiness dimensions (pleasure, engagement, meaning) and the predictors were the four types of experience (rated on 7-point scales; see Brakus et al., 2009). The results (see Table 3) suggest that behavioral (i.e., bodily) and sensory experiential ad imagery is particularly well suited to communicate pleasure (p < .05 and p < .001 respectively), which is consistent with Study 1. In


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

97

contrast to the results of Study 1, visual imagery that evokes affective or intellectual experience is not a significant predictor of the perceived pleasure. Behavioral experience, which is communicated through imagery that depicts interactions between body and environment, increases the perceived engagement (p < .02), which is consistent with the results of Study 1, but so does intellectual (imaginative) experience (p < .001). Finally, intellectual experiential imagery is best suited to communicate meaning (p < .001). However, consumers do not seem to derive meaning from behavioral (bodily) experiences when they are passive recipients of advertising messages. Recall that in Study 1 behavioral (bodily) experiences did contribute to meaning, but in that case consumers were actively engaged in consumption. CONCLUSION The results across the two studies are largely consistent. When consumers actively participate in consumption activities, they derive pleasure from affective and behavioral (bodily) experiences. Moreover, they are engaged when they experience a consumption activity with their bodies, and they derive meaning from intellectual and behavioral experiences. Ads that employ experiential imagery can contribute to the perceived happiness. In that case, the message needs to evoke a sensory experience (i.e., visual in the case of print ads) in order to increase the perceived pleasure and intellectual experience (e.g., imaginative thinking or daydreaming) in order to increase the perceived engagement. On the other hand, ads do not have to evoke behavioral (bodily) experiences in order for consumers to derive meaning from ads. Behavioral (bodily) experiences are essential part of meaning when consumers are actively engaged in consumption. Our results confirm the notion that happiness is not necessarily tied to consumption and possession of material goods (e.g. Kahneman et al., 1999). It is possible that individuals reject the idea of â&#x20AC;&#x153;happiness as possessionsâ&#x20AC;? when they reflect back upon their lives or they are asked to make long-term predictions about their happiness in the future. On the other hand, our results also indicate that individuals can derive happiness from mundane, ordinary everyday consumption activities, but what really leads to happiness in this case is the nature of the experience evoked by consumption. Not only that the experience construct is central to our understanding of happiness, but it also seems to be central to our understanding of communication effects of commercial messages that use happiness as a persuasion tool. From a practitionerâ&#x20AC;&#x2122;s point of view, knowing how to create and how to communicate experiences to make customers happy is, in our view, the next step of experience management (Pine and Gilmour, 1999; Schmitt, 1999). We believe that such knowledge can also provide a distinct competitive advantage for companies. Given the attention that the issues of experience and happiness have received and given that we are surrounded by commercial offers, it is important to understand the link between experience and happiness in a commercial marketing communication context.


98

förord: Kan man bygga förtroende för ett varumärke genom annonsering? På 1990-talet skedde ett fundamentalt skifte från transaktionsmarknadsföring (sälja en grej i taget) till relationsmarknadsföring (bygga en långsiktig relation mellan varumärke och konsument). I detta skifte visade sig förtroendet vara en nyckelfaktor. Chaudhuri and Holbrook (2001) undersökte 107 varumärken och fann att förtroendet för varumärket hade en stark påverkan på varumärkeslojalitet, både emotionellt och beteendemässigt, vilket ledde till ökade marknadsandelar och bättre möjligheter till prispremier. De två varumärken i Sverige som haft störst förtroende genom de senaste åren, bland annat i studier av Nordic Brand Academy, har också varit de mest framgångsrika oavsett om man mäter tillväxt, marknadsandelar eller lönsamhet. Om svaret är att förtroendet är en nyckelfaktor, så är frågan hur man etablerar förtroende för sitt varumärke i tider när konsumenterna blir allt mer kunniga och cyniska. Forskningen visade, föga förvånande, att varumärkesupplevelsen var den viktigaste faktorn. Och att förtroendet påverkas i positiv eller negativ riktning beroende på hur annonseringen genomförs. Denna studie fokuserar på vilka ingredienser som behövs för att porträttera ett pålitligt varumärke och att försöka ta reda på varför konsumenter har förtroende för vissa varumärken. Och inte andra. Och det tydligaste budskapet är var varumärkesägaren bör undvika för att inte riskera att urholka förtroendet för sitt varumärke genom illa genomtänkt annonsering.

Peter granström


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

99

TRUST ME I’M AN ADVERT! HOW TO CREATE A TRUSTING BRAND IDENTITY THROUGH ADVERTISING Dr. Kathleen Mortimer University of Northampton UK Dr. Annie Danbury University of Bedfordshire UK

PURPOSE OF STUDY The recognition of the importance of trust in Marketing was a result of the fundamental shift that took place in the early 1990’s from a transactional Marketing approach to Relationship Marketing. A seminal paper by Morgan and Hunt (1994) demonstrated the importance of commitment and trust in building a relationship with customers. They define trust as “when one party has confidence in an exchange partner’s reliability and integrity”. This definition is based on trusting either a person or an organisation and incorporates the idea that trust is not only linked to consistency (i.e. reliability) but also that there is a belief that the person or organisation would act in the appropriate way, i.e. in the other parties interest, if necessary. The reliance element could be considered a rational element of the construct while the integrity is more affective (Blois, 1999). Since the Morgan and Hunt (1994) paper, the topic of trust has been explored in some depth, with 35 papers being written on the subject in 2006 alone (Kenning, 2007). The discussion of trust has evolved from the business to business literature to more general marketing journals with the slow recognition that trust is also important in the business to consumer markets. This is obviously rather different as the trust is now being linked to an inanimate object. However it is recognised that customers can perceive brands as having personalities (Aaker, 1996) and can feel some loyalty towards them. As a result, brands are being personified and therefore it is possible to trust them. Indeed, these constructs may be linked. Chaudhuri and Holbrook (2001) examined 107 brands and found that brand trust had a strong impact on brand loyalty in both its forms i.e. attitudinal loyalty and purchase loyalty which then led to increases in market share and premium pricing. Delgado-Ballester and MunueraAleman (2004, 2005) also found a positive relationship between brand trust, brand loyalty and brand equity. So it is clear that brand trust leads to many positive outcomes for the brand owner. It is therefore important to establish how it can be created. Particularly as there is evidence to suggest that customers are being increasingly cynical and levels of brand trust are decreasing (Lantieri and Chiagouris, 2009). There is surprisingly very little work in this area. A study by Xingyuan, Li and Wei (2010) examined the impact of three sources of information on brand trust; user experience, word of mouth (WOM) and advertising. They proposed a strong relationship between brand knowledge and brand trust and thereby examined how these different sources of information contributed to brand knowledge. Perhaps, unsurprisingly, they found that user experience had the strongest influence. Advertising was good at increasing brand awareness, product knowledge and company knowledge but less effective at influencing perceived value. The important influence of consumption experience was also supported by Delgado-Ballester and Munuera-Aleman (2005).


