Immediate and Delayed Response - SSRN

1 downloads 0 Views 277KB Size Report
Send correspondence to Patrali Chatterjee, School of Business, Montclair State. University, Upper Montclair, New Jersey 07043, tel: 973-655-7935, fax: 973-655 ...
Please cite as: Chatterjee, Patrali (2011), “Can Unconscious-Conscious Processing Sequences Enhance Ad Exposure Outcomes?” Journal of Brand Management, 18, 506-515.

CAN UNCONSCIOUS-CONSCIOUS PROCESSING SEQUENCES ENHANCE AD EXPOSURE OUTCOMES?

Patrali Chatterjee, Montclair State University

Submission: May 2010 Revision: September 2010 Acceptance: October 2010

Send correspondence to Patrali Chatterjee, School of Business, Montclair State University, Upper Montclair, New Jersey 07043, tel: 973-655-7935, fax: 973-655 (Email: [email protected]).

Electronic copy available at: http://ssrn.com/abstract=1861324

2

Author bio: Patrali Chatterjee is Associate Professor of Marketing at the School of Business, Montclair State University. Prior to joining Montclair State University, Dr. Chatterjee was Vice Chair and Assistant Professor, Department of Marketing at Rutgers Business School, Rutgers University. She holds a Ph.D. in Marketing from Owen School of Management, Vanderbilt University. Prof. Chatterjee's research interests are online advertising and promotions, clickstream data modeling, and customer relationship management. Dr. Chatterjee's research has appeared in several books and academic journals including Advances in Consumer Research, Journal of Business Research, Journal of ComputerMediated Communication, Journal of Electronic Commerce Research, Marketing Science and the Review of Economics and Statistics. Her research has been reported in news and business media outlets like The Wall Street Journal and Stanford Innovation Review among others.

Electronic copy available at: http://ssrn.com/abstract=1861324

3

Abstract

Can the sequence in which ad exposures are generated across various ad formats enhance or inhibit communication outcomes in low involvement situations? This study compares the effect of self-selected and forced exposure (or intrusive) ad format sequences compared to repeated exposures of single ad formats on brand recall, brand attitude, and stimulus-based inclusion in consideration set. Analysis of experimental data show that the sequence of self-selected ad formats that induce pre-attentive processing of ad information first, followed by forced exposure ad formats that induce conscious processing lead to superior outcomes compared to others. Prior research on priming and supports the findings and implications for media planning are addressed.

Keywords: forced exposure, implicit memory, priming, preattentive processing, selfselected exposure, synergy

4 Introduction Academic researchers have unanimously demonstrated the existence of synergies in multiple media advertising (Dijkstra, Buijtels, and Van Raaij 2005; Edell and Keller 1999). Much of these demonstrations are in experimental conditions when respondent involvement with the task is high. In such situations, the ad is very salient, memory traces of the ad are very accessible, and impact on communication outcomes and attitudinal measures are strong (Chattopadhyay and Nedungadi 1992). However, most ad processing is under lowinvolvement conditions (MacInnis, Moorman and Jaworski 1991). Whether or not synergies attributed to usage of multiple media or ad formats occur in low-involvement or natural media consumption conditions is unclear. Prior research on multiple media effects has examined media types that differ in modality. The implicit assumption is that cross-media synergies accrue only when media differ in modality or number of sensory modes they engage (Edell and Keller 1989). Encoding variability (Tassavoli 1998) suggests that when a consumer receives the same message from several media sources, they encode the message in their memory in such a manner as to enhance the likelihood of recalling information correctly. This outcome may occur when consumers process ad messages centrally, in high involvement situations, however most advertising exposure in natural media setting occurs in low-involvement conditions. In low-involvement situations characterized by ad clutter and interference, consumers are likely to process ads preattentively. Advertisers can use intrusive ad formats to force consumers to attend to advertising (Shavitt, Vargas, and Lowrey 2004), however do negative effects of such exposures impact outcomes? The combined or joint impact of multiple-ad formats in media on memory-based communication outcomes is largely unknown (Vakratsas and Ambler 1999). Hence, examining the impact of multiple ad formats in a single ad

