Too Exciting to Fail, Too Sincere to Succeed: The Effects of Brand Personality on Sensory Disconfirmation

Abstract

Across four studies, the authors demonstrate that consumers intuitively link disconfirmation, specifically sensory disconfirmation (when touch disconfirms expectations by sight), to a brand’s personality. Negative disconfirmation is often associated with negative posttrial evaluations. However, the authors find that when negative sensory disconfirmation is introduced by an exciting brand, the source of disconfirmation can sometimes be perceived positively. This occurs because consumers intuitively view disconfirmation as more authentic of an exciting personality. Similarly, despite the wealth of literature linking positive disconfirmation to positive posttrial evaluations, the authors find that sensory confirmation is more preferred for sincere brands because consumers intuitively view confirmation as more authentic of a sincere personality. The authors conclude by demonstrating the intuitive nature of this phenomenon by showing that the lay belief linking brand personality to disconfirmation does not activate in a context where sensory disconfirmation encourages a more deliberative assessment of the product.

expectancy disconfirmation, brand personality, sensory marketing, touch

Touch comes before sight, before speech. It is the first language and the last, and it always tells the truth.

—Margaret Atwood (2000)

Extant research suggests that sensory marketing can be used subconsciously to trigger abstract notions of a product’s sophistication or quality ( Krishna 2012 ). For example, when Samsung and Nokia were coming out with cell phone cases made predominantly of plastic, Apple released phones that, despite looking plastic, were made with aluminum bands with chamfered edges wrapped around the case. Anyone who picked up the iPhone would immediately feel the temperature and weight distortion—a cooler, heavier feel—characteristic of metal ( Thorn 2013 ). This subtle conflict between initial expectations based on sight, and subsequent experience based on touch, leveraged a discrepancy in sensory inputs to augment perceptions of quality. Not surprisingly, the market was quick to respond. Nokia released its own aluminum phone, the Lumia, but despite rave reviews, the product underperformed. Although much of a product’s success or failure has to do with market dynamics, it is our contention that the success or failure of sensory marketing strategies may have a lot to do with how consumers perceive the brand.

A brand’s personality (i.e., a set of humanlike characteristics; Aaker 1997 ) is an important consideration when evaluating a product. Apple, for example, has been described as creative ( Fitzsimons, Chartrand, and Fitzsimons 2008 ), young ( Aaker 1997 ), and exciting ( Maehle, Otnes, and Supphellen 2011 ). Exciting brands are characterized as unique and attention getting ( Aaker 1997 ; Aaker, Fournier, and Brasel 2004 ). In contrast, Nokia has been described as sincere ( Van der Lans, Van den Bergh, and Dieleman 2014 ) and more down to earth ( Aaker et al. 2004 ). Sincere brands are characterized as consistent and trustworthy ( Aaker et al. 2004 ; Buss 1991 ; Robins, Caspi, and Moffitt 2000 ). Evidence suggests that conflicting cross-modal inputs from sight and touch often violate expectations and elicit surprise ( Ludden, Schifferstein, and Hekkert 2009 ; Noseworthy, Di Muro, and Murray 2014 ). Indeed, such conflict may be more fitting of an exciting personality, but it may also be unbecoming of a sincere personality. Thus it stands to reason that the effectiveness of sensory tactics may be highly subject to how consumers perceive the brand.

We explore the novel prediction that consumers intuitively link exciting brands to inconsistent actions, and sincere brands to consistent actions, and that marketers can use these lay beliefs to augment product preference. We do so in the context of what we refer to as sensory disconfirmation. A typical example is when touch reveals more inferior or superior quality than initially expected by sight. We find that exciting brands can benefit from sensory disconfirmation on a peripheral aspect of the product (e.g., on the packaging), even when the disconfirmation is perceived to be negative, whereas sincere brands benefit from peripheral sensory confirmation, more so than even positive peripheral sensory disconfirmation. Furthermore, we find that perceptions of authenticity underscore this phenomenon. That is, peripheral sensory disconfirmation is seen as more authentic of exciting brands, whereas peripheral sensory confirmation is seen as more authentic of sincere brands. Lastly, in isolating the intuitive nature of this phenomenon, we demonstrate that the link between sensory consistency and brand personality does not influence preference when the disconfirmation encourages a more deliberative assessment of the product’s functionality. Taken as a whole, the results suggest that consumers hold lay beliefs linking brand tactics to their broader preconceptions of the brand, and that these lay beliefs can be used to heighten brand preference.

CONCEPTUAL FRAMEWORK

Brand Personality

Extant research suggests that brands have personalities that serve a symbolic or self-expressive function ( Keller 1993 ). These personality dimensions form an inferential base by which consumers represent the brand as having a set of humanlike characteristics ( Aaker 1997 ; Johar, Sengupta, and Aaker 2005 ). By anthropomorphizing brands in this manner, consumers help create and sustain an intimate relationship with the brand ( Aaker 1997 ; Fournier 1994 ).

Brand personality consists of five dimensions: sincerity, excitement, competence, sophistication, and ruggedness ( Aaker 1997 ). These dimensions do not refer to the way consumers relate to the brand but rather how consumers view the brand ( Brakus, Schmitt, and Zarantonello 2009 ). Of the five brand personality dimensions, sincerity and excitement are considered fundamental to the marketing landscape because they constitute two of the three partner ideals in intimate personal relationships ( Aaker et al. 2004 ), and because they capture the majority of variance in brand personality ratings ( Aaker 1997 ; Capara, Barbaranelli, and Guido 2001 ). Aaker et al. (2004) note that, in general, consumers form stronger relationships with sincere brands (e.g., Hallmark, Ford, Coca-Cola) than with exciting brands (e.g., Yahoo, Virgin, MTV). This is consistent with the notion that exciting brands are more suitable for short-term relationships, whereas sincere brands are more suitable for long-term relationships ( Aaker 1997 ). Sincere brands are considered warmer (e.g., Gateway) and more down to earth (e.g., MetLife, Coke, and Pepsi), whereas exciting brands are associated with a younger demographic (e.g., Mountain Dew) and are viewed as possessing cultural vitality (e.g., BMW, Dr. Pepper). Hence it is not surprising that exciting brands are often viewed as attractive, appealing, and capable of generating interest and trial ( Altschiller 2000 ).

In essence, a brand’s personality influences how its actions and behaviors are perceived by the consumer ( Aaker et al. 2004 ). Such actions can be inferred from consumer reviews ( Dawes, Singer, and Lemons 1972 ), advertising ( Deighton 1984 ), past experiences ( Day 1977 ), and of particular interest in this article, how the product looks ( Bloch 1995 ; Chitturi, Raghunathan, and Mahajan 2007 ; Noseworthy and Trudel 2011 ). How a product looks helps maintain a brand’s standard and style, which is a key consideration in capturing the cultural essence of a brand in a given context ( Aaker, Benet-Martínez, and Garolera 2001 ).Yet, as the old adage goes, “looks can be deceiving.” This is certainly the case when it comes to products.

Sensory Disconfirmation

Consumption is often a function of combining multiple perceptions into a single integrated percept. Most products, at the minimum, incorporate visual and haptic information, and do so in temporal order—that is, consumers first see a product and then they subsequently interact with the product through touch. Such cross-modal perceptions need not align and indeed may often conflict ( Krishna 2012 ; Mattila and Wirtz 2001 ). This conflict arises from the need to validate sensory cues with other sensory inputs ( Gottfried and Dolan 2003 ). Researchers have explored cross-modal conflict between smell and sound ( Mattila and Wirtz 2001 ), sound and vision ( Russell 2002 ), sound and smell ( Spangenberg, Grohmann, and Sprott 2005 ), and touch and taste ( Krishna and Morrin 2008 ). To date, much of this work has focused on how cross-modal consistency can augment product evaluations ( Krishna, Elder, and Caldara 2010 ; Krishna and Morrin 2008 ). For example, Krishna et al. (2010) found that when cross-model sensory cues align (e.g., when something looks feminine and smells feminine), the overall information presented is processed faster and product evaluations increase. Indeed, one sensory input can augment the efficacy of the other ( Krishna et al. 2010 ). However, when cross-modal information is inconsistent, the conflicting sensory inputs can often lead to surprise ( Ludden et al. 2009 ; Noseworthy et al. 2014 ). The role of surprise in sensory inconsistency illuminates how sensory cues can set expectations.

When it comes to setting expectations, visual information often takes precedence. This phenomenon has a long history and has been demonstrated in judgments of an object’s curvature ( Gibson 1933 ), size ( Rock and Victor 1964 ), length ( Teghtsoonian and Teghtsoonian 1970 ), location ( Hay, Pick, and Ikeda 1965 ; Pick, Warren, and Hay 1969 ), depth ( Ho et al. 2009 ; Singer and Day 1969 ), and movement ( Klein and Posner 1974 ). According to the expectancy disconfirmation paradigm, when consumers interact with stimuli, they compare their experience with their initial expectations ( Diehl and Poynor 2010 ; Oliver and DeSarbo 1988 ; Tse and Wilton 1988 ). If their experience with a product exceeds their expectations, consumers often evaluate the experience as positive; in contrast, if the interaction falls short of their expectations, consumers often view the experience as negative ( Oliver and Swan 1989 ). However, there is reason to predict that there may be certain circumstances in which disconfirmation would not follow such a deliberative path to evaluation.

Research suggests that consumers process disconfirmation differently depending on the extent to which the expectancy violation challenges preconceptions of the product category. For instance, core violations alter attributes that are prototypical of a product category (e.g., altering the product’s shape, material, or core functionality), whereas peripheral violations alter attributes that are neither necessary nor sufficient for category membership (e.g., adding a feature that merely complements core functionality, adding a novel accessory, or altering packaging; Bagga, Noseworthy, and Dawar, forthcoming). For example, touching a seemingly plastic smartphone bumper case and finding that it is aluminum may indeed disconfirm sensory expectations, but it would not require that consumers consciously and deliberately reconfigure their beliefs about smartphones. By comparison, touching what looks like a plastic smartphone and finding that it is aluminum would require that consumers deliberately reassess the general belief that smartphones are plastic. Such reassessments tend to focus attention primarily on functionality (e.g., is this a more or less durable smartphone?). This is because core alterations are tightly coupled to functional subsystems, whereas peripheral alterations tend to be weakly coupled ( Gatignon et al. 2002 ). This fits the notion that consumers engage in more deliberative processing when cued to evaluate a product’s functionality ( Hoegg, Alba, and Dahl 2010 ; Page and Herr 2002 ), and that functionality is often inferred from a product’s form ( Veryzer 1995 ).