100

Research on the link between advertising and trust specifically consists of two main threads. One line of research is people’s perceptions of advertising generally and how trustworthy they perceive it to be. This is important because if people don’t trust advertising then they may be less influenced by individual brand attempts to build trust through that marketing communications tool. Indeed this negative perception of all advertising was identified by Li and Miniard (2006) as a challenge. In 2009 Soh, Reid and King developed the ADTRUST scale, the first scale to measure Trust in Advertising so that this area could be explored further. The scale identifies four main factors, reliability, usefulness, affect and willingness to rely on. It has not yet been adopted in other studies but may be a useful tool is future research. The other line of research examines how advertisements can be designed to create a trusting image for the brand. One well-researched tool employed to create this trusting image is the use of endorsers. It has been stated that 25% of all advertisements in the UK and USA use celebrity endorsers and this increases to 75% in Japan. Studies show that the credibility of an endorser, which should contain the three dimensions of expertise, trustworthiness and attractiveness, can have a direct impact on attitude to the advertisement and the perceived credibility of the brand through association (Lafferty et al 2002; Spry et al 2011). However a study that compared corporate credibility and endorser credibility found that endorser credibility has a stronger impact on attitude to the advertisement whereas corporate credibility had more impact on brand attitude (Goldsmith et al, 2000). The importance of the company image as well as the brand image is important to highlight here, particularly when looking at services where the corporate brand and product brand can be the same. Garretson and Niedrich (2004) specifically examined the role of fictitious spokes-characters in advertising, such as Mitchelin Man, and how they are perceived by consumers as trusting. Their results show that trust is linked to the perceived expertise of the character and the level of nostalgia that it brings and this has a positive effect on brand attitude. Li and Miniard (2006) examined how to create a trustworthy image of a brand through advertising by undertaking two advertising experiments using a fictitious car repair company, without the use of celebrity endorsers. It was found that the inclusion of the words “You can trust us to do the job for you” had an impact on the number of positive thoughts about the advertisement and the brand. The brand was perceived to be more trustworthy compared to the advertisement without those words and consequently purchase intention increased. Li and Miniard (2006:111) acknowledge that their study “only scratched the surface” and that much more work needed to be done in this area. Their findings were based on an experiment using a fictitious car repair company which is a high involvement service and consequently a high level of information processing would normally results due to high risk factors. Interestingly, a current radio advertisement for Country Vehicles, a car repair company in the UK, also adopts a similar slogan: “A service you can trust”. Our exploratory study follows on from the work of Li and Miniard (2006) by examining high involvement services further and is designed to answer the following research questions: » What ingredients do advertisements portraying a trusting brand image have that others don’t? » What makes people trust specific brands? At a more macro level, all these studies on advertising and trust have been undertaken in the USA, with the exception of the work of Delgado-Ballester and Munuera-Aleman. It is therefore worthwhile to undertake research in


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

101

Europe that reflects more closely our practitioners’ efforts due to differences in the advertising industry. METHODOLOGY In order to explore the subject further, three focus groups were conducted with Marketing 1st and 2nd year undergraduate students, containing a total of 38 student perceptions. They were shown nine advertisements which appeared in the Metro free London newspaper on Wednesday 15th February 2012. It was possible to group these advertisements into three relevant product categories; mobile telephone providers, financial services and holidays. It was not felt necessary for the participants to be familiar with all the brands as it was their response to the advertisements which was initially of interest to us. Each advertisement was presented to them on slides and they were given one minute to write down whether they trusted the advertisements and why or why not. They were then shown the three advertisements together and asked to compare them in terms of portraying a trusting image. Lastly they were required to step back and describe the brands they trusted within that category more generally. This process was repeated for all three product categories. When all advertisements had been examined the participants were asked to indicate a brand that they trusted the most and why. Finally, participants were asked how they felt about advertisements generally in terms of trust. Once the form had been completed there was a general discussion with the group about brand trust to pick up any issues that the form may not have captured. FINDINGS AND DISCUSSION The first three tables (see below) present the responses that were received from asking the question: “Do you trust this ad: why/why not?” Selections of both positive and negative comments have been included for each advertisement to represent the main issues that were raised. The highlighted brand in each table indicates the advertisement that was perceived to be most trustworthy out of the three. This decision was unanimous across all three focus groups. The examination of comments in the three categories revealed some interesting findings. Firstly, it would seem that an advertisement needs to be very simple and straightforward in its communications in order to be perceived as honest and trusting. When advertisements suggest that something is free or a good deal people become suspicious and are looking for the catch. All information needs to be presented clearly and be straight to the point. The amount of information that the advertisement should contain is difficult to define as advertisements were criticised for having too little as well as too much information. Certainly a large amount of small print is perceived negatively and needs to be handled with care. It is good to include prices, but they need to be precise and understandable. To present prices from is perceived as rather misleading. Visuals are generally regarded as helpful in creating a more trusting image. Seeing happy customers helps to communicate some of the main points and adds a more emotive message to accompany the facts and figures. Evidence of awards also has a positive impact as well as links to social media sites. However some advertisements were perceived as looking cheap which then influenced the participants’ perception of the brand. The tangibilisation of services through providing information in advertising is discussed in the services advertising literature but there seems a lack of evidence of it being adopted by practitioners (Mortimer, 2000). One finding that became clear during the focus groups was the strong link between previous brand knowledge and perception of the advertisements. As illustrated in the work of Xingyuan, Li and Wei (2010), user experience


102

has a significant impact on how the brand is perceived. Our study shows that it also has a strong impact on perceptions of brand communications, even if the brand in question is being promoted in a product category where personal experience has not taken place e.g. Tesco and Virgin. Such brand trust is obviously invaluable and contributes greatly to brand equity. It is why brands like Tesco and Virgin can successfully extend their portfolio into new product categories such as financial services. It is interesting to note the long list of “trusted brands” that were provided by the participants for each product category investigated. Again, the main reason given for each choice was user experience. This came through particularly strongly with financial services where a number of people talked about the brand being used by the family and the reference to “my bank”. This would suggest that it takes time to build a trustworthy identity. Lopamudra and Subhadip (2011) suggests that perceived credibility is a result of the company being perceived as consistently competent and honest. The length of the lists would suggest that the positioning of “trust” does not seem to be occupied by any particular brand in these product categories at the present time which may indicate an opportunity. Table 4 provides a list of the most trusted brands from our study. It is worth noting that they cover a wide range of different products and services as well as examples of high and low involvement purchases. This is despite evidence to suggest that trust is more important for high involvement product categories (Lantieri and Chiagouris, 2009). However, the reasons provided for trusting a brand were less varied. Consistent quality came through very strongly, based on reputation and personal experience. The size of the company, how long they have been established and their level of customer service were also highlighted. These are obviously all characteristics that can be communicated through advertising and will be explored in further research. Lastly the participants were asked whether they trusted advertising. Their perception of advertising was influenced by the type of product being advertised and the advertisement itself. As one participant put it “I only trust ads for trustworthy brands”. If a customer has a positive opinion of a brand then that influences their perception of that brand’s advertising. A number of participants used make-up and financial services as examples of products whose advertisements would be the least trustworthy. The fact that models were retouched and false eyelashes were used to promote mascara was cited as examples of how advertising tries to fool people. Indeed, an undercurrent prevailed in the responses which suggested a lack of trust and a feeling of being fooled e.g. “they rarely tell you the entire truth” and “don’t trust most of them – strings attached”. These findings support other evidence indicating that trust in brands is decreasing at a macro level (Lantieri and Chiagouris, 2009). Customers are becoming more cynical about brands and their claims of being the best. They see that the emphasis on short term goals within organisations has led to quality inconsistencies, product recalls, and poor service encounters resulting in not wishing to have a “relationship” with all the brands that they buy. CONCLUSION The purpose of this exploratory study was to examine advertisements for high involvement services and identify the ingredients needed to portray a trusting brand image. A more general objective was to explore why people trust certain brands. The impact of user experience on brand trust comes through very strongly and brand owners need to ensure that their brands are consistently meeting consumers’ expectations to encourage loyalty and positive WOM. However, the results from this initial research do indicate that advertising can assist in creating and maintaining a trusting image. It


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

103

would seem that consumers value advertising that is clear, simple and straightforward. They are generally wary of being tricked in some way and cynical when it comes to deals, offers and small print. Tools that advertising practitioners can utilise to address this seems to be more emotive messages, possibly through visuals, evidence of awards, links to social media sites and indications of the size of organisation and how long they have been in operation. More research is obviously necessary to explore these findings further as they are based on high involvement services and on a specific group of people, both limitations preventing generalisations to be made. Meanwhile perhaps advertisers should be more critical of their own efforts in attempting to build relationships with consumers based on some of the poor practices that have been identified in this study. It is a sobering thought that young people are so critical of the advertising they are exposed to even for brands that they buy.


