دیدن جنگل با وجود درختان : اثرات نام تجاری (برند) بر عدم قطعیت انتخاب

Prior research on brand equity suggests that consumers use brands as signals to reduce uncertainty and perceived risk. Erdem and Swait (1998) developed a conceptual framework based on information economics and signaling theory to explain how equity is created, maintained and transferred over time that involves seven theoretical constructs. This paper reviews the impact of brand-equity-associated brand utility on the scale of the indirect utility function (i.e., the inverse of the error variance); we argue that higher brand-equity-associated brand utility reduces the need for consumers to review previously formed preferences. We combine a brand utility experiment with a brand feature experiment to estimate the effects of brand-equity-associated brand utility scores on choice. We find that higher brand-equity-associated brand utility leads to higher choice consistency, which can drive increases in market share.

The concept of a brand (defined as “a name, term, sign, symbol or design, or a combination of them which is intended to identify the goods and services of one seller or a group of sellers and to differentiate them from those of competitors”; Kotler, 1997, p. 443) is widely regarded as a key marketing principle. Different research streams focus on different roles that brands play in consumer choices. One stream focuses on the impact of brands on consumer utility in random utility choice models (e.g., Kamakura and Russell, 1993, Louviere and Johnson, 1988 and Park and Srinivasan, 1994). A second stream focuses on price premiums that consumers are willing to pay (e.g., Kamakura and Russell, 1993 and Park and Srinivasan, 1994) or differences in price sensitivity for strong brands (e.g., Keller, 1993 and Sivakumar and Raj, 1997). Other streams focus on ways to measure brand health and brand satisfaction, among other issues (e.g., Ailawadi et al., 2003 and Bloemer and Kasper, 1995). In this paper, we adopt the perspective of Erdem and Swait (1998), who view brands in an information economics framework. They propose that markets are characterized by imperfect and asymmetric information and that “consumer uncertainty about product attributes may exist even after active information gathering (for experience attributes) or after consumption (for long-term exposure or credence attributes)” (Erdem & Swait, 1998, p. 138). Companies can reduce this uncertainty by sending signals about their product quality, for example, through advertising (Milgrom & Roberts, 1986) or manufacturers' warranties (Lutz, 1989). In the information economics perspective on brand equity, a brand name itself serves as a signal, and thus, “consumer-based brand equity is defined as the value of a brand signal to consumers” (Erdem & Swait, 1998, p. 133). For example, umbrella branding (in which firms use the brand name of established products for new products) is one way to send a quality signal about a new product to consumers. This perspective on brand equity ascribes costs to information acquisition processes that consumers use to resolve uncertainty. Higher brand equity, which is defined as a strong brand signal, reduces these information costs, in turn leading to higher brand utility (Erdem & Swait, 1998). Over the past several decades, product markets have become more variable, with a proliferation of SKUs in many product categories.1 Additionally, and more recently, major supermarket chains across the world have begun to move into private-label products, further increasing the number of choices and diversity available to customers. Not only are more signals and more diverse signals being sent by firms about an increasing number of brand offers, but the proliferation of these signals now also occurs across many more communication channels, such as Facebook, Twitter and other new media. Moreover, the availability of consumer product ratings via consumer-review websites and e-commerce websites, such as http://www.yelp.com/, http://www.walmart.com/ and http://www.amazon.com/, increases the complexity of brand signals. A potentially important issue that has been largely ignored in the brand equity literature is the impact of a brand name on choice consistency. Choice consistency is one of several possible unobserved utility components in random utility theory-based choice models.2 That is, the stochastic component of utility (the so-called “error component”) can be decomposed into several possible subcomponents, such as variability in choices due to mistakes, inattention, differences in familiarity with choice options and model misspecification. We propose that brands that provide strong signals to consumers are also likely to exhibit more consistent choices in scanner panel and choice experiment data sources. Strong brands help a “decision maker more strongly discriminate between that brand and others, because the evaluation of the former product may be less subject to idiosyncratic uncertainties (e.g., different levels of knowledge about attributes) compared with that of other brands” (Swait & Erdem, 2002, p. 307). This discrimination further suggests that stronger brands can simplify consumer decision processes and reduce the need to reevaluate products, making purchase decisions easier for consumers. One measure of choice consistency is the scale of the indirect utility function, which is inversely proportional to the standard deviation of the error component. This scale determines how “in different contexts and for different decision makers, the same systematic utility difference can result in more-extreme choice probabilities” (Swait & Erdem, 2007, p. 682). As Swait and Louviere (1993), Louviere, Hensher, and Swait (2001, Chapter 