Given the preceding observations, it is not surprising that peripheral violations tend to encourage a more automatic and perceptual system, whereas core violations encourage a more deliberative, rule-based system ( Ashby and Alfonso-Reese 1998 ; Malt and Johnson 1992 ; Noseworthy and Goode 2011 ). In one of the few marketing examples, Noseworthy and Goode (2011) found that the deliberative processing of core product violations can override the implicit processing of peripheral inputs. Of course, much of this work has looked at core and peripheral attributes within a single product concept; however, of the work that has explored core design adjustments as built-in components, and peripheral design adjustments as physically detached from the product, Ma, Gill, and Jiang (2015) found consistent results in that really new peripheral adjustments fare better than core adjustments, in part because peripheral adjustments do not command a deliberate reassessment of the base product category.

Thus we predict that if sensory disconfirmation occurs on a peripheral aspect of a product (e.g., on the product’s case or packaging), consumers should process the disconfirmation in a more automatic manner. However, if sensory disconfirmation appears as part of the product’s core design, consumers should process the disconfirmation more deliberatively. In the latter case, we should see the classic disconfirmation effect, whereby positive disconfirmation is favored to negative disconfirmation ( Oliver and Swan 1989 ). This is consistent with the notion that changes in a product’s core materiality tend to motivate consumers to actively seek out additional information ( Peck and Childers 2003b ), and why such changes, when negative, often lead consumers to infer cost-saving motives ( Wilcox, Kim, and Sen 2009 ). Nevertheless, given that peripheral violations encourage automatic processing, this type of disconfirmation may not behave in accordance with traditional accounts.

A Lay Belief Linking Brand Personality to Disconfirmation

A wealth of literature suggests that social judgments, such as those that govern relationship norms, are often processed in a top-down manner that can influence the perceiver’s expectations about the events being attended to ( Fiske and Taylor 1984 ; Hastie 1981 ; Srull and Wyer 1989 ). In the context of sensory marketing, this suggests that consumer expectations can derive not only from the alignment of immediate subordinate sensory cues, but also from more abstract and superordinate representations of the brand. In one of the few articles to lend credence to the possibility that subordinate brand behavior interacts with superordinate representations of the brand, Aaker et al. (2004) found that consumers evaluate service transgressions differently depending on the brand’s personality. Specifically, in the case of sincere brands, transgressions provide contrary evidence that disconfirms expectations of the partner. These expectations are believed to stem from consumers inferring partner trustworthiness and dependability ( Aaker 1999 ). One surprising result, as noted by the authors, was that relationship strength improved for exciting brands following a transgression. The authors speculated that the transgression may have reinvigorated the exciting brand relationship, such that consumers seemingly expected, if not welcomed, some degree of disconfirmation in the relationship exchange. These findings support the distinct possibility that consumers intuitively associate sincere brands with more consistent and dependable actions, and exciting brands with more unpredictable and inconsistent actions. This suggests that as long as a sensory violation is not too severe as to coerce deliberative processing, exciting brands may in fact benefit from sensory disconfirmation, whereas sincere brands may benefit more from sensory confirmation.

The importance of an intuitive lay belief governing expectations of brand behavior is bolstered by evidence that consumers evaluate a brand more positively when the brand’s actions are consistent, as opposed to inconsistent, with the norms governing the brand relationship ( Aggarwal 2004 ). Such communal norms are processed automatically and held at a more abstract level than those of an exchange relationship ( Aggarwal and Law 2005 ; Maheswaran, Mackie, and Chaiken 1992 ). This is why brand personality can differentiate even when consumers cannot articulate differences in attributes or benefits, or even in the absence of sensory differences ( Aaker 1997 ). Critically, extant literature suggests that people are more likely to rely on implicit and automatic associations when intuition is relevant to the task at hand, and when they are not cued to engage in more deliberative processing ( Clore et al. 2001 ; Higgins 1998 ; Schwarz 2001 ; Sundar, Kardes, and Wright 2015 ). Indeed, this fits the notion that if consumers maintain an intuitive lay belief linking brand personality to brand actions, then this belief should be more accessible when the consumers are processing the brand’s behavior automatically. Therefore, if consumers process peripheral violations more automatically, and core violations more deliberatively, this lay belief would more likely activate in the former than in the latter.

HYPOTHESES

Given the wealth of literature documenting the classic disconfirmation effect, the focus of the current work is primarily on identifying the circumstances when a lay theory leads individuals to act is ways that are inconsistent with disconfirmation theory. Thus studies 1, 2, and 3 exclusively explore the case of peripheral violations. In study 4, we investigate the distinction between core and peripheral violations to highlight how deliberative processing can override the intuitive lay belief and subsequently manifest the classic disconfirmation effect. We begin with the basic prediction that consumers prefer peripheral sensory disconfirmation when induced by exciting brands but prefer peripheral sensory confirmation when introduced by sincere brands. Stated formally,

H1: There will be a significant interaction between brand personality and sensory consistency, such that a sincere brand will benefit more when touch and sight are aligned (sensory confirmation) than when unaligned, whereas an exciting brand will benefit more when touch and sight are unaligned (sensory disconfirmation) than when aligned.

It is believed that people develop lay theories or knowledge structures, regarding the nature of their social world, in order to interpret, predict, and control the broader social environment ( Lickel, Hamilton, and Sherman 2001 ). Prior research suggests that when the actions of a brand are consistent with the brand’s personality, consumers tend to interpret more authentic and legitimate brand behavior ( Muniz and O’Guinn 2001 ; Spiggle, Nguyen, and Caravella 2012 ). Brand authenticity is a unique part of brand identity and reflects how consumers maintain a brand’s essence ( Beverland 2005 ). Thus, for example, if consumers see exciting brands as unorthodox and unpredictable, and sincere brands as consistent and trustworthy ( Aaker 1997 ; Aaker et al. 2004 ; Buss 1991 ; Robins et al. 2000 ), they should see an exciting brand as more authentic when associated with sensory disconfirmation, and see a sincere brand as more authentic when associated with sensory confirmation. Critically, Spiggle et al. (2012) demonstrated that brand authenticity through cultural contiguity can augment purchase intent. This is consistent with the notion that consumers prefer a brand when the brand’s actions are consistent with the norms governing the brand relationship ( Aggarwal 2004 ), and with recent evidence that consumers not only attend to, but also desire cues that convey authenticity ( Beverland and Farrelly 2010 ). Thus we predict that peripheral sensory consistency will interact with brand personality via perceptions of brand authenticity. Stated formally,

H2: Perceptions of brand authenticity will mediate the effect of sensory consistency on brand preference, such that a sincere brand will be seen as more authentic and thus more desirable when touch and sight are aligned (sensory confirmation) than when unaligned, whereas an exciting brand will be seen as more authentic and thus more desirable when touch and sight are unaligned (sensory disconfirmation) than when aligned.

In the four studies that follow, we examine how consumers intuitively link brand personalities to disconfirmation within the sensory marketing context. In study 1 we find support for hypothesis 1. In particular, we demonstrate that cross-modal peripheral sensory confirmation is perceived positively when associated with a sincere brand, and cross-modal peripheral sensory disconfirmation is perceived positively when associated with an exciting brand. In study 2, we replicate study 1 in another field experiment and we further identify perceived authenticity as the mechanism driving brand preference. In study 3, we reverse the nature of disconfirmation in a more controlled setting to show how positive or negative disconfirmation is rewarded when the brand’s personality is exciting, but even positive disconfirmation does little for a sincere brand. Finally, in study 4, we offer an important boundary condition whereby we show that if the source of disconfirmation encourages more deliberative processing, the intuitive link between the brand personality and sensory consistency does not manifest.

STUDY 1

The core objective of study 1 was to examine whether the impact of peripheral sensory consistency (i.e., whether touch aligns with sight) on consumption preference is highly dependent on a brand’s personality. For the sake of simplicity, we use the term sensory confirmation to define a product that visually aligns with how it feels (i.e., when a bag of coffee both looks and feels like burlap), whereas we use the term sensory disconfirmation to define a product that visually does not align with how it feels (i.e., when a bag of coffee looks like burlap but feels like paper). Specifically, we were interested in whether sincere brands would suffer sensory violations, given that such violations may be viewed as inconsistent with a sincere personality. Conversely, exciting brands might benefit from sensory violations because such violations are more fitting of an exciting personality. Critically, we are not predicting that exciting brands can do exciting things; rather, we predict that exciting brands benefit from violating norms, even if, all else being equal, the violation is perceived to be of inferior quality.

Method

Participants and Design

Consumers (N = 207; 50.7% female; M_age  = 33.03) were recruited on location at a local coffee shop and given a $5 gift card for participating in the study. Participants were told that a new coffee product was being introduced by a local company and that this study was commissioned on behalf of the company to gain feedback about the product. The coffee bag was presented as a prototype of what might be considered. Participants were randomly assigned to one of six conditions in a 3 (brand personality: control [no brand personality manipulation] vs. sincere vs. exciting) × 2 (peripheral sensory consistency: sensory confirmation vs. sensory disconfirmation) between-subjects factorial design.

Stimuli: Brand Personality

Three versions of a website were created to manipulate the brand personality of a fictitious brand JAUNT. For the control we used a white background with Ariel font, and the web page included no pictures, content, or a brand tagline. For the sincere and exciting brands, we followed Aaker et al.’s ( 2004 ) detailed procedures. Specifically, we varied five key elements of the website: color (soft brown vs. bright red), visuals (sitting St. Bernard dog vs. jumping Dalmatian puppy), font (Comic Sans vs. Jester), content (family picnic vs. rock climbing), and tagline (“Because Life Is Too Meaningful to Let You Pass It By” vs. “Because Life Is too Exciting to Let You Pass It by”; appendix A).

A pretest ( n  = 98) was conducted to evaluate the effectiveness of the three web designs relative to the control using a validated scale for brand sincerity and brand excitement ( Aaker et al. 2004 ). Participants were asked to rate the degree to which each brand had sincerity traits (“sincere, wholesome, sentimental, family oriented”; anchored: 1 = Not at all, 7 = To a great extent; α = .81) and exciting traits (“exciting, unique, young, trendy”; anchored: 1 = Not at all, 7 = To a great extent; α = .92). The pretest confirmed that participants presented with the sincere website rated JAUNT as more sincere ( M  = 5.49) than either those presented with the control website ( M  = 4.77; F (1, 95) = 10.53, p  < .005) or the exciting website ( M  = 4.23; F (1, 95) = 32.17, p  < .001). Although not predicted, the exciting website was also seen as less sincere ( M  = 4.23) than the control ( M  = 4.77; F (1, 95) = 5.57, p  < .05). Similarly, participants presented with the exciting website rated JAUNT as more exciting ( M  = 4.90) than either those presented with the control website ( M  = 4.47; F (1, 95) = 3.86, p  = .052) or the sincere website ( M  = 4.06; F (1, 95) = 16.56, p  < .001). The sincere website was seen as less exciting ( M  = 4.06) than the control ( M  = 4.47; F (1, 95) = 4.37, p  < .05). It seemed that the brand manipulations not only augmented the predicted corresponding inferences, but they also reduced the alternative inferences (e.g., the exciting brand was seen as less sincere). Although not predicted, these findings are consistent with the notion of instability being insincere and stability being unexciting.