104

FÖRORD: Att engagera sig i något större blir allt viktigare för företagen Att ”göra gott” har länge varit en viktig uppgift för företag. Syftet är naturligtvis att belysa ett större åtagande som sträcker sig utöver, kategorin, verksamheten och de produkter man säljer. Corporate Social Responsibility (CSR) har en positiv inverkan på konsumentbeteendet, det visste vi redan. Men vad vi har saknat är ett systematiskt angreppssätt för att förstå samspelet mellan CSR och konsumentens attityd till företaget och deras produkter. Historiskt har man inte undersökt huruvida varumärkets styrka har betydelse för konsumentens uppfattning när ett företag kommunicerar sitt CSR åtagande. CSR kan ses som ett sätt för företag att bidra till samhället. Forskningen visar också att konsumenten är mer kritisk innan de bildat sig en egen uppfattning om företagets intentioner. Motiven till varför företag väljer att kommunicera sitt CSR åtagande varierar. Det kan vara egoistiskt-, kommersiellt-, värderings-, strategiskt- eller intressentdrivet. Konsumentens uppfattning om företagens motiv kan därmed också påverka attityden till företagen. Resultatet av undersökningen är spännande och ger oss insikter i just konsumentens värderingsprocesser då det gäller CSR. Intressant är också att varumärkets initiala styrka spelar en central roll i hur konsumenten värderar företagets CSR åtagande- fast det kände vi redan till - för konsumenten tror på starka varumärken...

kim saxberg


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

105

THE INFLUENCE OF CORPORATE SOCIAL RESPONSIBILITY ON CONSUMER PRODUCT RESPONSES Christian Boris Brunner Newcastle Business School UK Xiaoming Lu Newcastle Business School UK

INTRODUCTION Corporate Social Responsibility (CSR) could be defined as a company’s activities and status related to its perceived societal or stakeholder obligations which may involve engaging in socially responsible employment and manufacturing practices, environmental protection, and socially responsible business practices or activities (Du et al., 2010; Lee et al., 2011). The attention to CSR in both, academia and practice has grown within the last decade (e.g. Lin et al., 2011). While most firms engage in CSR activities (Hess et al., 2002; Luo and Bhattacharya, 2009; McKinsey & Company, 2006), there is a large and growing number of publications, trying to find out, how companies could develop the right strategy in doing ‘good’ e.g. (Locket et al., 2006; Peloza and Shang, 2011; Vaaland et al., 2008). But even though there is still a wide array of studies regarding several CSR factors (see for an overview e.g. Peloza and Shang, 2010), the different studies are detached from each other mostly. PURPOSE OF THE STUDY In the literature it has been demonstrated that companies’ CSR activities have a positive influence on company’s financial performance (Barnett, 2007; Margolis and Walsh, 2003), brand’s image and loyalty (Basil and Herr, 2006; Bhattacharya and Sen, 2004; Du et al., 2007; 2010). Peloza and Shang (2010) carried out a literature review on CSR and collected 177 different studies of which 164 used quantitative methods. Most of those studies executed experimental designs, concentrating on a few success factors in CSR, but are mainly isolated from each other. Consequently, marketer’s knowledge of consumer’s responses to specific CSR activities is fairly limited (Peloza and Shang, 2010). The purpose of this paper is to develop a CSR model that demonstrates how perceived companies’ motives finally lead to the attitude towards the company and its products when communicating CSR to consumers. Many studies have only focused on the direct relationship between a CSR activity and company’s attitude (e.g. Groza et al., 2011; Nan and Heo, 2007). We presume that there are intermediate constructs present. Within the marketing literature it is often mentioned that the credibility of a message is very important, especially if consumers are confronted with advertising to persuade them (e.g. Eisend and Küster, 2011). As CSR in general is seen as a company’s contribution to society, we assume that consumers may be even more critically, asking about company’s motives before forming an attitude towards the company (e.g. La Ferle et al., 2011; Lin et al., 2011; Wagner et al., 2009). Furthermore, we propose that the fit between company and CSR activity and the perceived company’s motives as well as the consumer’s relation to the CSR activity have an impact on


106

the evaluation and the credibility of the CSR activity. These variables influence the evaluation of the company. Hence, we assume that the outcome of the company’s evaluation in consumer’s persuasion process via CSR advertising is depending on a complex interplay between different constructs in which credibility plays a key factor in consumer’s elaboration of CSR advertisements. To the best of our knowledge this relationship has been neglected in most studies, except a view studies (e.g. Alcañiz et al., 2011; Groza et al., 2011; Lin et al., 2011). Therefore, we aim to take a closer look into consumer’s evaluation process between perceived companies motives and company’s evaluation. THEORETICAL BACKGROUND AND HYPOTHESES Theoretical Underpinnings According to attribution theory (Jones and Davis, 1965; Kelley, 1972), human beings reason why others behave in a certain way and what factors are a causal role acting so. Assigning this idea to the context of CSR, consumers attribute company’s motivations for its engagement in CSR (e.g. Becker-Olsen et al., 2006; Du et al., 2007; Ellen et al., 2006; Groza et al., 2011; Sen et al., 2006, Walker et al., 2010; Yoon et al., 2006). Regarding the type of compa- ny’s motives, Vlachos and colleagues (2009) differentiate between egoistic-driven, values-driven, strategic-driven and stakeholder driven motives. In some studies it has been demonstrated that consumers attribute company’s motives as both altruistic and egoistic (e.g. Ellen et al., 2006; Rifon et al., 2004). Alcañiz et al. (2010) argue that perceived company’s motives such as altruistic attribution have an impact on company’s credibility. Similar findings are also reported regarding the relationship between sponsor attribution and sponsor credibility (e.g. Rifun et al., 2004). Furthermore, we include in our considerations the attitude towards the CSR activity itself and assume that the perceived company’s motives may also have an influence, how consumer’s evaluate the CSR activity. H1: Consumer’s perceived companies’ motives are positively related to the (a) evaluation of CSR activity and (b) credibility of CSR activity. Regarding to Social Identity Theory (Tajfel and Turner, 1986) salient group classifications enables people to order their social environment and locate themselves and other individuals within it (Turner, 1985). Here, selfdefinitions are important for individuals, because they help to situate others in the context and suggest what people should do, think, and feel (Ash- forth, 1998; Marin and Ruiz, 2006). Social identification means that individuals perceive themselves as belonging to a group, resulting in a high identification with this group (Tajfel and Turner, 1986). Assigning this idea on the context on CSR (e.g. Bhattacharya and Sen, 2003; Bhattacharya et al., 1995; Lii and Lee, 2012; Marin and Ruiz, 2006), we assume that if consumers identify with a CSR activity and (therefore) keep a strong relation to a CSR activity; this should also lead to a higher evaluation of the CSR activity itself. H2: Consumer’s relation to the CSR activity is positively related to the evaluation of CSR activity. It has been widely reported that the perceived fit between the company and a CSR activity as one of the main antecedents influences consumer behaviour (e.g. Alcañiz et al., 2010; Menon and Kahn, 2003). Whereas in many studies it has been demonstrated that there is a positive relationship between this factor and consumer responses (e.g. Becker-Olsen et al., 2006; Ellen et al., 2006; Pracejus and Olsen, 2004; Rifon et al., 2004), there are also results in which no significant differences between high and low fit have been found (e.g. Hamlin and Wilson,2004). These different