8), Salisbury and Feinberg (2010) and Fiebig, Keane, Louviere, and Wasi (2010) indicate, if individuals differ in their scales, there can be significant implications for marketing policies that rely on choice modeling results. For example, suppose there are two segments of consumers, one with considerable experience in a category that makes very consistent product choices, and a second that is new to the category and makes much less consistent choices. Even if the consumers in these two segments use identical decision rules (i.e., choice models and associated indirect utility specifications) to make choices, their observed (and predicted) choice probabilities (i.e., proportions) will differ. That is, the segment that makes choices more consistently should exhibit a wider range of choice probabilities in any particular context (e.g., choice set, choice occasion, etc.) than the less consistent segment. Indeed, at the extreme of almost perfect choice consistency, observed choice probabilities will be close to zero and one whereas when choice consistency decreases, observed choice probabilities will be close to 1/J, where J is the number of choice options offered. Thus, managers need to better understand that the range of predicted choice shares is likely to be smaller/larger for segments (individuals) with lower/higher error variances or choice consistency, and this should be taken into account when making marketing policy decisions. In turn, this implies that choice consistency or the scale of the utility estimates is likely to be useful for positioning and targeting. Prior literature identified several factors that are associated with differences in the scale of the indirect utility function, including information frames (e.g., Swait & Adamowicz, 2001), labeled alternatives in choice sets (e.g., brand names) and individual differences, such as education/literacy, age and involvement (see, e.g., Louviere et al., 2001, Chapters 8 and 13). Consequently, the purpose of this paper is to test whether brand names affect consumer choice consistency, which in turn affects consumer choices. We show that brands systematically affect consumer choice consistency in a series of discrete choice experiments, and our results suggest that if managers understand and can predict differences in effects on choices due to “pure preferences” and choice consistency, it should be possible to make more effective use of a brand's marketing mix so as to affect both preferences and choice consistency. For example, managers can manipulate the strength of a brand signal by carefully choosing suitable advertising and product line extension branding strategies, which send out consistent quality signals to the consumer and thus strengthen a brand's credibility. Similarly, different advertising and communications media also likely vary in their credibility, and thus, managers can likely improve brand credibility and other equity signals by appropriate media choices, a topic that we leave for future research. In a practical sense, differences in choice consistency matter in marketing research because, as noted by Fiebig et al. (2010) and Salisbury and Feinberg (2010), if scales differ across contexts, segments and individuals (among other things), these scale differences must be accounted for in traditional choice models (including models that attempt to capture preference heterogeneity) to avoid confounding partworth estimates with these differences. Scales are likely to differ by brand and attribute as well as by the individual, as shown by Louviere and Eagle (2006) and Louviere and Meyer (2007). If the estimated utility of a particular brand is due in part to pure preference for the brand and in part to the choice consistency associated with that brand, existing choice models such as conditional logit or its extensions (i.e., random effects or latent class models) can potentially yield incorrect and misleading results. That is, ignoring error variability differences in choice models may result in biased parameter estimates rather than a simple loss of efficiency (Swait & Erdem, 2002). This paper tests the hypothesis that brand-equity-associated brand utilities affect unobserved choice consistency. To empirically examine this proposition, we rely on the Erdem and Swait (1998) theory of brand equity in a discrete choice setting using respondents' choices from two related discrete choice experiments (DCEs) (Louviere and Woodworth, 1983 and Street et al., 2005). One DCE asks participants to choose among alternatives described by the Erdem and Swait theory constructs (a brand equity DCE), and a second DCE asks them to choose among choice options described by attributes/features (a traditional DCE). Thus, the brand equity DCE is used to measure brand equity as defined by signaling theory, and we test the effects of the brand equity measures on choice consistency in the second DCE. To determine the magnitude of the choice consistency effect, we use recent advances in random utility models to estimate a more flexible and general model that simultaneously allows for a distribution of scale and a distribution of preferences. In particular, we develop and apply a model that combines the Generalized Multinomial Logit, or G-MNL model (Fiebig et al., 2010), with a variant of the Random Coefficients Generalized Scale Multinomial Logit (RGCSMNL) model by Swait and Erdem (2002). The remainder of the paper is organized as follows. We first describe the two discrete choice experiments and then discuss conceptual and statistical model considerations, followed by the presentation of results from six empirical datasets. We conclude by discussing the managerial implications of this study and directions for future research.