In a follow-up posttest ( n  = 100), we utilized the brand manipulations to explore whether consumers intuitively expect exciting brands to do different things, and expect sincere brands to maintain the status quo. That is, we wanted to confirm the assumption that consumers intuitively see inconsistency as an integral part of an exciting brand, but they see consistency as integral to sincere brands. To accomplish this, we first showed participants the brand personality manipulation and explicitly conveyed the product category. We then asked participants to describe in writing what they expect a new product in this category to be like from this brand. Participants were asked to focus on color, packaging, taste, and so on. Independent coders coded the thoughts that were prototypical for the coffee category as 1, and thoughts that were atypical of the coffee category as−1. We then created a consistency index by taking the number of prototypical (i.e., consistent) thoughts and subtracting the number of atypical (i.e., inconsistent) thoughts, divided by total thoughts. A positive number indicates greater thoughts of consistency, whereas a negative number indicates greater thoughts of inconsistency. The results confirmed that participants were more likely to list consistent thoughts when the brand was viewed as sincere ( M  = .92) and more likely to list inconsistent thoughts when the brand was viewed as exciting ( M  =−.97, t (97.6) = 8.03, p < . 001). The next step was to see what would occur if we could harness these general beliefs.

Stimuli: Sensory Consistency

Each coffee bag was custom designed, manufactured, and printed with the brand characteristics of the actual brand in mind (appendix A). The coffee bag consisted of either a finish that looked and felt like burlap (sensory confirmation) or a finish that looked like burlap but felt smooth like paper (sensory disconfirmation). All other information regarding the brand was identical between the two coffee bags.

To ensure that participants could perceive texture through incidental touch, we followed Picard’s (2006) specified procedure, in which pretest participants ( n  = 120) were presented with one of the two coffee bags. One group of participants was asked to review the bag only by sight, and the other group was asked to review the bag using both sight and touch. Participants were asked to indicate the extent to which they can discern three basic dimensions of tactile perception: softness, thickness, and relief (i.e., roughness; anchored: 1 = Very incapable; 7 = Very capable). The results revealed that when participants were asked to evaluate the product using sight alone, there was no difference in perceived tactile texture across smoothness ( M_conf  = 1.31 vs. M_disconf  = 1.50; p  = .16), thickness ( M_conf  = 2.15 vs. M_disconf  = 2.01; p  = .31) or relief ( M_conf  = 2.69 vs. M_disconf  = 2.83; p  = .14). However, when participants were asked to evaluate the product using sight and touch, there was a significant difference in perceived smoothness ( M_conf  = 1.73 vs. M_disconf  = 2.56; t (119) = 4.71, p < . 001), thickness ( M_conf  = 2.51 vs. M_disconf  = 1.83; t (119) = 3.92, p < . 001) and relief ( M_conf  = 2.40 vs. M_disconf  = 1.57, t (119) = 4.71, p < . 001). Further, in order to rule out the possibility that the haptic manipulation influenced perceptions of brand personality, we presented a separate group of participants ( n  = 120) with either one of the two coffee bags but with no brand description and asked them to speculate about the personality of the brand. In each case, the same measures used to pretest the brand manipulation were captured. The results revealed that there was no difference in perceived sincerity (α = .89; M_conf  = 3.55 vs. M_disconf  = 3.54; p  = .99) or excitement (α = .85; M_conf  = 3.25 vs. M_disconf  = 3.10; p  = .52).

Furthermore, to ensure we were not confounding sensory disconfirmation with perceived innovativeness—which may make for a relatively intuitive link to exciting brands—and to ensure we were manipulating negative disconfirmation, pretest participants were also asked to rate the innovativeness of the packaging (anchored: 1 = Strongly disagree, 7 = Strongly agree) on four 7 point items: “The coffee packaging is highly innovative”; “Relative to other brands in this product category, this coffee packaging is very innovative”; “The coffee packaging I reviewed is unique to its users”; and “The technology adopted in this coffee package is new and cutting edge” (adapted from Ali, Krapfel, and LaBahn 1995 ; Calantone, Chan, and Cui 2006 ), and to rate the overall quality of the product (anchored: 1 = Inferior quality, 7 = Superior quality; Kirmani and Wright 1989 ). The results revealed that perceived innovativeness did not vary between the coffee bags (α = .86; M_burlap  = 2.71 vs. M_paper  = 2.85; p  = .70). However, participants rated the quality of the paper coffee bag as generally inferior ( M  = 1.63) to that of the burlap coffee bag ( M  = 1.96; t (120) = 2.65, p < . 05). Feedback from participants indicated that the majority who rated the quality low viewed the paper coffee bag as cheap and gimmicky. Thus the manipulation of sensory consistency was carried forward into the main study.

Procedures and Dependent Measures

The study was conducted by three confederates, located at three different locations near a local coffee shop on a large university campus. Participants were presented with a tablet that directed them to JAUNT’s beta website (i.e., the brand personality manipulation) under the guise that we wanted consumers to get a feel for the company. Participants were then presented with a physical bag of coffee beans (i.e., the peripheral sensory consistency manipulation). They were first shown the coffee bag to set expectations by sight, and then they were encouraged to handle the product. After interacting with the coffee bag, participants were asked to refer back to the tablet and fill out a brief survey about the product. Purchase intent was captured on a three item, 7 point scale (“unlikely/likely,” “improbable/probable,” and “impossible/possible”; Chattopadhyay and Basu 1990 ). We also collected a manipulation check for brand personality (discussed in the pretest). The instrument concluded with basic demographic information.

Results and Discussion

Manipulation Check: Brand Personality

An analysis of sincerity (α = .86) and excitement (α = .89) as a function of brand personality and peripheral sensory consistency, yielded a main effect of brand personality, such that participants perceived the sincere brand to be more sincere ( M  = 4.64) than both the exciting brand ( M  = 3.88; F (1, 201) = 3.77, p  < .05), and the control brand ( M  = 4.28; F (1, 201) = 3.90, p  < .05). The exciting brand was seen as less sincere ( M  = 3.88) than the control ( M  = 4.28; F (1, 201) = 3.77, p  < .05). Similarly, participants perceived the exciting brand as more exciting ( M  = 4.78) than both the sincere brand ( M  = 4.30; F (1, 201) = 3.96, p  < .05) and the control brand ( M  = 4.37; F (1, 201) = 4.68, p  < .05). The difference between the control and the sincere brand was not significant ( p  = .76). Critically, no other main effect or interaction approached significance ( p’ s > .25).

Hypothesis Testing

We hypothesized that sincere brands will benefit more so from sensory confirmation than sensory disconfirmation, and exciting brands will benefit more so from sensory disconfirmation than sensory confirmation. In support of this, an analysis of purchase intention (α = .96), as a function of brand personality and peripheral sensory consistency, yielded a significant two-way interaction ( F (2, 201) = 10.86, p  < .005). Simple effects confirmed that when JAUNT was positioned as sincere, participants indicated greater intent to purchase in the sensory confirmation condition ( M  = 5.00) than in the sensory disconfirmation condition ( M  = 3.66; F (1, 201) = 15.43, p  < .001. Note that in this and all other comparisons, the degrees of freedom shown in text represent a pooled error term; thus they should not be mistaken for indication of how many individuals made up the comparison). Conversely, when JAUNT was positioned as exciting, participants indicated greater intent to purchase in the sensory disconfirmation condition ( M  = 4.48) than in the sensory confirmation condition ( M  = 3.63; F (1, 201) = 5.59, p  < .05). Not surprisingly, participants in the control condition generally favored confirmation ( M  = 4.70) to what was pretested as negative disconfirmation ( M  = 3.82, F (1, 201) = 5.86, p  < .05).

The results as just described provide initial evidence to suggest that the influence of peripheral sensory disconfirmation on product preference may be highly contingent on brand personality. If so, the question that needs to be answered is why? As theorized, we believed this may have something to do with perceptions of authenticity; specifically, with respect to how the sensory marketing tactic matches with broader preconceptions of the brand. Beverland (2005) noted that consistency of a brand’s standards, as manifested in production processes, components, ingredients, and other attributes, can signal legitimacy and reinforce (or undermine) a brand’s identity ( Beverland 2005 , 2006 , 2009 ). Although it has been speculated that inferior quality can cause consumers to perceive a brand as inauthentic ( Spiggle et al. 2012 ), we suggest that because consumers intuitively link disconfirmation to specific brand personalities, negative disconfirmation may actually enhance perceptions of authenticity, but only for exciting brands, whereas sensory confirmation will be seen as more authentic of a sincere brand. Therefore, as stated in hypothesis 2, we predict that brand authenticity will explain the observed asymmetric effects of peripheral sensory violations on brand preference.

STUDY 2

The core objective of study 2 was to isolate whether the phenomenon observed in study 1 was due to perceptions of authenticity. Specifically, we posit that perceptions of authenticity will predict preference for sincere brands that offer sensory confirmation rather than sensory disconfirmation, whereas authenticity will predict preference for exciting brands that offer sensory disconfirmation rather than sensory confirmation.

Method

Participants and Design

Participants (N = 176; 62.5% female; M_age  = 32.69) were recruited through mall intercept and given a $5 gift card for participating in the study. Participants were told that a new board game was being released and that this study was commissioned on behalf of the local company. The packaging was presented as a prototype of what might be considered. Participants were randomly assigned to one of six conditions in a 3 (brand personality: control vs. sincere vs. exciting) × 2 (peripheral sensory consistency: sensory confirmation vs. sensory disconfirmation) between-subjects factorial design.