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

107

results are caused by the fact of missing intermediated constructs, e.g. consumer’s relation to CSR, consumer’s processing depth and perceived company’s motives, which have been mostly neglected: As Becker-Olsen and colleagues (2006) demonstrated, a low fit between company and CSR activity negatively impacts consumer attitudes (independent from the perceived firm’s motivation). Similar results occurred in the case of a high fit that are profit-motivated (Becker-Olsen et al., 2006). Hence, there should be an interaction between fit and perceived company’s motives. Menon and Kahn (2003) found out that consumers evaluated a company better in case of cause promotions compared to advocacy adver- tising. They explain these results with the fact that consumers could not figure out ad hoc a direct relationship between cause and company in case of advocacy advertising, which trigger them to think about the fit. As Friestad and Writh (1994) point out, the more consumers think about persuasion tactics, the less likely this results in a favourable evaluation. Consequently, the higher elaboration process led to a decreased attitude towards the company (Menon and Kahn, 2003). Assessing these results, we assume that depending on the fit between company and CSR activity, consumers may be motivated to think about the company’s motivation to engage in CSR. Hence, we expect that there is a direct impact of fit on perceived companies’ motives, which is consistent with findings from Rifon and colleagues (2004) as well as Barone and colleagues (2007). Moreover, we assume that the fit also influences the credibility of the CSR activity indirectly through the perceived companies’ motives. H3: The fit between the company and CSR activity is positively related to (a) consumer’s perceived companies’ motives and (b) the credibility of CSR activity. It has been researched that the influence of corporate credibility has a positive effect on the attitude towards brand (Goldsmith et al., 2000; Lafferty et al., 2002; Rifon et al., 2004). Hence, we assume that the consumer’s perceived company’s credibility to engage in the CSR activity has a positive impact on the attitude towards the company. Furthermore, we expect the evaluation of CSR activity to have a positive effect on the evaluation of company (Luo and Bhattacharya, 2006; H4: The (a) evaluation of CSR activity and (b) credibility of CSR activity is positively related to the evaluation of the company. Brown and Dacin (1997) demonstrated that “what consumers know about a company can influence their reactions to the company’s products” (Brown and Dacin, 1997, p. 79). Even though they observed positive and negative relationships between corporate evaluations and a (new) product evaluation in their studies, other researchers demonstrated positive effects between a company or brand and its products (e.g. Berens et al., 2005; Biehal and Sheinin, 2007; Goldberg and Hartwick, 1990; Sheinin and Biehal, 1999; Wansink, 1989). Hence, we adopt a positive relationship between the evaluation of the company and the evaluation of the product. H5: The evaluation of the company is positively related to the evaluation of the product. Within the literature to brand extensions, it is often mentioned that consumer’s perceived similarity between brand and new product is a key driver when consumers form an evaluation of the fit (e.g. Aaker and Keller, 1990; Czellar, 2003). The higher the number of salient attributes between brand and new product, the higher the perceived similarity (Broniarcyk and Alba, 1994). In the case of a weak brand, consumers keep only a view associations to the brand and the brand awareness as well as the brand image are low (Keller, 1993). However, to be able to constitute an evalu-


108

ation of fit between brand and new product, consumers need to have enough brand knowledge. Consequently, it should be easier for consumers to form an evaluation of the fit, if brand strength is high. Assuming this idea on the context of CSR, we adopt that consumers need to have specific knowledge about the company to form an appropriate evaluation of fit between company and CSR activity. Therefore, we assume that in the case of weak brands, the influence of fit between company and CSR activity on the credibility of CSR activity as well as on perceived consumer’s motives will be lower compared to strong brands. Furthermore, we presume that because of less company brand knowledge, consumers will reason less about the company’s motives. This effect should lead to a reduced impact of perceived company’s motives on the credibility of CSR activity in the case of weak brands compared to strong brands, too. In addition to this, because of the mere exposure effect (Zai- chowski, 1985), consumers are more familiar with strong brands. As Erdem and colleagues (1998, 2006) point out, strong brands signal product positions credibly. Hence, uncertain product attributes could be reduced when signalling a strong brand (Erdem et al., 1998; 2006). Therefore we conclude the assumption that in case of a strong company the influence of fit between company and CSR activity on both, perceived companies motives and on credibility of CSR activity (direct and indirect) is higher than in case of a weak company. H6: In case of (a) strong, i.e. (b) weak company brands, perceived fit between company and CSR activity is (a) positively, i.e. (b) negatively related directly to the credibility of CSR activity and indirectly (a) positively, i.e. (b) negatively through perceived companies’ motives. METHODOLOGY Experimental Design and Data Collection The data for the study were collected online at a German university and via online social networks within three week from December 2010 to January 2011. The sample size in this research setting also consists of two simulations (e.g. well-known brands via not well-known brands). Of received 800 usable questionnaires (71% female), 400 responses were for both different well-known brands and not well-known brands, respectively. Automotive sector was chosen as the focal area of this study due to its wider use for studying durable goods in the context of CSR (e.g. Becker-Olsen et al., 2006; Irwin et al., 2009). Measures Scales from prior research provided measurement sources for the seven constructs of the hypothesised model using 7-Likert scales. “Evaluation of the Company” and “Evaluation of the Product” were each measured with three items adapted from Gardner et al. (1985). “Anticipated Company’s Motives” was measured with 4 items adapted from Ellen et al. (2006). Measures about “Credibility of CSR Activity” and “Fit between company and CSR-Activity” were adapted from Tax et al. (1998) and Menon and Kahn (2003), respectively. To measure “Evaluation of CSR Activity” and “consumers’ relation to CSR activity”, we used Nan and Heo (2007)’s and Bhattacharya et al. (2004)’s three-item scale. Appendix 2 provided the measures employed in this study along with the scale reliabilities. The scales used in this study showed some acceptable reliabilities (Nunnally and Bernstein, 1994). Confirmatory Factor Analysis The data were subjected to a confirmatory factor analysis using the AMOS 18 structural equation modelling software. The confirmatory factor analysis was completed with maximum likelihood estimation. In both well-known brands and not well-known brands setting, the measurement models offered an acceptable fit to the data (χ²=331; df,188; p<.001; AGFI=.91; CFI=.94; RMEA=.004) and (χ²=257; df,188; p<.001; AGFI=.93; CFI=.97;


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

109

RMEA=.003), respectively. As we report in Appendix 3, the factor loadings provided some evidence for convergent validity (Bagozzi and Yi, 1988). Support for convergent validity is also demonstrated by the average variance extracted (AVE) for all seven constructs (Bagozzi and Yi 1988). Discriminant validity was assessed by calculating AVE for all pairs of constructs and comparing this value to the squared correlation between the two constructs of interest (Fornell and Larcker 1981). Results Model 1: The first model (in Appendix 4) included 400 responses, which was based on the evaluation of well-known brands. As strong brand could reduce consumer’s perceived risk regarding uncertain product attributes (Erdem et al., 1998, 2006), and consumer’s brand trust is high (DelegadoBallester et al., 2003; Doney and Cannon, 1997). For a company with a well-known brand, consumers’ perceived company’s motives did not have any influence on their evaluation of CSR activities. In contrast, consumers’ evaluation of the company was only mediated by credibility of CSR activity from two sources (perceived company’s motives and the fit between company and CSR activity). Model 2: The second model (in Appendix 4) included 400 responses, which was based on the evaluation of weak company brands. (χ²=304; df, 201; p<.001; AGFI=.92; CFI=.95; RMEA=.004). Although the overall fit to the data in the second model is identical to that of the first model, path analysis revealed very different picture. There were two non-significant paths (perceived company’s motives -> evaluation of CSR activity and perceived company’s motives -> credibility of CSR activity). Other paths were significant. The result indicated that consumers’ evaluation towards a company with a weak company brand was not initially developed based on their perceived company’s motives. Instead, consumers’ relation to CSR activity and the fit between company and CSR activity had relatively strong influence on consumers’ perception that were mediated by evaluation of CSR activity and credibility of CSR activities. DISCUSSION, FURTHER RESEARCH AND MANAGERIAL IMPLICATIONS Discussion and Further Research Our results provide an in-depth insight in consumer’s evaluation process of companies communicating CSR. In general, the credibility as well as the evaluation of CSR activity have a significant influence on the evaluation of the company. However, the strength of the company brand plays a central role to influence this interplay: Brand strength seems to be responsible for reduced impact of evaluation of CSR on company’s evaluation. Beyond that, we could observe a direct and indirect effect of fit between company and CSR activity of CSR credibility. Besides consumer’s relation to the CSR activity plays a limited role, because consumers believe in strong brands. However, in case of not well-known brands the fit still has a significant influence on credibility of CSR. How do consumers compose and evaluate the fit, if there is little brand knowledge? As there are several types of fit (Berger et al., 2004), future research should investigate how this process is developed in consumer’s minds. Furthermore, it will be interesting for future research to examine the factors such as the source of information or the quality of products and to see their impact on consumers’ perceptions and decision making. Managerial Implications Marketing managers who communicate CSR advertising to consumers should be aware of the interrelationships of the constructs in our proposed model and especially consider the strength of their company brand in the consumer’s minds. A well-known company like Nestlé should focus even more on the fit to the selected CSR activity, while not well-known company brands should also take into account consumer’s relationship to the CSR issues.