5 campaign on ad-evoked attitudes, brand recall and inclusion in consideration set under low involvement conditions is an important research goal. Ad formats within each media differ significantly in the degree to which they induce conscious or unconscious processing based on whether exposure is self-selected or forced/ intrusive (Shavitt, Vargas, and Lowrey 2004). The integrated information response model (Smith and Swinyard 1982) proposed that consumers’ acceptance of beliefs communicated through advertising can be affected by message sequencing. However, in low-involvement situations, consumers are unlikely to notice self-selected ad formats (like print, billboards, and Internet banner ads, Dreze and Husherr 2003) even though they may process it preattentively. Does this mean ad campaigns should only use intrusive or forced exposure ad formats (e.g., Internet pop-up, interstititial ads or TV)? This research explores an alternate route that sequences of conscious and non-conscious processing, induced by self-selected or forced exposure ad formats within the same media, can lead to improved communication outcomes compared to those induced by a single ad format alone. Specifically, we examine whether self-selected (Internet banner) and forced exposure (Internet interstitial ads) ad formats complement each other in the route to persuasion because of their differential effects on implicit versus explicit memory compared to single ad format usage. While self-selected ads do not generate explicit memory, pre-attentive processing leads to perceptual analysis of ad information leading to implicit memory traces that drive familiarity based brand preference (Janiszewski 1998). Pre-attentive processing due to selfselected ads likely plays a preparatory role in increasing consumer’s motivation to process ad information on forced exposure ads. The choice of ad formats reflects current commercial practice and the fact that the ad formats do not differ in modality this controlling for potential confounding effects on the findings.

6 The next section discusses prior research on conscious and unconscious information processing induced by ad formats considered in this research in low involvement conditions. Section three presents hypotheses for unconscious-conscious, conscious-unconscious processing sequences compared to conscious and unconscious ad processing only. Section four presents a test and findings in an experimental study and compares brand recall, brand attitude, and stimulus-based brand consideration when using multiple versus single ad formats.

Theoretical Background & Hypothesis Consumer involvement with the brand or product category, availability, and complexity of ad information affects ad information processing intensity. According to the elaboration likelihood model (ELM; Petty and Cacioppo1986) under high involvement situations (high relevance or risk), people are motivated to engage in diligent deliberation of attribute-relevant information (e.g., arguments) in the ad. In contrast, when viewers lack sufficient motivation (because of low relevance or risk) or ability (because of personal traits or external interference), persuasion follows a peripheral route by which people base their attitude and evaluation on attribute-irrelevant information or peripheral cues or physical attributes of the ad (e.g., a picture’s attractiveness, source characteristics, music, message sidedness, etc.). If an ad induces elaboration, recognition and recall will improve (Krishnan and Chakravarti 1999). Advertising formats are associated with different degrees of self-selected versus forced ad exposure experiences which impact psychological processes underlying evaluations of the advertised brands. Self-selected ad formats (e.g., print magazine, catalogs, billboard, Internet banners, Internet search ads), where perceived interest in the ads is the basis for attention to them, receive more favorable evaluations by consumers than forced exposure ads

7 (Shavitt, Vargas, and Lowrey 2004). Self-selection allows consumers to obtain information most relevant to their needs or interests and allocating extensive processing resources to ads of interest but also allows consumers to easily ignore or avoid ads that are not relevant when involvement is low. In contrast, forced exposure ad formats (e.g., TV, radio, Internet pop-up and interstitial ads) are difficult to avoid. Advertising formats that are forced onto consumers without regard to their choice are perceived to be intrusive and elicit negative feelings of irritation, reactance, and irritation with adverse effects on subsequent brand attitudes (Shavitt, Vargas, and Lowrey 2004). In low-involvement situations repeated ad exposures are used to increase the likelihood that a brand name is remembered and strengthen the association between the brand name and the advertised product benefit in memory. Prior literature on repetition effects in print and television media suggests a non-monotonic or inverted U-shaped relationship between message repetition and message effectiveness (cf. Cacioppo and Petty 1979). Exceptions to the inverted U-shaped relationship (Nordheilm 2002) occur when presumably respondents have little or no motivation or opportunity to elaborate on the stimuli being presented, similar to conditions of incidental exposure to banner ads. In fact, no downturn in affective responses are observed when respondents’ opportunity to process stimuli consciously is constrained (Bornstein and D’Agostino 1992). The mere exposure effect (Zajonc 1968) suggests that a mere increase in exposure to stimuli would increase positive affect toward those stimuli. The perceptual fluency/ misattribution model posits that repeated exposure to a stimulus will create a feature-based representation of a stimulus (i.e., feature analysis) in memory that will facilitate encoding (perceptual fluency) and processing of the stimulus when viewed at a later time (Nordheilm 2002).