We studied the impact of Erdem and Swait's (1998) brand-equity-associated brand utility measures on product choices. We combined two DCEs to test the association between a brand-anchored DCE (used to measure brand utility) and a brand-feature DCE (which varied key features of real brands). For five out of the six categories, we found a significant positive impact of the brand-equity-associated brand utility rankings on the scale of utility in the choice data. Thus, our results indicate that higher brand-equity-associated brand utility measures significantly increased respondents' utility scale. Thus, stronger brands led to more consistent choices, supporting Erdem and Swait's (1998) view of brands as information signals. In turn, this finding suggests that stronger brand signals can help simplify consumer decision processes and/or make consumers more confident about their choices, leading to more consistent choices. However, our results also showed that model fit only improved in two of the six categories, implying that the impact of brand equity on scale may have limited importance in explaining choice behavior. Similarly, given that we could not completely rule out priming effects that may arise from the brand equity experiment DCE 1, the extent to which brands affect scale may be lower in real-world choice settings. The last two issues should be the subject of further research to clarify and generalize this study's findings. Together with prior work on the impact of brands in consideration set formation (Erdem & Swait, 1998), our results suggest that stronger utility brands can help guide consumers' purchase decisions. Retailer assortment sizes have greatly increased in many markets, possibly making choices more difficult for some consumers (Botti and Iyengar, 2006, Chernev, 2003 and Huffman and Kahn, 1998). Our results imply that marketing managers not only can increase brand-equity-associated brand utilities in ways shown by previous research but also may be able to deal with choice complexity by developing and maintaining stronger utility brands and transmitting consistent and credible signals about their quality. Given that credibility, as well as the other underlying brand equity constructs, is influenced by managerial actions affecting the brand, one take-away from our results is that marketing actions by managers can also affect the choice consistency of a market. However, it is important to note that higher choice consistency—or “preference discrimination”, as Swait and Erdem (2007) refer to it—does not imply that consumers are more likely to purchase a stronger utility brand. Rather, it means that (all else equal), stronger brands will exhibit more extreme choice probabilities. Thus, consumers will be more likely to choose higher utility brands and will be less likely to choose lower utility brands. This finding implies that it will be easier for strong brands to retain consumers as “loyal” customers and harder for competitors to switch those consumers' loyalties. Our results also suggested that brand-equity-associated brand utility measures positively affect scale and that this finding was robust across several data sources. This result is consistent with the findings of Swait and Erdem (2002), who showed a positive impact of marketing mix consistency (theoretically linked to brand credibility) on scale for a frequently moving consumer good (FMCG) product. Thus, our results suggest that marketing managers should focus on building and maintaining strong brands that send consistent and credible product quality signals, which can build brand loyalty via increasing choice consistency. As Erdem and Swait (1998) noted, consumer brand equity exists regardless of a brand's quality position; hence, low- or high-quality brands can have relatively high equity with a consistent low-/high-quality position. Thus, focusing on consistent and credible signals should benefit any brand. The signaling perspective on equity also suggests that firms can build stronger utility brands by communicating a long-term commitment to them, again requiring consistency within and across marketing mix components. For example, consistency is likely to be particularly important for product line extension decisions that are implicated in SKU proliferation. That is, increased choice complexity may be associated with within-brand SKU proliferation because brands assumed to be strong from the firm's perspective can overstretch their signals via product line extensions. Thus, our work suggests that managers should