110


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

reception of Advertising

111


112

FÖRORD: Varning, programmet innehåller reklam Vi människor är generellt skeptiska till reklam, det gäller delvis även sponsorerat innehåll. Visst kan vi erkänna för oss själva och andra att vi gillar en viss typ av reklam, men i det stora hela är vi reklamtrötta. Reklam handlar lite förenklat om att övertala och övertyga, och det är inget som vi inte kan leva utan. Faktum är att vi människor inte gillar att känna oss övertalade och det i sin tur kan leda till en negativ attityd till ett sponsorerat innehåll. När reklamen ”bäddas in” i icke kommersiella sammanhang så blir det svårare för tittaren att särskilja redaktionellt innehåll från kommersiellt, vilket minskar tydligheten för annonsören. Att informera tittaren om sponsorskap är obligatoriskt i EU. Samma regelverk håller på att utvecklas för USA. Den här studien handlar om effekten av ett tydligt sponsorskap och hur det påverkar människors sätt att hantera annonsörens övertalningsförsök samt hur vi reagerar på varumärket som avsändare. Studien ger dessutom insikter i hur sponsorskap påverkar övertygelseprocessen, men även mottagareffekten, beroende på antal sekunder som varumärket exponeras. Experimentet som genomfördes i studien har enbart fokuserat på en typ av sponsorskap, ett varumärke, samt en typ av exponering. Det kan tyckas begränsande och lite generellt men samtidigt ger det en tydlighet då det gäller resultat och slutsatser.

kim saxberg


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

113

â&#x20AC;&#x153;THIS PROGRAM CONTAINS ADVERTISINGâ&#x20AC;?: EFFECTS OF SPONSORSHIP DISCLOSURE ON PERSUASION KNOWLEDGE AND BRAND RESPONSES Sophie Boerman University of Amsterdam The Netherlands Eva van Reijmersdal University of Amsterdam The Netherlands Peter Neijens University of Amsterdam The Netherlands

INTRODUCTION The proliferation of sponsored content (i.e., that integration of brands, products and persuasive messages into traditionally non-commercial context) has raised the concerns of policy makers and consumer organizations. Since sponsored content (i.e., brand placement and brand integration) is embedded in non-commercial content, the source and persuasive intent of the message is not obvious. This makes it harder for viewers to distinguish between editorial and commercial content. Therefore, sponsored content is often referred to as deceptive (Cain, 2011; Nebenzahl and Jaffe, 1998). As a result, sponsorship disclosure on television is now obligatory in the European Union, and regulations are being developed in the United States (Cain, 2011, Schejter, 2006; Woods, 2008). Sponsorship disclosures explicitly inform audiences when content is sponsored. This way regulators attempt to guarantee fair communication and avoid persuasion without audience awareness. The few studies that have investigated the effects of sponsorship disclosures on television (i.e., Campbell et al., 2007; Dekker and Van Reijmersdal, 2010; Van Reijmersdal and Tutaj, 2010) mainly focused on how disclosures alter the effect of sponsored content on brand memory and persuasion. However, the primary goal of disclosures is to discourage deception by activating persuasion knowledge (Cain, 2011). Persuasion knowledge comprises a general understanding of persuasion and knowing how to cope with persuasive attempts (Friestad and Wright, 1994). There is a lack of research investigating the effect of sponsorship disclosures on persuasion knowledge. Moreover, most persuasion knowledge research has only directed the cognitive aspect of persuasion knowledge and does not consider its attitudinal dimension. This study examines how sponsor-ship disclosure influences persuasion knowledge (both conceptual and attitudinal) and brand responses (i.e., brand memory and brand attitude). Moreover, current regulations obligate broadcasters to display a disclosure for at least three seconds (e.g., Ofcom, 2011). However, we do not know whether three seconds is enough to inform viewers of sponsored content and its persuasive intent. Therefore, we test whether extending the duration of the disclosure could influence its impact. The Persuasion Knowledge Model (Friestad and Wright, 1994) describes how people develop general knowledge about persuasion and how people use this knowledge to interpret, evaluate and respond to persuasion attempts. Following Rozendaal et al. (2011), we believe that persuasion


114

knowledge research and advertising theories should not only focus on the cognitive aspects of persuasion knowledge but also should take into account its attitudinal aspect. Hence, we divide persuasion knowledge into a conceptual and an attitudinal dimension. Conceptual persuasion knowledge embraces the recognition of advertising, its source and audience, and the understanding of the advertisingâ&#x20AC;&#x2122;s persuasive intent, selling intent, and tactics. As the main goal of disclosure is to raise awareness of the commercial nature of content, this study focuses on one aspect of conceptual persuasion knowledge, i.e., the recognition of advertising. Supposing disclosure has its intended effect, we expect a sponsorship disclosure to enable viewers to recognize content as advertising and to activate conceptual persuasion knowledge. The attitudinal dimension comprises critical attitudes, such as skepticism and disliking, applied to a specific persuasion attempt (Rozendaal et al., 2011). It involves the critical feelings, such as honesty, trustworthiness and credibility. As the majority of consumers are quite skeptical toward advertising (Calfee and Ringold, 1994; Obermiller and Spangenberg, 2000), the disclosure may activate these general skeptical feelings toward advertising which are consequently applied to the sponsored content. Prior research showed that general critical attitudes toward advertising can automatically transfer negative affect to a specific advertisement (Lutz, 1985; MacKenzie and Lutz, 1989). In addition, due to the disclosure viewers may realize that the program and the brand are not neutral and are trying to persuade. As people usually do not want to be persuaded (Reactance theory by Brehm, 1966), this could result into critical feelings toward the sponsored content. Therefore, we expect that sponsorship disclosure activates attitudinal persuasion knowledge. In addition, research has shown that longer message duration provides more opportunities for viewers to attend to a message (Janiszewski, 1993) and to cognitively process it (Buijzen et al., 2010; Mackworth, 1963). As people need cognitive resources to process messages and these resources are limited, we cannot process all information at once (Buijzen et al., 2010; Lang, 2000). On television, viewers are exposed to sponsorship disclosure and program content simultaneously. Hence, when watching a television program, viewers allocate resources to its content and consequently may not have the necessary cognitive resources available to process the disclosure. Extending the duration of the disclosure provides viewers more opportunities and time to allocate resources to the disclosure and to process it, making it more likely that the disclosure will have an effect. This leads to the following hypothesis: Sponsorship disclosure activates conceptual and attitudinal persuasion knowledge; this effect increases with disclosure duration (H1). Brand Responses Next to activating persuasion knowledge, a sponsorship disclosure could alter the effect of the sponsored content and hence elicit consequences that are not its primary goal. Since a disclosure informs viewers about the presence of sponsored content, it puts more emphasis on the sponsored content and the brand that is making the persuasive attempt. Consequently, the disclosure functions as an additional prime for the brand and increases the chance viewers attend to the sponsored content and the brand, process it and store it in memory. Therefore, we propose that disclosure of sponsored content leads to greater brand memory. Furthermore, disclosures could elicit better brand memory indirectly through persuasion knowledge. The activation of persuasion knowledge requires attention to and awareness of the sponsored content and brand (Buijzen et al., 2010). Hence, when the disclosure activates persuasion knowledge, this increases attention to the brand even more, leading to greater brand memory. Therefore pose the following hypotheses: Sponsorship disclosure increases brand memory; this effect increases with disclosure duration (H2a) and The effect of the disclosure and its duration on brand memory is mediated by persuasion knowledge (H2b).