Stimuli: Brand Personality

We created three versions of a pamphlet to manipulate the brand personality of a fictitious brand YULE. Each pamphlet was exactly 72 words in length and designed to be similar in content (adapted from Aaker et al. 2004 ). Key differences were written to operationalize brand personality as sincere or exciting (appendix B). A control condition was included whereby participants read a generic commentary about the brand. A pretest ( n  = 100) confirmed that participants presented with the sincere pamphlet rated YULE as more sincere ( M  = 5.40) than either those presented with the control pamphlet ( M  = 4.72; F (1, 97) = 8.02, p  < .01) or the exciting pamphlet ( M  = 4.23; F (1, 97) = 32.71, p  < .001). The exciting pamphlet was seen as less sincere ( M  = 4.23) than the control ( M  = 4.72; F (1, 97) = 4.47, p  < .05). Similarly, participants presented with the exciting pamphlet rated YULE as more exciting ( M  = 4.88) than either those presented with the control pamphlet ( M  = 4.41; F (1, 97) = 4.50, p  < .05) or the sincere pamphlet ( M  = 4.00; F (1, 97) = 16.36, p  < .001). The sincere pamphlet was seen as less exciting ( M  = 4.00) than the control ( M  = 4.41; F (1, 97) = 3.14, p  = .08).

Similar to study 1, a follow-up posttest ( n  = 103) using the same brand manipulations revealed that participants were more likely to list consistent (i.e., prototypical) thoughts when describing a new product from the sincere brand ( M  = .44) and more likely to list inconsistent (i.e., atypical) thoughts when describing a new product from the exciting brand ( M  = − .51, t (91.92) = 5.69, p < . 01). Thus given the brand manipulations worked as intended, we incorporated the three pamphlets into the main study.

Stimuli: Sensory Consistency

The YULE game boxes were custom designed and manufactured with actual dominos game pieces inside (appendix B). All game information was adopted from a comparable game on the market. The boxes consisted of either a finish that looked and felt like soft cork (sensory confirmation) or a finish that looked like cork but felt like cheap plastic (sensory disconfirmation). All other information regarding the brand and content was identical. We conducted a pretest ( n  = 120) for haptic perception using the same procedures and measures discussed in study 1. The results revealed that when participants were asked to evaluate the product using sight alone, there was no difference in perceived tactile texture across smoothness ( M_conf  = 1.57 vs. M_disconf  = 1.80; p  = .14), thickness ( M_conf  = 2.30 vs. M_disconf  = 2.20; p  = .47), or relief ( M_conf  = 1.21 vs. M_disconf  = 1.26; p  = .61). However, when participants were asked to evaluate the product using sight and touch, there was a significant difference in perceived smoothness ( M_conf  = 1.42 vs. M_disconf  = 2.47; t (119) = 2.70, p < . 001), thickness ( M_conf  = 2.68 vs. M_disconf  = 1.11; t (119) = 6.96, p < . 001), and relief ( M_conf  = 2.77 vs. M_disconf  = 1.73; t (119) = 6.17, p < . 001). Furthermore, as expected, when participants evaluated the product using sight and touch, they rated the quality of the plastic case (sensory disconfirmation) as generally inferior ( M  = 1.69) to that of the cork game box (sensory confirmation; M  = 2.07; t (119) = 2.94, p < . 001). Lastly, in order to rule out the possibility of the haptic manipulation influencing perceptions of brand personality, we presented participants ( n  = 118) with one of the two packages with no brand description and asked them to speculate about the personality of the brand. The results revealed no difference in perceived sincerity (α = .88; M_conf  = 3.93 vs. M_disconf  = 4.22; p  = .21) or excitement (α = .89; M_conf  = 3.23 vs. M_disconf  = 2.96; p  = .26). Thus the two stimuli were carried forward to the study.

Procedures and Dependent Measures

The study was conducted by two confederates, located at two different locations in a mall. Participants were informed that a local game company was conducting a study to evaluate the market’s response to a new boxed game. Each participant was provided with a pamphlet of the company’s profile and asked to review the pamphlet to get a feel for the company’s mission and goals (i.e., the brand personality manipulation). Participants were then presented with the game box and then asked to review the product including the dominos pieces and the game instructions inside. This procedure ensured that people handled the product after seeing it. Participants were then asked to rate purchase intention and brand personality (described in study 1). In addition, brand authenticity was captured on a 12 item, 7 point scale (1 = Strongly disagree, 7 = Strongly agree; Spiggle et al. 2012 ; appendix C). Although several alternate measures of brand authenticity exist (e.g., Morhart et al. 2015 ; Napoli et al. 2014 ), we selected Spiggle et al. (2012) , because several of their subscale dimensions tap not only the degree to which the brand’s behavior matches the brand’s standard and style, but also whether individuals were processing whether the brand was engaged in activities solely for commercial gain (both of which are highly influenced by sensory information). The instrument concluded with basic demographic information.

Results

Manipulation Check: Brand Personality

An analysis of sincerity (α = .89) and excitement (α = .88) as a function of brand personality and peripheral sensory consistency yielded only a main effect of brand personality ( F (2, 170) = 9.72, p  < .05); participants perceived the sincere brand to be more sincere ( M  = 5.21) than both the control ( M  = 4.81; F (1, 170) = 5.29 p  < .05) and the exciting brand ( M  = 4.27; F (1, 170) = 35.68, p  < .001). The exciting brand was seen as less sincere ( M  = 4.27) than the control ( M  = 4.81; F (1, 170) = 6.22, p  < .05). Similarly, participants perceived the exciting brand to be more exciting ( M  = 4.96) than both the control ( M  = 4.45; F (1, 170) = 4.76, p  < .05) and the sincere brand ( M  = 4.00; F (1, 170) = 14.25, p  < .01). The sincere brand was seen as less exciting ( M  = 4.00) than the control ( M  = 4.45; F (1, 170) = 3.70, p  < .05). Once again, no other main effect or interaction approached significance ( p’ s > .17).

Hypothesis Testing

An analysis of purchase intention (α = .93), as a function of brand personality and peripheral sensory consistency, yielded a significant two-way interaction ( F (2, 170) = 5.74, p  < .005). Consistent with the pretest and study 1, simple effects in the control revealed that confirmation was generally preferred to negative disconfirmation ( M_conf  = 5.16 vs. M_disconf  = 4.42; F (1, 170) = 4.04, p  < .05; figure 1 ). This effect was slightly exacerbated when the brand personality was sincere ( M_conf  = 5.28 vs. M_disconf  = 4.43; F (1, 170) = 5.37, p  < .05). However, when brand personality was exciting, participants indicated greater intent to purchase the game box in the sensory disconfirmation condition ( M  = 5.14) than in the sensory confirmation condition ( M  = 4.41; F (1, 170) = 3.94, p  < .05). It was predicted that these effects may have something to do with perceptions of authenticity.

FIGURE 1

PURCHASE INTENT RESULTS FOR STUDY 1

An analysis of the 12 item brand authenticity scale (α = .70), as a function of brand personality and peripheral sensory consistency, yielded a significant two-way interaction ( F (2, 170) = 6.32, p  < .05). Simple effects confirmed that when brand personality was sincere, participants indicated higher perceptions of authenticity in the sensory confirmation condition ( M  = 4.50) than in the sensory disconfirmation condition ( M  = 4.17; F (1, 170) = 4.10, p  < .05). Conversely, when the brand personality was exciting, participants indicated higher perceptions of authenticity in the sensory disconfirmation condition ( M  = 4.47) than in the sensory confirmation condition ( M  = 4.05; F (1, 170) = 6.42, p  < .05). There was no significant difference in the control ( M_conf  = 4.47 vs. M_disconf  = 4.21, p  = .11). To determine whether brand authenticity accounted for participants’ purchase intention, we conducted a mediated moderation analysis ( Hayes 2012 ; model 8; bootstrapped with 10,000 draws). Consistent with the preceding information, we found no effect of brand authenticity in the control condition (95% confidence interval [CI],−.52 to .03). However, as predicted, authenticity mediated the increase in purchase intent for confirmation in the sincere brand condition (95% CI, .10–70), and it mediated the increase in purchase intent for disconfirmation in the exciting brand condition (95% CI, −.13 to −.29).

Discussion

Study 2 replicated the findings from study 1, suggesting that peripheral sensory confirmation indeed favors sincere brands, and peripheral sensory disconfirmation favors exciting brands. Importantly, as hypothesized, we found evidence that brand authenticity mediates these effects, whereby exciting brands are seen as more authentic when they engage in sensory disconfirmation, and sincere brands are seen as more authentic when they engage in sensory confirmation. In both instances, the sensory tactic intuitively aligned with the brand’s personality and thus consumer preference increased.

Up to this point, we have shown how an exciting brand can benefit from peripheral sensory disconfirmation, even when that disconfirmation is negative. However, we have yet to explore fully the role that sincerity plays in predicting preference for sensory violations. In both studies 1 and 2, the sincere condition mirrored the control. Thus there is the question of whether sincere brands truly benefit from confirmation or whether, as one might expect, they merely suffer negative disconfirmation. There is a way to test whether this is indeed the case.

As discussed, disconfirmation can be positive. For example, if an object looks like it is made of plastic, a consumer may be delighted to find out that it is actually made of metal after touching it. It is in this respect that touch can reveal more rare or desirable material ( Gentry et al. 2001 ). Of course, there is a wealth of research supporting the benefits of positive disconfirmation ( Peck and Childers 2003a ; Peck and Wiggins 2006 ). However, if this is truly about consumers intuitively linking disconfirmation to exciting brands and confirmation to sincere brands, and if evaluations are more positive when a brand’s actions are consistent with the norms governing the brand relationship ( Aggarwal 2004 ), then the valence of the disconfirmation should not matter. This is critical because it would suggest that, at least when processed automatically, sincere brands may not benefit from even positive disconfirmation. Stated formally,

H3a: A sincere brand will benefit more from sensory confirmation than from either positive or negative sensory disconfirmation.

H3b: An exciting brand will benefit more from positive or negative sensory disconfirmation than from sensory confirmation.

STUDY 3

The objective of study 3 was to evaluate whether reversing the valence of the disconfirmation would reverse the effects on brand preference. To evaluate this, we created three sets of stimuli representing peripheral sensory confirmation (where sight and touch align), negative sensory disconfirmation (where touch reveals less quality than expected by sight), and positive sensory disconfirmation (where touch reveal more quality than expected by sight).

Method

Participants, Design, and Procedure

Students (N = 276; 66.3% female; M_age  = 22.19) were recruited in a behavioral lab and participated in exchange for course credit. Participants were told that a retailer was interested in their evaluations of a new gourmet popcorn being targeted to the participants’ specific demographic. Participants were randomly assigned to one of nine conditions in a 3 (brand personality: control vs. sincere vs. exciting) × 3 (peripheral sensory consistency: sensory confirmation vs. positive sensory disconfirmation vs. negative sensory disconfirmation) between-subjects factorial design.