Multiple Exposures to Self-selected-only or Forced-exposure-only Ad Formats

8 Self-selected ad formats like print and Internet banner ads are embedded within editorial content and allow consumers to process information at their own pace. Prior research shows that Internet banner ads are more likely to be implicitly processed in low involvement situations (Dijkstra, Buijtels, and Van Raaij 2005, Chatterjee, Hoffman, and Novak 2003). Research in low involvement and unconscious processing suggests ad information in incidental exposure condition is processed at a subconscious level without awareness or recognition (Janiszewski 1993). It is manifested through ―memory without perception‖, i.e.,

the presence of implicit memory but absence of explicit memory. If Internet banner ads are used in an entire ad campaign, after a certain level of repetition consumers may not recall having seen the preattentively processed ad stimulus before yet find it familiar. Conventional wisdom holds that conscious processing of ads is necessary for ads to impact consumers’ preferences. In low-involvement conditions, when conscious processing is induced by forced exposures to ads immediate attitudinal measures depend on the ad exposure context and ad-induced feelings and cognitive thoughts generated by the ad execution, ad copy, brand claims and attribute-level information. Since, the impact of the valence of attribute-level information is not under investigation in this study, the valence or quality of ad copy and attribute-level information is held constant across all ad formats. This research considers the effect of interstitial ads as forced-exposure ads. Interstitials (or ―transition‖) ads load in the time between two content pages while the consumer is waiting for the next Web page to download. It completely fills the browser as if it is a page itself. Interstitial ad pages have a link to skip the ad and a count-down message. Unlike pop-ads, an interstitial is usually designed to move automatically to the page the user requested after allowing enough time for the message to register or the ad(s) to be read. Since consumers have no choice but view the interstitial ad, positive habituation can set in early, at the first exposure itself.

9 Pechmann and Stewart’s (1988) two-stage learning model explains the inverted-U curve response of advertising in terms of "wearin" and "wearout." They suggest that wearin occurs during approximately the first three exposures with diminishing returns on affect due to each additional exposure, and at the third exposure, positive thoughts finally outnumber negative thoughts. Wearout begins with approximately the fourth exposure as message recipients become bored. As a result, message recipients generate negative repetition-related thoughts, undermining the persuasive impact of the ad. Repeated exposures to the same forced exposure or interstitial ad will lead to tedium and result in inattention to the stimulus and disliking of the ad stimulus at later exposures. Further, the declining marginal impact of exposure to repeated interstitial ads on brand attitude will occur earlier than repeated banner ads.

Exposures to Self-selected and Forced Exposure Ad Format Sequences When both self-selected and forced exposure ad formats appear in the same ad campaign, processing intensity at the time of initial ad format exposure affects what is stored in the ad memory trace (e.g., brand name, product category, brand claims, ad execution elements, and affective and cognitive reactions). The organization of the ad memory trace interacts with the available cues at subsequent exposures to determine the retrievability of information from memory (Edell and Keller 1999). Retrieval processes affect encoding processes during the successive exposures to the ad. That is, when a person receives an exposure to an ad later, the ad may serve as a retrieval cue for the stored ad memory trace or as another encoding opportunity. A person may elaborate on and have affective and cognitive reactions to either the current ad, the ad memory trace, or to both. Because limited processing resources are available, these three processing activities—comprehension, retrieval, and elaboration—must compete for the available processing resources during subsequent ad