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

115

Although consumers can enjoy and like specific advertisements (Smit and Neijens, 2000), the majority of consumers are quite skeptical toward advertising (Calfee and Ringold, 1994; Obermiller and Spangenberg, 2000). Reminding them that the program they are watching contains sponsored content could therefore negatively influence what people think of the brand. A disclosure may serve as an extra cue for viewers to attend to the sponsored content and brand. Higher levels of attention to advertising have shown to enable people to activate persuasion knowledge (Campbell, 1995) and use their cognitive defenses, leading to diminished persuasion (Buijzen et al., 2010; Greenwald & Leavitt, 1984; Petty et al., 1981). This is because people tend to resist persuasion attempts when they recognize them as such (Brehm, 1966, Sagarin et al., 2002; Wei et al., 2008). When resistance of a persuasive message occurs, it is unlikely that attitudes become more favorable (Tormala and Petty, 2002). Several studies have demonstrated that the activation of persuasion knowledge leads to a less favorable brand attitude (e.g., Campbell, 1995; Lee, 2010; Wei et al., 2008). Based on reactance and forewarning literature we propose: Sponsorship disclosure has a negative effect on brand attitude; this effect increases with disclosure duration (H3a) and The effect of the disclosure and its duration on brand attitude is mediated by persuasion knowledge (H3b). METHOD We conducted an experiment (N = 209 college students, Mage = 22.21, 77% female) in which we compared the effects of no disclosure compared to a three-second and a six-second disclosure. The disclosure “This program contains advertising by Alive Shoes” was inserted into an edited episode of the television program MTV Was Here. This program consisted of three items about lifestyle, music and gadgets. The second item incorporated the sponsored content as it discussed a new brand of sneakers, Alive Shoes, which links shoes with social networking. Conceptual persuasion knowledge (conceptual PK) was measured by asking participants to indicate on a seven-point scale (1 = strongly disagree, 7 = strongly agree) to what extent the item about Alive Shoes was advertising (M = 5.62, SD = 1.41). Attitudinal persuasion knowledge (attitudinal PK) was measured by asking participants whether they agreed (1 = strongly disagree, 7 = strongly agree) with the statement, “I think the item about Alive Shoes in MTV Was Here is…” followed by five attributes based on ad skepticism scales (Obermiller & Spangenberg, 1998; 2000): “honest” (reversed),” “trustworthy,” (reversed) “convincing,” (reversed) “biased” and “not credible”. High scores of attitudinal PK correspond to high levels of skepticism and more disbelief, whereas low scores correspond to less critical feelings (EV = 2.78; explained variance = 55.64%; α= .79; M = 3.76, SD = 1.09). Brand memory was measured by asking participants whether they recalled seeing any brands in the episode of MTV Was Here and if yes, which brands. Brand recall is coded 1 (mentioned Alive Shoes) and 0 (did not mention Alive Shoes) and 51% of the participants mentioned Alive Shoes. Brand attitude was measured using six seven-point semantic differential scales: bad/good, unpleasant/pleasant, unfavorable/favorable, negative/positive, dislike/like and poor quality/high quality (e.g., Bruner, 2009; Campbell, 1995) (EV = 4.10; explained variance = 68.27%; α = .90; M = 4.81, SD = 1.11). RESULTS Fifty-two per cent of the participants (n = 93) that were exposed to a disclosure did not recall the disclosure. Because we focus on the effects of disclosure duration, only participants that recalled the disclosure were included in the analyses, leaving a sample of 116 participants (control n = 30, three-second disclosure n = 46, six-second disclosure n = 40). There was no significant difference in recall of disclosure between the threesecond and six-second conditions, χ² (1) = 0.68, p = .409. The experi-


116

mental groups did not differ with respect to sex, program familiarity, program viewing frequency, product interest, and program involvement (all p’s > .05). In addition, 96% of the participants were not familiar with Alive Shoes before participating in the study, and no participant owned the shoes. We included program involvement and product interest as covariates in all analyses, to make sure that they do not play any role in the effects. Persuasion Knowledge A MANCOVA with the experimental conditions as independent variables, conceptual PK and attitudinal PK as dependent variables reveals a statistically significant main effect of disclosure duration on conceptual PK, F(2, 110) = 11.60, p = .047, and a marginally significant effect on attitudinal PK, F(2, 110) = 4.68, p = .067. Pairwise comparisons demonstrated that participants who were exposed to the six-second disclosure (M = 5.92, SD = 1.01) had significant higher conceptual PK than participants that were not exposed to a disclosure (M = 5.07, SD = 1.44; p = .016). There was a marginal significant difference in conceptual persuasion knowledge between the control condition and the three-second condition (M = 5.72, SD = 1.44; p = .058). There are no significant difference between the threesecond condition and the six-second condition with respect to conceptual PK (p = .516). Attitudinal PK was significantly higher for the six-second disclosure condition (M = 4.10, SD = 0.96) compared to the no disclosure condition (M = 3.58, SD = 1.21; p = .048) and compared to the threesecond disclosure condition (M = 3.60, SD = 1.07; p = .042). Results showed no significant difference between the no disclosure and the threesecond condition (p = .875). These results demonstrate that disclosure does enhance both conceptual and attitudinal persuasion knowledge, especially when it is displayed for six seconds, supporting H1. Brand Responses To test the effect of any disclosure compared to no disclosure and the effect of disclosure duration we used orthogonal contrast coding. We constructed two contrast variables, the first corresponding to a contrast of no disclosure versus the three- and six-second disclosures, and the second corresponding to a contrast between the three- and six-second disclosures. Logistic regression analysis with disclosure duration as categorical predictor reveals that exposure to any disclosure significantly increase the odds of recalling the brand compared to no disclosure (see Table 1). After being exposed to a disclosure participants are more likely to recall the brand (-2LL = 144.24, Nagelkerke R² = .18, χ² (4) = 16.54, p = .002). Results show no significant difference in brand recall between the three- and six-second disclosure conditions. This means that sponsorship disclosure leads to greater brand memory, regardless of its duration. Hence, H2a is partly supported. To test for a possible mediation effect of the effect of disclosure duration on brand recall through persuasion knowledge we used Preacher and Hayes’ (2008) method of calculating standard errors and 95% confidence intervals. Because the confidence intervals for both conceptual PK and attitudinal PK did contain zero (see Table 2), results show no significant mediation. H2b is not supported. An ANCOVA demonstrated no statistically significant main effect of disclosure duration on brand attitude, F(2, 111) = 0.39, p = .675. Brand attitudes did not differ between the control condition (M = 4.91, SD = 1.19), the threesecond disclosure condition (M = 4.74, SD = 1.05) and the six-second disclosure condition (M = 4.82, SD = 1.15). H3a is not supported. Indirect effect analyses (Table 2) demonstrate a significant indirect effect when comparing the six-second to the three-second disclosure. The six-second disclosure results in significant less favorable brand attitudes (b = -.36, p < .001) as a result of higher attitudinal PK (b = .40, p = .046) compared to the three-second disclosure. An additional analysis with indicator dummy variables comparing the six-second disclosure to the control condition,