Participants were seated at a computer with the product next to it. They were then asked to pick up the product and review it at their leisure, and once they believed they could make a valid assessment, they were to provide input about the product to the retailer. The survey began with a brief introduction about the product on a secure website. The website showcased a promotional advertisement (the brand personality manipulation). Using the same measures discussed in study 2, participants were then asked to list their purchase intention, perceived authenticity, inferences of brand personality, and brief demographics.

Stimuli: Brand Personality

We created three versions of a web page to manipulate the brand personality of a fictitious brand MAKKA. Similar to study 1, we manipulated brand personality by adjusting key design elements ( Aaker et al. 2004 ; appendix D). Once again, we conducted a pretest ( n  = 93) to evaluate the effectiveness of the three web pages using the previously discussed scales for brand sincerity (α = .84) and brand excitement (α = .87). The pretest confirmed that participants presented with the sincere website rated MAKKA as more sincere ( M  = 5.24) than those presented with either the control website ( M  = 4.63; F (1, 90) = 5.07, p  < .05), or the exciting website ( M  = 4.27; F (1, 90) = 19.63, p  < 0.01). The difference between the exciting manipulation and the control condition was not significant ( p  = .14). Participants presented with the exciting website rated MAKKA as more exciting ( M  = 5.11) than those presented with either the control website ( M  = 4.46; F (1, 90) = 6.78, p  < .05) or the sincere website ( M  = 4.15; F (1, 90) = 13.05, p  < .01). The difference between the sincere manipulation and the control condition was not significant ( p  = .24). Once again, a follow-up posttest ( n  = 100) revealed that participants were more likely to list consistent (i.e., prototypical) thoughts when describing a new product from the sincere brand ( M  = .93) and more likely to list inconsistent (i.e., atypical) thoughts when describing a new product from the exciting brand ( M  = −.86; t (101) = 8.41, p < . 001). Thus the three websites were incorporated into the main study.

Stimuli: Sensory Consistency

The MAKKA packaging was custom designed and manufactured with real popcorn inside (appendix D). All packaging information was consistent with similar products on the market. Given that the plastic condition in study 2 represented the inferior condition, we strategically built a stimulus where the plastic condition served as confirmation in that pretest participants generally expected the product to be plastic (i.e., it looked like a plastic tube). This allowed us to rule out the possibility that inferior inferences were being tagged to the material as opposed to the disconfirmation process. The positive disconfirmation product looked identical to the plastic product but consisted of an aluminum finish that felt heavy and cold to the touch, and the negative disconfirmation product looked identical to the plastic product but consisted of a paper finish that felt rather thin and flimsy. All other information regarding the brand and content was identical among the products.

We conducted a pretest ( n  = 120) to establish the sensory consistency manipulation using the same procedures and measures discussed in the previous studies. The results revealed that when participants were asked to evaluate the product using sight alone, there was no difference in perceived tactile texture across smoothness ( M_conf  = 2.55 vs. M_pos-disconf  = 2.60 vs. M_neg-disconf  = 2.20; p’ s > .22), thickness ( M_conf  = 3.37 vs. M_pos-disconf  = 3.57 vs. M_neg-disconf  = 3.52; p’ s > .44) or relief ( M_conf  = 4.75 vs. M_pos-disconf  = 4.52 vs. M_neg-disconf  = 4.57; p’ s > .36). However, when participants were asked to evaluate the product using sight and touch, there was a significant difference in perceived smoothness ( M_conf  = 3.52 vs. M_pos-disconf  = 4.47 vs. M_neg-disconf  = 2.52; p’ s  <  .001), thickness ( M_conf  = 3.65 vs. M_pos-disconf  = 2.85 vs. M_neg-disconf  = 2.45; p’ s  <  .05), and relief ( M_conf  = 1.45 vs. M_pos-disconf  = 1.25 vs. M_neg-disconf  = .90; p’ s  <  .05). Furthermore, the results revealed that the form of the package in the three material variations did not vary on perceived innovativeness (α = .92; M_neg-disconf  = 3.37 vs. M_pos-disconf  = 3.35 vs. M_conf  = 3.05; p’ s > .47). In fact, the motivation to adopt a novel form (i.e., a cylinder tube) for popcorn was to maintain a standard perception of innovativeness while allowing for plausible sensory violations. Critically, when participants were asked to evaluate the product using sight and touch, participants rated the quality of the negative disconfirmation manipulation as inferior ( M  = 3.50) to that of the confirmation manipulation ( M  = 4.62; F (1, 119) = 43.39, p < . 001), and, in turn, rated the quality of the confirmation manipulation as inferior ( M  = 4.62) to that of positive disconfirmation manipulation ( M  = 5.18; F (1, 119) = 86.29, p < . 001). Lastly, the results revealed that the sensory violation alone caused no discernable differences in perceived sincerity (α = .88; M_neg-disconf  = 3.18 vs. M_pos-disconf  = 2.95 vs. M_conf  = 3.23; p’ s > .33) or excitement (α = .87; M_neg-disconf  = 3.37 vs. M_pos-disconf  = 3.35 vs. M_conf  = 3.05; p’ s > .55).

Results

Manipulation Check: Brand Personality

An analysis of sincerity (α = .86) and excitement (α = .83) as a function of brand personality and peripheral sensory consistency, yielded only a main effect of brand personality ( F (2, 267) = 7.25, p  < .05); participants perceived the sincere brand to be more sincere ( M  = 4.52) than both the control ( M  = 4.19; F (1, 267) = 4.50, p  < .05) and the exciting brand ( M  = 3.86; F (1, 267) = 4.04, p  < .05). The exciting brand was seen as less sincere ( M  = 3.86) than the control ( M  = 4.19; F (1, 267) = 4.29, p  < .05). Similarly, participants perceived the exciting brand to be more exciting ( M  = 3.73) than both the control ( M  = 3.39; F (1, 267) = 4.57, p  < .05) and sincere brand ( M  = 2.99; F (1, 267) = 5.21, p  < .05). The sincere brand was seen as less exciting ( M  = 2.99) than the control ( M  = 3.39; F (1, 267) = 5.21, p  < .05). No other main effects or interactions approached significance ( p’ s > .12).

Purchase Intentions

An analysis of purchase intention (α = .95), as a function of brand personality and peripheral sensory consistency, yielded a significant two-way interaction ( F (2, 267) = 5.33, p  < .01). As illustrated in figure 2 , simple effects confirmed that participants indicated greater intent to purchase the sincere gourmet popcorn in the sensory confirmation condition ( M  = 3.57) than in either the positive sensory disconfirmation ( M  = 2.84; F (1, 267) = 4.43, p  < .05) or negative sensory disconfirmation conditions ( M  = 2.51; F (1, 267) = 11.63, p  < .001). The difference between positive and negative disconfirmation was not significant ( p  = .31). Conversely, participants indicated greater intent to purchase the exciting gourmet popcorn in both the positive ( M  = 3.41) and negative ( M  = 3.30) sensory disconfirmation conditions than in the sensory confirmation condition ( M  = 2.62; F (1, 267) = 5.37, p  < .05, and F (1, 267) = 4.89, p  < .05, respectively). The difference between positive and negative disconfirmation conditions was not significant ( p  = .71). As expected, in the control condition, positive disconfirmation ( M  = 3.64) was preferred more so than either confirmation ( M  = 3.00; F (1, 267) = 2.93, p  < .05) or negative disconfirmation ( M  = 2.53; F (1, 267) = 10.36, p  < .01), and confirmation ( M  = 3.00) was preferred more than negative disconfirmation ( M  = 2.53; F (1, 267) = 5.30, p  < .05).

FIGURE 2

PURCHASE INTENT RESULTS FOR STUDY 3

Brand Authenticity

An analysis of brand authenticity (α = .81), as a function of brand personality and peripheral sensory consistency, yielded a significant two-way interaction ( F (2, 267) = 3.67, p  < .01). Simple effects confirmed that when brand personality was sincere, participants indicated higher perceived authenticity in the sensory confirmation condition ( M  = 4.08) than in either the positive sensory disconfirmation condition ( M  = 3.58; F (1, 267) = 5.76, p  < .05) or the negative sensory disconfirmation condition ( M  = 3.66; F (1, 267) = 5.21, p  < .001). The difference between positive and negative sensory disconfirmation was not significant ( p  = .67). Conversely, when brand personality was exciting, participants perceived higher brand authenticity in the positive sensory disconfirmation condition ( M  = 4.08) and negative sensory disconfirmation condition ( M  = 4.04) than in the sensory confirmation condition ( M  = 3.62; F (1, 267) = 5.53, p  < .05, and F (1, 267) = 5.32, p  < .02, respectively). The difference between the positive and negative sensory disconfirmation was again not significant ( p  = .81). We conducted a follow-up mediated moderation analysis ( Hayes 2012 ; model 8; bootstrapped with 10,000 draws). Consistent with study 2, authenticity did not mediate purchase intent for either positive or negative sensory disconfirmation relative to sensory confirmation in the brand control condition (positive: 95% CI,−22 to .29; negative: 95% CI, −.53 to .01). However, as predicted, authenticity mediated the increase in purchase intent for sensory confirmation relative to both positive and negative sensory disconfirmation in the sincere brand condition (positive: 95% CI, .80–18; negative: 95% CI, .03–42). Conversely, authenticity mediated the increase in purchase intent for both positive and negative sensory disconfirmation relative to sensory confirmation in the exciting brand condition (95% CI, −.15 to −.09 and 95% CI, −.03 to −.41, respectively).

Discussion

The results from study 3 further support the prediction that exciting brands benefit from sensory disconfirmation more so than sensory confirmation, regardless of the valence of the disconfirmation. Conversely, and rather noteworthy, sincere brands benefit from sensory confirmation more so than sensory disconfirmation, even if that disconfirmation is positive. Moreover, we further corroborate that perceptions of authenticity mediate these effects. Critically, in support of the notion of a lay theory processed at an automatic level, the difference between positive and negative disconfirmation only manifested in the control condition, where no brand personality was activated.