10 exposures. For example, a person may not attempt to elaborate on the current ad, instead concentrating solely on retrieving the stored ad trace. Differences in the sequencing of initial versus later ad formats imply that either implicit processing leads to explicit processing or vice-versa which may have a synergistic or depletion effect on brand attitude, brand recall and stimulus-based brand consideration due to repeated ad exposures. Retrieval activities are more likely to occur in the mixed ad format conditions than in single-ad format conditions because the subsequent ad formats may be easier or harder to ignore, even if the information on the ad is the same. If processing motivation is low, subjects viewing the forced exposure ad format or interstitial ad first in a self-selected-then-forced exposure sequence will be forced to view ad information and process information peripherally. However, the negative attitude towards the forced exposure ad is likely to lead subjects to ignore the self-selected (banner and print) ad formats on later exposure occasions. Explicit memory of the initial ad exposure context is less likely to endure over later exposures since subjects expend processing effort retrieving the stored ad trace during exposure to subsequent print and banner ads and have less capacity available for encoding information on the banner and print ads. In contrast, consumers exposed to a self-selected-then-forced-exposure ad sequence will preattentively process information on banner ads first. Research on priming suggests that activation or the establishment of implicit memory by the first exposure to an ad (i.e., the primed stimulus – banner ad) will increasingly ease processing during later ad exposures. This perceptual fluency or ease of processing will lead to positive brand attitude (Lee and Labroo 2004). Even if consumers don’t notice the banner ads or attend to information on them, memory traces from Internet banner ad processing will prime the brand depicted within the ad as well as the semantic information for the brand and the words appearing in the print

11 and later email ads, even if there is no explicit memory of the ad immediately after Internet banner ad exposure. H1: Brand attitude associated with self-selected-then-intrusive ad format campaign will be significantly higher than that of (a) intrusive ad format-only campaign, (b) intrusivethen-self-selected ad format campaign, and (c) self-selected ad format-only campaign for the same number of ad exposures. A brand that is highly accessible in memory and thus comes to mind easily or recalled, has a greater chance of being considered and selected than a less accessible brand. Thus, forced ad exposures that lead to greater explicit memory enhances brand accessibility or brand recall. Pre-attentive processing, however, would produce limited or no explicit memories of the exposed ad information, and consumers may not be able to retrieve target brand names from explicit memories. Accordingly, unconscious processing of Internet banner ads would not benefit brand recall. H2: Brand recall associated with forced exposure ad format-only campaign will be significantly higher than that of (a) self-selected ad-format only campaign, (b) intrusive-thenself-selected ad format campaign, and (c) self-selected-then-intrusive ad format campaign. The term consideration set refers to the subset of all available brands brought to a consumer’s mind on a particular choice occasion and can be memory-based or stimulusbased, or both. Enhanced perceptual and/or conceptual fluency due to unconscious processing of self-selected ad formats like Internet banner ads will benefit stimulus-based brand choice despite the lack of recall. Chattopadhyay and Nedungadi (1992) demonstrated that an increase in brand accessibility through exposures results in a greater probability of choice without changing brand preference. Similarly, Shapiro, MacInnis, and Heckler (1997) report the likelihood that an advertised product would be included in the consideration set is greater in the pre-attentive

12 processing condition than in the control group. A consumer can choose a brand among other alternatives in the product category in a stimulus-based choice situation, based on information in memory they are less likely to do so when they are negatively predisposed towards the brand. Hence, despite higher levels of brand recall for forced exposure-only ads, consumers are less likely to select the brand because of negative attitudes towards the ad exposure context relative to other ad formats. H3: Subjects are less likely to include the advertised brand in the stimulus-based consideration set in forced-exposure ad-only campaign compared to (a ) self-selected adformat only campaign, (b) intrusive-then-self-selected ad format campaign, and (c) selfselected-then-intrusive ad format campaign.