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

117

controlling for the three-second disclosure, demonstrated the same significant indirect effect of the six-second disclosure compared to no disclosure (Indirect effect = -.16, SE = 0.10, 95% CI -.40 to -.01). A six-second disclosure increases the critical feelings toward the sponsored program item, leading to a less favorable brand attitude. Hence, H3b is supported, but only via attitudinal PK and only for a six-second disclosure. DISCUSSION This study shows that sponsorship disclosures can activate both conceptual and attitudinal persuasion knowledge, but only when the viewer is exposed to a six-second disclosure. After exposure to a six-second disclosure, viewers are better able to distinguish commercial from editorial content compared to viewers who are not exposed to a disclosure. In addition, viewers exposed to a six-second disclosure show higher rates of attitudinal persuasion knowledge (i.e., more disbelief in the sponsored content) than viewers exposed to no disclosure and to a threesecond disclosure. These results emphasize the necessity to take both the cognitive and attitudinal dimension of persuasion knowledge into account, in theory and in future research. This is a valuable finding, as no earlier research has empirically investigated both the conceptual and attitudinal dimension of persuasion knowledge. Theoretically, our findings provide new insights into how sponsorship disclosure influences the persuasion process and the role of persuasion knowledge within this process. In addition, in accordance with the limited capacity theory (Lang, 2000), our results show that only a longer disclosure duration leads to these effects. Regarding brand responses, we found that sponsorship disclosure elicits two processes. The first process is a direct effect of sponsorship disclosure on brand memory. Sponsorship disclosure, regardless of its duration, leads to greater brand memory. The disclosure functions as an additional prime and enhances brand recall. The effect of a disclosure on brand memory is not mediated by persuasion knowledge. The second process is an indirect effect of the disclosure on brand attitude through attitudinal persuasion knowledge. Sponsorship disclosure does not directly alter consumersâ&#x20AC;&#x2122; perception of the brand. However, a six-second disclosure leads to higher rates of attitudinal persuasion knowledge (i.e., more disbelief in the sponsored content), which consequently results in resistance to persuasion, as demonstrated by more negative brand attitudes. This is in accordance with forewarning literature that demonstrated that revealing the persuasive intent of a message results in skepticism (Jacks and Devine, 2000; Wood and Quinn, 2003) and resistance (e.g., Lee, 2010; Petty et al., 1981; Quinn & Wood, 2004). Conceptual persuasion knowledge does not have the same effect. Contrary to earlier findings (Sagarin et al., 2002; Wei et al., 2008) and the reactance theory (Brehm, 1966), people do not resist the persuasive appeal of the sponsored content as soon as they recognize the persuasive attempt. As this study has only focused on one type of sponsored content, one brand and one disclosure, we need to use caution in generalizing the results. In addition, we focused on one specific disclosure, although the characteristics of disclosures are different between and within countries. Future research is needed to understand the effects of different types and content of disclosures, with different durations, in other contexts. Our findings show that the current obligated three-second disclosure (e.g., Ofcom, 2011) does not achieve its goal but that longer disclosure durations are able to affect both conceptual and attitudinal persuasion knowledge. Therefore, based on our outcomes, we recommend extending the obligated duration of disclosures.


118

förord: ”This time it’s personal” – den personliga faktorn i kommunikation I sin artikel ”It matters who ’they’ think you are” så pekar Magnus Söderlund och Claes-Rober Julander på ett intressant och förbisett fenomen – hur implicita antaganden om en kommunikationsmottagares identitet kan påverka resultat av kommunikation. Söderlund och Julander finner i studien att beroende på vilken av tre ”etiketter” som sätts på mottagaren i en inbjudan till en modevisning – arbetslös, student eller beslutsfattare på ett företag – så kan ett i allt annat identiskt och ganska attraktivt erbjudande få väldigt olika effekt. Om mottagaren upplever att de klassificerats rätt (student), så överväger de erbjudandet i högre grad än om de upplever att de klassificerats fel (arbetslös eller beslutsfattare). Och om de de klassificeras fel, så ger en positiv felklassificering (beslutsfattare) ett bättre resultat än en negativ felklassificering (arbetslös). I korthet - som kommunikatör ska du i bästa fall träffa rätt när du gör antaganden om vem mottagaren är, och i näst bästa fall hellre fria än fälla. Givetvis är studiens generaliserbarheten bortom själva experimentet ännu inte fullständigt prövad, men implikationerna av studien är extremt intressanta för alltifrån reklam till CRM-arbete. För implikationen är att det inte bara har betydelse vad ett varumärke erbjuder, utan också hur varumärket speglar mottagarens självbild och upplevda identitet. Det verkar överlag vara en bra ambition för ett varumärke att vara personligt, och söka spegla självbilden hos den man kommunicerar med. Men det är viktigt att vara medveten om att den ambitionen kan slå fel, och att ett ogenomtänkt försök kan göra mer skada än nytta. Och om man vet med sig att man riskerar att göra fel, är det helt enkelt bättre att klassificera till mottagarens fördel än nackdel. Till sist, så pekar också studiens resultat på att relationen mellan ett varumärke och en människa kan tänkas ha ett visst släktskap med vanliga sociala relationer. Man vill känna sig sedd. Man vill känna att man blir lyssnad på. Och man vill definitivt inte att någon tar fel på vem man är, vad man vill och vad man heter.

karl wikström


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

119

IT MATTERS WHO “THEY” THINK YOU ARE: CATEGORIZING THE CUSTOMER IN MARKETING COMMUNICATIONS AND ITS EFFECTS ON THE CUSTOMER Magnus Söderlund Stockholm School of Economics Sweden Claes-Robert Julander Stockholm School of Economics Sweden

INTRODUCTION Categorization of customers is a fundamental point of departure for marketing communications and it is incorporated in many firm’s segmentation practices. The prevalent use of segmentation, however, is likely to encourage the receiver of marketing messages to make inferences about who the sender (i.e., the categorizer) thinks that the receiver is in terms of category membership. This is indeed a hitherto neglected issue in marketing communications, yet given that others’ categorizations of oneself are intimately linked to one’s identity (Burke, 1991; Jenkins, 2000), we expect that such inferences are capable of influencing both evaluations and behavior vis-à-vis the categorizer. The purpose of this paper is to take some initial steps in examining the customer’s reactions to receiving a marketing message signaling how the customer has been categorized. In specific terms, we are interested in the potential for differences in customers’ evaluative reactions when a marketing message indicates that the customer has been classified either (a) correctly, (b) incorrectly in a positively charged category, or (c) incorrectly in a negatively charged category. An examination of this type, we argue, would contribute to the vast segmentation and marketing communications literature, which is focused on how the marketer should categorize customers – with little attention devoted to the possibility that the customer’s perceptions of the outcome of the categorization itself may affect the customer’s reactions. The examination would also complement the scope of categorization research, in the sense that few studies have examined the outcome of being correctly or incorrectly categorized in this or that category from the categorized person’s point of view – particularly when the source of the categorization inference is a message delivered by marketers. CONCEPTUAL FRAMEWORK AND HYPOTHESES Our conceptual framework rests on the notion of the individual’s social identity. It is a set of meanings applied to the self in a social role; social identity defines what it means to be who one is, and it is developed and negotiated by comparisons with others – and by assessments of others’ view of one’s identity (Burke, 1991; Jenkins, 2000). Social categories play an important role in this process, because the individual is assumed to continuously assess his/her self-categorization vis-à-vis how others’ categorize him/her (Barreto and Ellemers, 2003; Burke, 1991; Hogg and Terry, 2000).