Given that study 3 established the link between brand personality and sensory consistency, we next wanted to isolate the fundamental assumption that these effects only occur with peripheral violations. As outlined previously, like most lay beliefs, this phenomenon should exclusively operate at an automatic level ( Evans and Stanovich 2013 ; Raghunathan and Suher 2012 ). Given that core violations tend to encourage a deliberate assessment of an object’s utility ( Ashby and Alfonso-Reese 1998 ; Bagga et al., forthcoming ; Malt and Johnson 1992 ), and given that conscious assessments tend to override automatic association that coincide with peripheral violations ( Noseworthy and Goode 2011 ), a core violation should render intuitive judgments as less influential. Thus we expect that core disconfirmation will behave the way the literature on disconfirmation predicts (i.e., people will prefer positive disconfirmation to negative disconfirmation). Conversely, we expect that peripheral disconfirmation will be processed more automatically, and thus preference will be contingent on the consumer’s broader preconceptions of the brand (replicating the previous three studies). Stated formally,

H4: Type of disconfirmation will moderate the influence of brand personality on sensory consistency, such that the previously stated findings will only manifest when the source of disconfirmation occurs on a peripheral aspect of the product. When the source of disconfirmation appears on the core product concept, consumers will prefer positive disconfirmation to negative disconfirmation, regardless of the brand’s personality.

STUDY 4

The goal of study 4 was to test whether sensory violations that appear directly on the product (i.e., a core violation) would encourage more deliberative processing of the product’s functionality, rendering intuitive judgments less influential. This predication is predicated on the notion that core violations command a more deliberative assessment, which overrides the automatic associations that would otherwise coincide with peripheral violations ( Noseworthy and Goode 2011 ). To accomplish this, we selected a peripheral manipulation that mirrored the packaging manipulations from our prior studies (a smartphone bumper case), and we pitted this against the exact same manipulation on the product itself (a smartphone).Critically, we strategically selected a peripheral manipulation that was still inherently functional (i.e., bumper cases are built to protect smartphones). In this respect, we sought to ensure that the observed effects were not being relegated to things that have no bearing on utility, but more so on things that do not encourage a deliberate assessment of the target product.

Method

Participants and Design

Participants (N = 400; 55.8% female; M_age  = 36.19) were recruited by mall intercept and were compensated $5 for their participation. Participants were randomly assigned to a 2 (brand personality: sincere vs. exciting) × 3 (sensory consistency: sensory confirmation vs. positive sensory disconfirmation vs. negative sensory disconfirmation) × 2 (type of disconfirmation: peripheral vs. core) fractional factorial design. The fractional nature of the design was such that only 10 of the 12 full factorial conditions were collected. This was done because the peripheral confirmation condition was redundant with the core confirmation condition (both the phone and case looked and felt exactly as one would expect). Moreover, we excluded the brand personality control to further simplify the design.

Stimuli: Brand Personality

We employed an alternative manipulation of brand personality from Aaker et al. (2004) , whereby we selected existing brands that retained the desired characteristics. In step with our opening dialogue about Apple altering the material in their phones to elicit stronger inferences of quality, we used Nokia and Apple as the sincere and exciting brands, respectively. Doing so afforded the additional benefit of allowing us to use an iconic form as our base of manipulation without having the phone be more strongly linked to one specific brand. Specifically, we adopted a form based on the Nokia Lumia 925 and the iPhone 6, which look almost identical ( http://forums.windowscentral.com/phone-wars/309824-doesnt-new-iphone-6-especially-black-one-look-like-lumia-925-a.html ). A pretest ( n  = 90) confirmed that with minor adjustments to the width of the phone, we were able to develop a prototype that was no more associated with Apple ( M  = 4.27) than with Nokia ( M  = 4.22; t (89) = .31, p  = .75).

Beyond allowing for prototypical expectations for the product’s form, the use of real brands afforded an alternative means of tapping brand personality without requiring that we contrive a web page or pamphlet. The selection of Apple and Nokia was confirmed via a pretest ( n  = 103) of several smartphone manufacturers. As in the prior studies, participants were asked to rate the degree to which each brand had sincerity traits (α = .95) and exciting traits (α = .84). The results confirmed that participants perceived Nokia as more sincere ( M  = 4.77) than Apple ( M  = 3.91; t (100.7) = 2.95, p  < .005), and perceived Apple as more exciting ( M  = 5.51) than Nokia ( M  = 3.99; t (89.38) = 6.60, p  < .001). Finally, as a measure of the extent to which consumers’ beliefs about sensory consistency relate to their perceptions of the brand, we included an open-ended question where participants were asked to describe how they picture a new product from each respective brand. The results confirmed that participants were more likely to associate an inconsistent design with Apple ( M  = −.94) and a more consistent design with Nokia ( M  = .95; t (96.7) = 8.70, p  < .001). Thus adopting these two brands not only served to tap the core constructs of interest, but it also further solidified the general expectation that people inherently link exciting brands to inconsistency and sincere brands to consistency.

Stimuli: Peripheral Sensory Consistency

Evidence from our pretest debriefs revealed that people tend to perceive a smartphone bumper case as peripheral accessories for the smartphone category. Moreover, it seemed that people have strong prototypical expectations that these cases are made predominantly of plastic (or silicone). Given this insight, we had three different versions of bumper cases manufactured as a means to operationalize incidental sensory consistency (appendix E). The cases consisted of a finish that pretested as either looking like and feeling like plastic (sensory confirmation), looking like plastic but feeling like aluminum (positive sensory disconfirmation), or looking like plastic but feeling like rough cardboard fabric (negative sensory disconfirmation). All visual information was identical across the three cases. As with the prior studies, all pretesting confirmed that discernable differences in perceived tactile perception only occurred when participants were permitted to handle the case ( p’ s < .001). Moreover, the cases did not vary in perceived innovativeness (α = .86; p’ s > .16); nor did the cases independently alter perceptions of sincerity (α = .87; p’ s > .26) or excitement (α = .91; p’ s > .33). Critically, participants rated the cardboard fiber case as generally inferior ( M  = 3.68) to the plastic case ( M  = 4.06; F (1, 115) = 4.59, p < . 05), and rated the plastic case as inferior to the metal case ( M  = 4.55; F (1, 115) = 9.10, p < . 005). Thus the brand logos were printed on the three smartphone bumper cases, and the cases were carried forward to the study.

Stimuli: Core Sensory Consistency

Our manipulation of core sensory consistency followed the same procedures as the peripheral manipulation with the one exception that the phone itself was made of either plastic (sensory confirmation), aluminum (positive sensory disconfirmation), or cardboard fiber (negative sensory disconfirmation). The phones were presented as hardware prototypes of a new smartphone concept (i.e., there was no working software interface; appendix E). Participants were told that we were interested in exploring their thoughts and first impression of the new concept. The lack of a software interface was unavoidable given the difficulties with crating working prototypes. Nevertheless, the lack of explicit functionality allowed us to isolate whether core violations would cause people to question what the functionality would be like. Also, from a pragmatic perspective, this discouraged participants from estimating their preference based on the software platform. All visual information was identical across the three phones.

Once again, discernable differences in perceived tactile perception only emerged when participants were permitted to handle the phone ( p’ s < .001). Moreover, the phones did not vary in perceived innovativeness (α = .84; p’ s > .18) nor did they independently alter perceptions of sincerity (α = .79; p’ s > .27) or excitement (α = .84; p’ s > .16). As with the cases, participants rated the cardboard fiber phone as generally inferior ( M  = 2.63) to the plastic phone ( M  = 3.48; F (1, 115) = 27.10, p < . 001), and they rated the plastic phone as inferior to the metal phone ( M  = 4.62; F (1, 115) = 48.40, p < . 001). Thus the brand logos were printed on the three smartphone prototypes, and they too were carried forward to the main study.

Procedures

The study was conducted by three confederates located at three different locations in a mall. Participants were told that the study aimed to explore a new smartphone designed to reach all segments of the market. The smartphones were first presented intact with their bumpers (appendix E). Participants were told that the product will ship with the bumper as a free accessory to promote the product (emphasizing the marketing tactic). Participants were then encouraged to remove the bumper and get a feel for the phone itself. This procedure ensured that participants handled both the bumper and the phone after viewing both together. Participants were then directed to an electronic questionnaire. Participants were asked to assume that they came across this smartphone in the mall. Using the same measures discussed in study 2, participants were asked to rate perceptions of brand personality, list their perceived authenticity, and then rate their purchase intent. Additionally, to ascertain whether functionality weighs more heavily when sensory disconfirmation is directly linked to the product, participants filled out a measure of perceived functional efficacy on three 7 point items (“Please indicate the extent to which the product . . . [is likely to perform poorly/well; will offer few/a lot of advantages; will add a little/a lot of value]; Mukherjee and Hoyer 2001 ; Thompson, Hamilton, and Rust 2005 ). The survey instrument concluded with basic demographic information.

Results

Manipulation Check: Brand Personality

An analysis of sincerity (α = .71) and excitement (α = .76) as a function of brand personality, sensory consistency, and type of disconfirmation yielded only a main effect of brand personality; participants perceived Nokia to be more sincere ( M  = 4.21) than Apple ( M  = 3.84; F (1, 390) = 9.76, p  < .001). Conversely, participants perceived Apple to be more exciting ( M  = 4.55) than Nokia ( M  = 4.23; F (1, 390) = 4.18, p  < .05). Once again, no other main effects or interactions were significant ( p’ s > .11).

Purchase Intentions

Given the fractional nature of the design, we first analyzed purchase intent (α = .98) when there was a variation in the peripheral sensory consistency condition (i.e., variation in the smartphone case with no variation in the phone). The results yielded a significant second-order brand personality × sensory consistency interaction ( F (2, 234) = 19.03, p  < .001). Simple effects confirmed that Nokia was preferred more in the sensory confirmation condition ( M  = 4.29) than in either the positive ( M  = 3.22; F (1, 234) = 8.10, p  < .005) or the negative ( M  = 2.60; F (1, 234) = 19.24, p  < .001) sensory disconfirmation conditions. The difference between positive disconfirmation and negative disconfirmation was marginal ( p  = .08). Conversely, Apple was preferred more in the positive ( M  = 4.23) and negative ( M  = 4.01) sensory disconfirmation conditions than in the sensory confirmation condition ( M  = 2.54; F (1, 234) = 17.03, p  < .001, and F (1, 234) = 12.54, p  < .001, respectively). The difference between the positive and negative disconfirmation conditions was not significant ( p  = .61).

As predicted, the results took on a very different pattern of effects in the core sensory consistency condition (i.e., variation in the phone but no variation in the case). Consistent with disconfirmation theory, there was a main effect of sensory consistency on purchase intent such that participants generally preferred positive disconfirmation ( M  = 4.43) to confirmation ( M  = 3.41; F (1, 234) = 14.47, p  < .001), and preferred confirmation to negative disconfirmation ( M  = 2.71; F (1, 234) = 5.67, p  < .01). Although not predicted, this main effect was qualified by a significant brand personality × sensory consistency interaction ( F (2, 234) = 12.35, p  < .001). The nature of the interaction was such that the same pattern of effects held across Apple and Nokia; however, the preference for positive disconfirmation over confirmation was more pronounced for Apple than Nokia ( table 1 ). Nevertheless, despite the attenuation, the hypothesized (hypothesis 4) classic disconfirmation effect held across both brands.