Experimental Design and Method The experimental design for the lab study had five media use conditions (self-selected ad format only; intrusive ad format only; self-selected then intrusive ad format sequence; intrusive then self-selected ad format sequence and control group). One hundred sixty undergraduate students (approximately thirty in each group) at a major Northeastern university completed the study. To simulate a natural media consumption experience, banner and interstitial ads were embedded in a modified version of the online student newspaper. To prevent boredom by exposing subjects to the same content in a row, three versions were created with different news articles in each version. Articles from the forthcoming issues of campus newspaper were used as editorial content. Links to external web sites were removed and the experimental web site and email newsletter pages were launched from an internal server in a computer lab to ensure page download times were same. Online questionnaires were administered at the end of each session to collect dependent measures. Filler ads for ten fictitious sponsors were used to mimic ad clutter. We induced involvement in the activity by

13 instructing subjects that the student association was interested in student opinions of the format, campus articles, and types of ad sponsors and was planning a redesign of campus publications based on their feedback similar to other research (Janiszewski 1998). Selection of Ad Stimuli: Prior to development of experimental stimuli, a different set of 41 undergraduate students answered a questionnaire to measure their involvement with fifteen product categories, and relevance and likelihood of the product category being advertised in the campus weekly publications (Rodgers 2004). Credit card and car rental were chosen as product categories for ad stimuli because they were associated with high involvement, and both were equally rated as relevant to be advertised in the campus newspaper. Ads for two product categories were used to examine generalizability; credit card (i.e., Apply for Victory Mastercard, low APR, www.victorymastercard.com) and rental car agency (Supervalue Rental Car Agency, same low rate all week, www.supervaluerental.com). Since prior familiarity with the advertised brands could potentially confound results (Dahlen 2001), we used fictitious brands for focal and filler ads. All ads had the same pictorial execution in all treatment conditions and neither quantity nor quality of information was varied across the ad formats. Procedure Every student had a numeric code to access questionnaires and web site across stages and receive course credit. Subjects were randomly assigned to one of the conditions according to the experimental design. All subjects were asked to carefully browse through the campus newspaper web site for 10 minutes and answer questions at the end of the session. Subjects in the interstitial condition were given the experimental campus newspaper, asked to read for 10 minutes and answer questions on a computer at the end of the session. In the single media condition, students were allowed to browse for 20 minutes. In the mixed media version,

14 students were asked to switch to the other media after 10 minutes. Control group subjects were exposed to filler ads only and not the experimental ads. Subjects (other than those in control group) were exposed to the experimental ad stimuli for one focal sponsor (counterbalanced) and filler ads two times in each media, for a total of 6 times during the session. After 30 minutes at end of media use, subjects were taken to an online questionnaire to provide dependent and manipulation check measures. Demographic, Internet use information, awareness and frequency of use of the campus newspaper, student newspaper site, credit card ownership and frequency of rental car usage were collected. Participants were thanked and debriefed before leaving. Following exposure to the treatment data were collected to measure the three dependent measures. Brand recall was calculated by considering each advertisement to have four elements: a brand name, ad claim, image used and a web address (or URL). Subjects scored 0.25 for each element correctly recollected or recognized. In the recall measure, subjects were asked to list all ad information they remember seeing. Stimulus-based brand consideration was measured using sequential two-alternative forced choice tests of brand name, claim, image and URL conditional on a product-category need. It was collected after the recall measure but before attitude measures. Brand attitude was assessed by 3 seven-point scales anchored by phrases ―good-bad,‖ ―like-dislike,‖ and ―favorable-unfavorable.‖ The  coefficient was 0.92 indicating a high degree of internal consistency. In addition, subjects responded to two new statements on seven-point scales anchored by strongly agree and strongly disagree. The statements were ―Victory (Supervalue) is a good credit card (rental car agency)‖ and ―Victory (Supervalue) offers favorable interest (rental) rates.‖ Manipulation Checks: Manipulation check for information quality in terms of relative importance of claims in the ad formats, was not successful, hence the relative importance of

15 claims across the media were similar. The perceived proportion of screen space occupied by interstitial ads was significantly higher than banner ads (69% vs. 19%, t=7.97, p0.05) in initial analyses, and dropped from further analyses. Initial ANOVA analyses revealed no significant main or interaction effects of product type for each of the dependent variables. Consequently, data was pooled across the two products for the reminder of the analyses. Table 1 presents the means and significance tests for the ad format conditions on brand attitude, brand recall and stimulusbased brand consideration.

Place Table 1 about here.

Hypothesis 1 considers the effect of ad format usage on brand attitude. The control group was added and included in the statistical analysis. The ANOVA indicates main effect of ad format (F(4,152)=5.96, p