ensure the consistency of quality signals generated by new extensions matches those of their parent brands so as not to confuse consumers and dilute brand signal (see also Wernerfelt (1988) for how umbrella branding may serve as a signal of a new product's quality). Prior research on “brand fits” when extending into new categories clearly suggest that consumers can become confused and not “see” the fit of some extensions unless great care is taken (Aaker and Keller, 1990 and Völckner and Sattler, 1996). Our work also contributes to the ongoing discussion about positioning private labels in retailing. In particular, recent research suggest that “consumers use their experience with one private label brand to update their beliefs about rival retailers' brands, and [that] these effects are quite sizeable” (Szymanowski & Gijsbrechts, 2012); hence, controlling consumer perceptions of signals from private labels should be more difficult as the number of categories with such labels increases. Even though Szymanowski and Gijsbrechts (2012) find that the consumption of one private label brand reduces consumers' uncertainty of others, they note that this result may be due to the similar positioning of the private labels in their study. Indeed, such a “familiarity spillover” may lead to more consumer uncertainty if consumers compare experiences with different quality tiers of private label brands. Thus, the recommendation of Geyskens, Gielens, and Gijsbrechts (2010), which cautions against cannibalizing effects of private label introductions (with variable levels of quality) being perceived as similar, might also reduce the perception of noisy signals by consumers. That is, retailers may wish to delink different quality-tier private labels by positioning them in different shelf areas or creating stand-alone brands rather than sub-brands under a retailers' name. Our work also provides important insights for the management of online consumer ratings. More than 60% of the respondents in the 2007 Forrester Research online survey reported that they sought user ratings, and as a response, consumer-review websites and e-commerce websites, such as http://www.yelp.com/, http://www.walmart.com/ and http://www.amazon.com/, make the distribution of ratings available on their websites (Sun, 2012). From an information economics perspective, these ratings provide important signals to consumers about a product's quality. Interestingly, heterogeneity in the ratings distribution does not necessarily mean that consumers receive noisier signals. Instead, Sun (2012) shows that such a high variance may enable consumers with niche preferences to find a better matching product. This finding suggests that, in the case of product ratings and reviews, signal noise has to be conceptualized at the preference segment level and that only signals within one segment should be consistent. Thus, marketing managers may find it beneficial to carefully monitor the variance of consumer reviews and to sometimes rely on reporting only averages rather than the full rating distribution. Our work also suggests future research opportunities. First, one should test if the effect of brand-equity-associated brand utility measures on scale also holds for choices in real markets. For example, one can combine data from experiments such as DCE 1 with data on actual choices reported in surveys or scanner panels (see, e.g., Horsky et al. 2006). Moreover, combining experiments such as DCE 1 with longitudinal purchase data would permit one to test some of the issues related to how brand-equity-associated brand utilities affect brand loyalty. Similarly, it would also be interesting and important to track the evolution of brand-equity-associated brand utility measures and their associated underlying constructs over time to determine whether and how different marketing campaigns associated with the brands involved affect each construct, as well as total brand utility. Finally, it would also be useful to determine in what ways and to what degree characteristics of consumers and product categories influence the impacts of brand signals. For example, the signaling process is less likely to be successful if a signal receiver is not looking for the signal (e.g., in markets where consumers have sufficient experience with the product offerings) or does not know what to look for (e.g., in completely new product categories) (see, e.g., Connelly, Certo, Ireland, & Reutzel, 2011). Research involving a broader range of categories may be able to determine these relationships.