Table 1

Means, Standard Deviations, and Cell Counts for Study 4

	Control		Peripheral sensory disconfirmation				Core sensory disconfirmation
	Confirmation		Positive disconfirmation		Negative disconfirmation		Positive disconfirmation		Negative disconfirmation
	Nokia	Apple	Nokia	Apple	Nokia	Apple	Nokia	Apple	Nokia	Apple
Sincere personality	4.13 (1.26)	3.79 (1.13)	4.15 (1.15)	3.96 (0.97)	4.27 (0.98)	4.00 (1.28)	4.43 (0.89)	3.74 (1.12)	4.06 (0.89)	3.71 (0.86)
Exciting personality	4.05 (1.31)	4.61 (1.11)	4.38 (1.16)	4.74 (1.23)	3.95 (1.12)	4.41 (1.21)	4.56 (1.10)	4.65 (1.26)	4.20 (1.05)	4.35 (0.94)
Functionality	4.42 (1.67)	3.64 (1.78)	4.84 (1.51)	5.23 (1.27)	4.43 (2.08)	4.73 (1.52)	4.64 (1.70)	5.33 (1.09)	2.93 (1.42)	3.69 (1.48)
Purchase intent	4.29 (1.82) a, b	2.54 (1.77) c, d	3.22 (1.53) a	4.23 (1.79) c	2.60 (1.62) b	4.01 (1.98) d	4.53 (1.35)	4.34 (1.51)	2.28 (1.45)	3.10 (1.78)
Authenticity	4.27 (0.63) e, f	3.42 (0.85) g, h	3.83 (0.61) e	4.12 (0.88) g	3.22 (0.74) f	4.06 (0.84) h	3.93 (0.85)	3.51 (0.78)	3.74 (0.76)	3.34 (0.83)
Cell size	40	40	40	40	40	40	43	40	37	40

NOTE.—Standard deviations are reported in parentheses; superscript represents the key significant simple effects replicating the prior studies.

Brand Authenticity

As with the purchase intent results, we first analyzed brand authenticity (α = .72) when there was variation in the peripheral sensory consistency condition. The results yielded a significant second-order brand personality × sensory consistency interaction ( F (2, 234) = 24.77, p  < .001). Simple effects confirmed that Nokia was seen as more authentic in the sensory confirmation condition ( M  = 4.27) than in either the positive ( M  = 3.83; F (1, 234) = 10.07, p  < .001) or the negative ( M  = 3.22; F (1, 234) = 46.69, p  < .001) sensory disconfirmation conditions. The positive disconfirmation condition was seen as more authentic ( M  = 3.83) than the negative disconfirmation condition ( M  = 3.22; F (1, 234) = 16.18, p  < .01). Conversely, Apple was seen as more authentic in both the positive ( M  = 4.12) and negative ( M  = 4.06) sensory disconfirmation conditions than in the sensory confirmation condition ( M  = 3.42; F (1, 234) = 12.23, p  < .001, and F (1, 234) = 10.68, p  < .001, respectively). As with the purchase intent results, the difference in perceived authenticity between the positive and negative disconfirmation conditions was not significant ( p  = .75).

Not surprisingly, the results of brand authenticity in the core sensory consistency condition revealed no significant second-order brand personality × sensory consistency interaction ( F (2, 234) = 1.99, p  = .14). The only effect that occurred was a main effect of sensory consistency ( F (2, 234) = 2.92, p  = .055). The nature of the effect was such that confirmation was seen as more authentic ( M  = 4.27) than either the positive ( M  = 3.93) or negative ( M  = 3.74) sensory disconfirmation conditions ( F (1, 234) = 4.13, p  < .05, and F (1, 234) = 11.53, p  < .001, respectively). The positive and negative disconfirmation conditions did not differ ( p  = .15).

To determine whether brand authenticity accounted for participants’ purchase intention, we once again conducted a follow-up mediated moderation analysis ( Hayes 2012 ; model 8; bootstrapped with 10,000 draws). Specifically, looking exclusively at the peripheral sensory consistency condition, authenticity mediated the increase in purchase intent for confirmation relative to both positive and negative disconfirmation for Nokia (positive: 95% CI, .36–1.20; negative: 95% CI, .07–1.14). Conversely, authenticity mediated the increase in purchase intent for both positive and negative disconfirmation relative to confirmation for Apple (95% CI,−.89 to −.109 and 5% CI, −.15 to −.40, respectively). As expected, when we explored the core sensory consistency condition, authenticity did not mediate purchase intent for either positive or negative sensory disconfirmation relative to confirmation for either Nokia (positive 95% CI, −.34 to 2.36; negative: 95% CI, −.14–1.60) or Apple (positive 95% CI, −.81 to .008; negative: 95% CI, −.94 to .11).

Functional Inferences

We then turned our focus to the underlying role of functional inferences. We first analyzed functional inferences (α = .98) when there was a variation in the peripheral sensory consistency condition. The results revealed no significant main effects or interactions ( p’ s > .12). This was expected given the observation that authenticity underlies the influence of sensory disconfirmation on brand preference when the disconfirmation is processed more automatically. However, it was predicted that concerns about the product’s functionality would underlie the classic disconfirmation effect in the core sensory consistency condition. In support of this prediction, and in step with the purchase intent results, there was a significant main effect of sensory consistency ( F (2, 234) = 23.58, p  < .001). Pairwise comparisons confirmed that consumers felt the phone would function better in the positive disconfirmation condition ( M  = 4.97) than in the confirmation condition ( M  = 4.03; F (1, 234) = 13.73, p  < .001), and, in turn, they felt the phone would function better in the confirmation condition than in the negative disconfirmation ( M  = 3.32; F (1, 234) = 7.41, p  < .01). As with the purchase intent results, this effect was qualified by a significant brand personality × sensory consistency interaction ( F (2, 234) = 6.31, p  < .005). Once again, the nature of the interaction was such that the same pattern of effects held across both brands, but the discrepancy between positive disconfirmation and confirmation was more pronounced for Apple than for Nokia ( table 1 ).

When perceived functionality was added to the previous mediation model, the results revealed no effects in the peripheral sensory consistency condition for either Nokia (positive 95% CI,−.16 to .14; negative: 95% CI, −.46 to .02) or Apple (positive 95% CI, −.43 to .03; negative: 95% CI, −.42 to .14). This was expected given the prediction that perceived functionality would only play a role in consumer preference when the sensory violation was directly linked to the product concept. Consistent with this prediction, perceived functionality mediated the observed main effect of sensory consistency on purchase intent in the core sensory consistency condition (95% CI, .03–.45). No other effects approached significance.

Discussion

Study 4 revealed an important and relevant boundary condition. By manipulating whether the source of disconfirmation appeared on the core product concept or on a peripheral aspect of the product, we were able attenuate the combined influence of brand personality and sensory consistency on brand preference, and we did so in a predictable manner that is consistent with the literature on disconfirmation. Specifically, in accordance with the previous studies, when sensory violations appeared on an peripheral aspect of the product (i.e., on the case), consumers indicated higher purchase intent for a sincere brand accompanied by confirmation rather than disconfirmation (even positive disconfirmation), and indicated higher purchase intent for an exciting brand accompanied by disconfirmation rather than confirmation (even negative disconfirmation, which critically was favored no less than positive disconfirmation). Authenticity once again mediated these effects. However, when sensory violations appeared on the core product concept (i.e., the phone), brand personality had little effect on purchase intent, and the results mirrored exactly what one would expect from the disconfirmation literature—positive disconfirmation was favored overall to negative disconfirmation. Unsurprisingly, concerns about the product’s functionality mediated this effect. Thus the results of study 4 suggest that the relationship between sensory disconfirmation and brand personality only intuitively manifests when the disconfirmation does not call for a deliberate assessment of the product’s functionality.

GENERAL DISCUSSION

Across four studies, we demonstrate that the success of a given sensory marketing tactic is highly contingent on how consumers perceive the brand. The findings show that people can perceive negative disconfirmation as positive when it is associated with exciting brands. This occurs because, regardless of positive or negative outcome, the violation is viewed as more authentic of an exciting personality. Similarly, despite the wealth of literature linking positive disconfirmation to positive post trial evaluations, we find that if the brand is perceived as sincere, positive disconfirmation fails to aid the brand. This occurs because, regardless of the positive outcome, the violation is perceived as less authentic of a sincere personality. Critically, in isolating the intuitive nature of this association, we find that the relationship between sensory violations and brand personality only augments product preference when the violation does not command a deliberate assessment of the product’s functionality.

Research on brand personality has focused on the connection between brand personality and the consumer. Fournier and Alvarez (2013) called for research on the negativity in brand relationships. Much of the work on the role of personality in consumer behavior has focused on the development of models of brand personality and, in particular, on the relationship between brand and human personalities ( Aaker 1997 ). Several research streams have demonstrated that branding and advertising shape the image of the brand in the minds of consumers. One area of research has focused specifically on the connections between brand personalities and consumer personalities (i.e., using the big five traits; Fournier and Alvarez 2013 ; Keller 1993 ). We contribute to this area by integrating the brand personality literature into sensory marketing. Researchers have focused on a single dimension of visual branding (e.g., the use of color, font, or images to express brand personality; Aaker et al. 2004 ). This research is the first to employ cross-modal sensory cues as a means to leverage preconceptions of the brand.

As a whole, we offer some of the first evidence to suggest that consumers possess a lay theory linking exciting brands to inconsistent actions and sincere brands to consistent actions. Consistent with the literature suggesting that brand authenticity drives preference for products with higher brand congruence ( Morhart et al. 2015 ), we find that preference coincides with inferences of authenticity. This lay theory suggests that, like people, the actions of brands are evaluated relative to consumers’ broader expectations of the brand’s behavior.

Like most lay theories, our findings are predicated on automatic judgments. When the violation commanded a more deliberative mindset, we observe the classic disconfirmation effect, whereby positive disconfirmation is favored to negative disconfirmation ( Oliver and DeSarbo 1988 ; Oliver and Swan 1989 ; Tse and Wilton 1988 ). In the current research, however, we demonstrate that when individuals make automatic judgments linking negative disconfirmation to sources other than the object or event that is being disconfirmed, even negative disconfirmation can result in positive outcomes. In such cases, the intuitive belief can be more important than the valence of the outcome. Furthermore, past research using the expectancy disconfirmation paradigm has investigated the role of violating expectations in service relationships. We further this stream of research and explore the paradigm in a context in which consumers had little or no experience with the product. In such cases, consumers use different strategies to create meaningful experiences with the product.