120

A central tenet in social identity theory is that to be seen by others (external categorization) in a manner consistent with one’s stable self-views (internal categorization) reflects a fundamental need for psychological regularity and coherence. When a person is correctly classified by others, it is therefore expected that this should result in a positive state of mind for the classified person. And when a person is incorrectly classified by others it is expected that the misclassification elicits unsettling feelings of insecurity, as well as distress and anxiety (Burke, 1991; Campbell and Troyer, 2007). Some empirical findings indeed indicate that being misclassified leads to negative effects. For example, males who see themselves as manly react more negatively, in terms of negative affect, when they are misclassified by others as being less manly compared to manly males who are correctly classified as manly (Schmitt and Branscombe, 2001). Indeed, it has even been argued that our appetite for category correctness is so strong that we actively resist information that challenges our view of ourselves (Markus and Kunda, 1986). Others have argued that misclassifications leads to a strive to refute the classification, given an opportunity to do so (Swann and Hill, 1982). Nevertheless, individuals sometimes intentionally choose to masquerade as members of a category to which they do not belong, yet this is likely to lead to psychological strain and inner turmoil (Barreto and Ellemers, 2003; Clair et al., 2005). It is important to add that negative reactions are expected even when the misclassification by others has been done in terms of a positively charged category (Burke, 1991; Campbell and Troyer, 2007). Despite the fact that a social category in which an individual is placed by another person may have a more positive charge than the individual’s “real” category, the correctness of the categorization is thus assumed to override the valence of the category in producing a reaction to the categorization outcome. In tune with the predictions of social identity theory, and in a situation in which the classification of the customer can be inferred by the customer from a marketing message, then, we hypothesize the following: H1: When the customer is correctly classified, this outcome generates a higher level of evaluations of the sender compared to when the customer is incorrectly classified. Yet we expect that the two types of incorrect classifications are not psychologically equivalent when it comes to the reactions they induce. The main rationale behind this assumption is that many social categories have a valenced charge (Howard, 2000). And several studies have shown that this charge per se seems to be a potent causal agent. For example, the activation of negative self-stereotypes, referring to the content of the category to which the individual belongs, reduces the individual’s performance in an academic setting, while the activation of positive self-stereotypes may enhance performance (Wheeler and Petty, 2001). Given the valenced aspect of social categories, we expect that the charge of the category per se influences the receiver’s view of the outcome when he or she has been subject to categorization; we expect that it is more dissatisfying to be categorized in a negatively charged category than in a positively charged category (even if both categorizations are incorrect; cf. Campbell and Troyer, 2007). This outcome is derived from the premise that people have a basic desire to think favorable of themselves (Schrauger, 1975). Moreover, in cases in which it is possible to “move” between categories, an incorrect classification can be seen as an indication of what we may actually become (i.e., a possible self; cf. Markus and Nurius, 1986), and this ego-involving potential is likely to further boost the impact of the valence of the category.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

121

In a marketing setting – in which a message is providing clues for the receiver of how he or she has been categorized – we therefore expect that the valence of the category has a potential to affect the reactions to a misclassification. We hypothesize the following: H2: When the customer is incorrectly classified in a positively charged category, this outcome generates a higher level of customer evaluations of the sender compared to when the customer is incorrectly classified in a negatively charged category RESEARCH METHOD We used a between-subjects experimental approach in which each participant was either (a) correctly classified, (b) incorrectly classified in a positively charged category, or (c) incorrectly classified in a negatively charged category. The message in our case was an e-mail inviting the receiver to the opening of a new clothing concept store. The sender was a fictitious fashion firm (“Vercelli”), and the invitation promised an interesting evening with fashion shows, mingling with international models, and a free buffet. The three categorizations were manipulated by using a design in which we (a) selected participants with a given social identity, and (b) used this identity to create one correct and two incorrect categorizations (representing two “possible selves” with different valence; cf. Markus and Nurius, 1986). More specifically, we selected business school students as our participants. For the correctly classified participants, the invitation stated that “You who are a business school student are a part of our future. Therefore you are invited to celebrate the official opening with us”. For the incorrectly classified participants allocated to positively charged category, we used the category “decision-maker in a firm”; for the incorrectly classified participants who were allocated to a negatively charged category, the category “unemployed” was used. Besides this difference, the content of the e-mail message, in total 172 words, was identical for all participants. We thus assumed – for a business school student – that it would be positively charged to be seen as a decision-maker in a firm and negatively charged to be seen as unemployed. The three message versions were randomly allocated to the participants (N = 150; 80 men and 70 women; Mage = 21.17), who were undergraduate students enrolled in business administration courses. After having read the e-mail, the participants were asked to respond to a set of questionnaire items following the message. With regards to our measures, and to obtain a check variable for the manipulation of the correctness of the classification, we asked the participants about the extent to which they perceived that the sender (the firm behind the invitation) had made a correct analysis of their identity. We used three items scored on a scale ranging from 1 (do not agree at all) to 10 (agree completely): “The company’s analysis of my identity was correct”, “The company was good at identifying my social category”, and “The company was right about who I am”. Cronbach’s alpha for this scale was .92. For the evaluation of the sender, we measured the attitude toward the firm with four adjective pairs (bad-good, dislike-like, unpleasant-pleasant, and negative impression-positive impression) scored on a dimension ranging from 1 to 10 (alpha = .97). Moreover, as an additional evaluation variable, we captured the intention to accept the invitation; this measure comprised three adjective pairs (unlikely-likely, improbable-probable, and impossible-possible), again scored on a 10-point scale (alpha = .94). For each of these scales, we used the unweigthed average of the responses to the individual items as a variable for the analysis reported below.


122

ANALYSIS AND RESULTS To assess the manipulation, we conducted a one-way ANOVA with treatment group as the factor and the perception of the correctness of the company’s analysis of the receiver’s identity as the dependent variable. The omnibus result (i.e., the results of the global F test) was significant (F = 22.97, p < .01). More specifically, this analysis resulted in a higher correctness mean for correctly classified receivers (M = 6.05) compared to both incorrectly classified receivers in the positively charged category (M = 3.79) and incorrectly classified receivers in the negatively charged category (M = 3.21). The correctness means for the two incorrectly classified receivers were significantly lower (p < .01 in both cases) than the mean for the correctly classified receivers. Moreover, in the correctly classified group the perception-of-correctness mean was significantly higher (p < .05) than the scale midpoint (i.e., 5.5), while this mean was significantly lower than the scale midpoint for the two incorrectly classified groups (p < .01 in both cases). The manipulation thus behaved as intended. The resulting means (and standard deviations) for our two evaluation indicators, (1) the attitude toward the firm and (2) the intention to accept the invitation, in the three conditions are presented in Table 1. To test the hypotheses, we employed a one-way ANOVA in which treatment group membership served as the factor (we assessed the pairwise differences with the Scheffé test). In the first step, we performed the ANOVA with the attitude toward the firm as the dependent variable. The omnibus result was significant (F = 19.71, p < .01). The attitude for correctly classified receivers (M = 6.91) was significantly (p < .01) higher than for incorrectly classified receivers in the negatively charged category (M = 4.74), yet it was not significantly (p = .57) higher compared to incorrectly classified receivers in the positively charged category (M = 6.51). Hypothesis 1 was thus only partly supported. Moreover, the attitude toward the firm for incorrectly classified receivers in the positive category (M = 6.51) was significantly (p < .01) higher than for incorrectly classified receivers in the negatively charged category (M = 4.74), thus supporting Hypotheses 2. As for the intention to accept the invitation, the omnibus F test produced a significant result (F = 11.6, p < .01). The intention for correctly classified receivers (M = 6.70) was significantly (p < .01) higher than for incorrectly classified receivers in the negatively charged category (M = 4.53), but it was not significantly (p = .48) higher compared to incorrectly classified receivers in the positively charged category (M = 6.13). Again, Hypotheses 1 received only partial support. Hypotheses 2, however, was supported; the intention to accept the invitation for incorrectly classified receivers in the positive category (M = 6.13) was significantly (p < .01) higher than for incorrectly classified receivers in the negatively charged category (M = 4.53). In sum, then, our results show that a correct classification was rewarded by a more positive attitude toward the sender, and higher intentions to act on the message, compared to an incorrect classification in a negatively charged category. A correct categorization also produced more positive reactions than an incorrect classification based on a positively charged category, yet this difference was not significant. Furthermore, the results show that the two types of incorrect categorizations were not psychologically equivalent – an incorrect classification in a positively charged category produced higher attitudes and intentions than an incorrect classification in a negatively charged category. Indeed, given the scale midpoint (i.e., 5.5) as a demarcation point between negative and positive evaluations, the positively charged misclassification produced positive evaluations while the negative misclassification produced negative evaluations.


THE CHANGING ROLES OF ADVERTISING / THE 11TH ICORIA / STOCKHOLM 2012

tack!

stockholm design lab

Monika Ocieczek

123


124

ICORIA 2012  

http://icoria.org/

Advertisement