Beyond the theoretical implications, this article offers critical insights for brand marketers and product and package designers. Prior research has advised marketers to focus on personality branding in the way that connects the brand to the consumer. For example, Sundar and Noseworthy (2014) demonstrated that visual cues on packaging can influence consumption decisions. In this research, we demonstrate how marketers can leverage expectancy disconfirmation to increase brand preference. In a recent call for further research, Krishna (2012) highlights the potential uses of cross-modal violations to enhance consumer memory and even arousal. Although a stream of research has explored when touch is salient and diagnostic in purchases ( Peck and Childers 2003a ), the current research suggests that sensory marketing can be used beyond improving consumer experience or adding to the experience of a brand. Specifically, by meeting expectations set by a brand’s personality, downstream effects of desirability can be influenced. In this way we integrate literature on sensory marketing to illuminate how managers can leverage disconfirmation to their advantage. Critically, we also demonstrate a cautionary note for managers who have sustained a strong image of sincerity. In such cases, tactics used to elicit surprise or interest could backfire given their inconsistency with the brand’s broader image.

Of course, this work is not without its limitations and unanswered questions. One question revolves around the brand’s personality itself. In none of the studies did sensory disconfirmation augment perceptions of excitement for exciting brands; nor did sensory confirmation augment perceptions of sincerity for sincere brands. We believe this may have something to do with the automatic processing of baseline expectations. As noted in our pretesting, if consumers associate a sincere (exciting) brand with consistent (inconsistent) behavior, activating this association will not augment perceptions of sincerity (excitement), inasmuch as asking someone a benign question and them answering would not necessarily augment perceptions of honesty. One would need to be in a scenario where the probability of inconsistency or consistency would be anticipated. That said, on occasion we did observe directional support for half of the effect (i.e., sincere brands were sometimes seen as less sincere when engaged in disconfirmation). The interaction washed because the exciting brands were not seen as more exciting when they engaged in disconfirmation. What might have contributed to this were the painstaking efforts undertaken to develop stimuli that were not seen as exciting. We did this to guard against the possibility that we were merely tapping whether exciting brands are normatively linked to exciting things.

On a related note, the only empirical inconsistency to emerge across the four studies was that we sometimes observed a reversal in inference for each respective brand personality condition. For example, in study 2 the brand positioned as exciting was not only seen as more exciting than the control—as intended, but it was also seen as less sincere. The latter was an unanticipated reversal. Although the data generally trended toward this reversal for both the sincere and exciting brand manipulations, these contrasts were only sometimes significant. Nevertheless, such a reversal fits the notion that the two broader personalities used in these studies are somewhat antithetical. Critically, the predicted interaction held regardless, but we do believe this could lend some interesting insights into how different personality traits interact.

Another question that emerged from this research is whether the observed effects generalize to other sensory inputs beyond sight and touch. We chose this specific combination primarily because it represents one of the most common forms of disconfirmation at the point of purchase, and from a more pragmatic perspective, sight and touch were the easiest to manipulate without introducing confounds. That said, we were cognizant that sight has a privileged position when it comes to setting expectations ( Gibson 1933 ; Hay et al. 1965 ; Ho et al. 2009 ; Klein and Posner 1974 ; Pick et al. 1969 ; Rock and Victor 1964 ; Singer and Day 1969 ). With this in mind, we initially attempted to manipulate more than one sensory combination (e.g., sight and scent). Although we generally found a similar pattern of effects, the creation of the stimuli and execution of the experiments proved rather challenging. In particular, we often struggled to isolate a sensory violation from category ambiguity (i.e., scents are often associated with a different category, whereas material is often category neutral). The consequence was that different inferences kept creeping into the data, leading to severe variance. It did seem, however, that regardless of which sensory cue was used for disconfirmation, the violation must be processed at the automatic level for this effect to hold. It is in this respect that some senses, like taste, would prove more problematic than others, given that taste is more integral to the functional efficacy of ingestible products.

The line of reasoning just described raises another question about the extent to which the observed effects may hold under instances of severe negative peripheral disconfirmation. Our results suggest a possible tipping point where the violation would command a deliberate assessment. That said, we did try to push the limits with some of our stimuli while staying within the realm of plausibility. For example, the negative disconfirmation condition in our popcorn study (study 3) was so flimsy that participants often crushed the stimuli just by handling it, and the oil residue from the popcorn eventually permeated through the paper packaging, requiring that we recreate the stimuli every few rounds of data collection. The challenge with going more severe—as initially attempted with scent—was that consumers often inferred that the violation was the result of indirect contact or an after-market defect. Participants, in essence, became more deliberate in their processing in an effort to rationalize the violation as opposed to accessing their intuition about the brand’s intent. Future research could explore the specific boundaries around the various sensory combinations. Indeed, the latitude of acceptance may be greater from some cues, like sight, as opposed to other more adaptive and emotionally laden cues, like scent.

Last, in accordance with the preceding, there were some additional insights to further support the intuitive nature of this lay belief. There was little to no movement in the brand exploitation subscale for authenticity, which pertains to whether the manipulation was seen as a tactic solely for commercial gain. This is interesting, given that we were specifically altering materiality, which is often linked to perceived cost savings ( Wilcox et al. 2009 ). The only instance where we witnessed movement on the brand exploitation scale was when there was no brand manipulation in the first three studies (i.e., in the controls), and in study 4 when the manipulation directly appeared on the product (note: the other subscales showed little movement in this condition). In these two circumstances, participants tended to infer, as one would expect, that negative disconfirmation was done purely to save money. Furthermore, the finding that authenticity did not move at all in the core condition in study 4 fits the notion that consumers were intuitively linking the peripheral tactic to the brand but linking the core tactic to the product. Although not predicted, these findings corroborate the notion that more deliberative assessments of the product rendered automatic judgments to be less influential. This raises not only a very interesting substantive question around exploiting cost efficiencies in accordance with a brand’s personality, but it also raises some interesting questions regarding the extent to which other forms of disconfirmation would achieve similar results. Future research could go a long way in exploring the implications of resolving various marketing tactics within a broader preconceptions of the brand. We certainly encourage more work in this domain.

DATA COLLECTION INFORMATION

The data were collected at Virginia Tech, Blacksburg, VA, and in Cincinnati, OH (Strategic Intelligence Research Services [SIRS]). The data were collected at Virginia Tech was done so by the first author’s research assistants and PhD students, supervised by the lab manager at Virginia Tech. Data collected by SIRS were approved by the institutional review board (IRB) at Virginia Tech and was collected by the employees at SIRS. All studies were IRB approved at Virginia Tech and by the University of Oregon. The first author with the help/support of the second author analyzed the data.

Appendix A

Brand Personality and Sensory Consistency Manipulation for Study 1

Appendix B

Brand Personality Manipulation for Study 2

Control

YULE is made for every possible audience. The YULE brand has everything you’d expect in a regular brand. This brand promises a complete gaming experience. YULE is a leading maker of board games, card games, and puzzles. Since their first year of operation in 1995, YULE has forged a close relationship with one of the world’s largest game companies. As a result, many long-lived board games are now made exclusively by YULE!

Sincere Manipulation

YULE is made for a heartfelt audience. The YULE brand is considered candid, honest, and a sensible brand. This practical brand promises a fulfilling experience. YULE is a leading maker of board games, card games, and puzzles. Since their first year of operation in 1995, YULE has forged a close relationship with one of the world’s largest game companies. As a result, many long-lived board games are now made exclusively by YULE!

Exciting Manipulation

YULE is made for a thrill-seeking audience. The YULE brand is considered to be electrifying, stylish, and one-of-a-kind. This sensational brand promises a breathtaking experience. YULE is a leading maker of board games, card games, and puzzles. Since their first year of operation in 1995, YULE has forged a close relationship with one of the world’s largest game companies. As a result, many long-lived board games are now made exclusively by YULE!

Sensory Consistency Manipulation for Study 2

Appendix C

^a Note: Originally, all 16-items were adopted from Spiggle et al. (2012) . However given that the scale consistently achieved poor loadings (α = .43 to .63) in all studies, 12 items were used in the studies. These 12-items corresponded with the final scale used by Spiggle et al. (2012) .

• Maintaining Brand Standards and Style :

• The standards of YULE are apparently contained in this product ^a

• The style of this product seems to reflect that of YULE ^a

• This product appears to reflect the quality I associate with YULE ^a

• This product appears to be inferior to what I expect from YULE [reverse coded]

• Honing Brand Heritage

• YULE appears to connect with what I know about its origins ^a

• There is no link between YULE and what I know about YULE’s legacy [reverse coded] ^a

• YULE seems to have abandoned its roots with this extension [reverse coded] ^a

• Preserving Brand Essence

• This product is not consistent with my image of YULE [reverse coded] ^a

• This product seems to embody the essence of YULE

• This product preserves what YULE means to me ^a

• This product captures what makes YULE unique to me ^a

• What make YULE distinct to me is lost in this product [reverse coded]

• The key association I have with YULE are missing in this product [reverse coded]

• Avoiding Brand Exploitation

• This product likely trades off the essence of YULE strictly for profit [reverse coded] ^a

• This product likely sacrifices what I think makes YULE special in exchange for commercial gain [reverse coded] ^a

• With this product, it seems that YULE was more concerned about preserving the brand rather than growing the market ^a

Appendix D

Brand Personality and Sensory Consistency Manipulation for Study 3

Appendix E

Sensory Consistency Manipulation for Study 4

Author notes

This work is the result of equal-authorship collaboration. Aparna Sundar is an assistant professor of marketing in the Lundquist College of Business, University of Oregon, Eugene, OR ( [email protected] ). Theodore J. Noseworthy is an associate professor of marketing and Canada Research Chair in Entrepreneurial Innovation and the Public Good in the Schulich School of Business, York University, 99 Ian Macdonald Blvd., Toronto, ON M3J 1P3, Canada ( [email protected] ). The authors would like to thank the Dean Walton at the Science Library at the University of Oregon and Jeffrey Garman at the University of Oregon machine shop for their help in developing the stimuli. This research was partially funded by the Social Sciences and Humanities Research Council of Canada (SSHRC Insights Grant #435‐2013‐0235). Please address correspondence to Aparna Sundar.