Silent persuasion. Incidental use of promotional merchandise benefits unfamiliar brands

Abstract Many brands distribute promotional merchandise like branded pens, flash drives or mugs in the hopes that it will improve consumers’ reactions to the brand. Yet, consumers often use them purely incidentally, i.e. they casually use them and pay no attention to the brand, such as when borrowing a pen from a colleague. We query whether even such incidental use of promotional merchandise affects consumer response. Drawing on the specifics of merchandise use and a combination of prior insights on persuasion, we suggest that incidental use of promotional merchandise affects in particular unfamiliar brands. In two controlled lab experiments, we find that consumers react more positively to an unfamiliar brand after incidentally using its merchandise. The effect emerges even though consumers do not consciously notice the brand’s logo but does not extend to a familiar brand. Equivalent incidental visual exposure to advertising stimuli, i.e., posters, was less effective than the haptic incidental use of promotional merchandise. Durable promotional merchandise may be a worthwhile investment, in particular for unfamiliar brands.


Introduction
People often borrow someone else's pen when, for instance, taking notes during a meeting or signing a credit card receipt. They drink coffee from any mug available in the coffee kitchen at their work place and they use or borrow whatever flash drive they can find when they want to share files. All of these are instances where an item, be it a pen, a mug or a flash drive, are used incidentally. That is, consumers temporarily use these items for a specific purpose but the item itself is secondary and consumers do not pay much attention to it. Notably, all of these items often are promotional merchandise, that is 'useful or decorative items of merchandise that are imprinted with an advertiser's identification, message, and/or logo' (Cooper 2009, 376) and distributed free of charge (Nelson and Kanso 2002).
Company spendings on promotional merchandise have been constantly rising and in 2018, US companies spent $24.7 billion on this particular communication tool (Promotional Products Association International 2019). In contrast to other, more costly, advertising tools, promotional merchandise is used by a broad range of companies, 'from the street market or corner shop to the largest multinationals' (Rudzki and Li 2007, 181). Nearly every household and office contains some promotional merchandise. A study in the US, for instance, showed that 89% of consumers have received some sort of promotional merchandise in the previous six months (Das 2017;Promotional Products Association International 2017). Most consumers use promotional merchandise regularly and, as highlighted in the introductory examples, pass it on to others (Promotional Products Association International 2017). However, using promotional merchandise does not equate to noticing that one just has used an item meant to promote a brand. Consumers often receive and use promotional merchandise incidentally. They may pay little to no attention to the message on it, may use it only briefly for a specific purpose and, if asked, might not be able to tell what exactly the item they just used was like. Given the prevalence of incidental use of promotional merchandise, the question arises whether advertisers will benefit from promotional merchandise also in such situations of incidental use. Existing literature has addressed what happens when consumers receive merchandise as a gift (e.g. Kendrick 1998;McCarthy and Fram 2008). This equates to contexts of deliberate exposure to the merchandise and the brand it promotes. Whether merchandise can affect consumer responses to a promoted brand, even when used once and incidentally, is still an open question.
We aim to advance current insights on promotional merchandise by addressing this question. Specifically, we make several theoretically and practically relevant contributions. We propose and show that promotional merchandise can subtly persuade consumers even if they pay no attention to the logo featured. This is an important extension to the known persuasive powers of merchandise, which are currently restricted to their reception as a gift (e.g. Kendrick 1998;McCarthy and Fram 2008).
Additional contributions arise for the theoretical basis our propositions build on. We theorize that the incidental effectiveness of merchandise arises because it uniquely combines elements of incidental exposure (e.g. Fitzsimons et al. 2002;Janiszewski 1993;Wilson 1979) and haptic interaction (Brasel and Gips 2014;Wolf, Arkes, and Muhanna 2008). We add to both these streams of literature. To research on incidental exposure (Fitzsimons et al. 2002;Janiszewski 1993;Wilson 1979) we add the insight that even surrogate (a logo on the merchandise) single exposure can improve responses to a presented stimulus. This insight simultaneously advances the literature on haptic interaction. Specifically, we show that even a logo on a stimulus can benefit from haptic interaction (Brasel and Gips 2014;Wolf et al. 2008).
This matters practically because among the different promotional opportunities, merchandise is uniquely suited for haptic interaction. Accordingly, we additionally propose and show in our studies that incidental use of promotional merchandise may outperform purely visual incidental exposure in the form of posters. This is an important practical insight when deciding about how to spend limited advertising budgets.
Finally, we also provide practitioners with insights on when the persuasive power of incidental use of merchandise may be limited. Drawing on literature suggesting that the existence of prior attitudes hinders persuasion (Cacioppo et al. 1992;Gibson 2008), we propose and show that in particular unfamiliar brands may benefit from promotional merchandise.
Together, these insights advance our understanding of the effects of promotional merchandise by providing new insights on the frequent use case of incidental use. We briefly discuss prior insights on the effectiveness of promotional merchandise, before developing our propositions and testing them in two controlled laboratory experiments and discussing the ensuing theoretical and practical implications.

Theoretical background
Despite its prevalence, research on whether, how, and when promotional merchandise is effective is scarce (Buil, de Chernatony, and Montaner 2013;McCarthy and Fram 2008) and marketing textbooks tend to dedicate little space to the topic. For example, Kotler's and Armstrong's 654 pages 2016 edition of 'Principles of marketing' only devotes seven lines to the topic of promotional merchandise (Kotler and Armstrong 2016).
The existing evidence suggests that promotional merchandise renders consumers' reactions to a brand more positive. In a field experiment, Beltramini (2000) found that gifting existing customers with promotional merchandise increases actual purchase behavior as a function of its price level. Similarly, McCarthy and Fram (2008) showed that customers who received a promotional pen had a higher willingness to engage with the promoted brand in the future. Kendrick (1998) likewise found that promotional pens and mugs are as effective as coupons when it comes to stimulating sales while avoiding the so-called 'coupon trap', i.e. purchases cease once coupons are no longer available. Yet other research found that customers evaluate key product attributes more favorably after receiving promotional leather business card cases (Beltramini 1992).
Notably, all of this research has focused on situations where respondents were deliberately handed promotional merchandise in the course of a promotion and had the chance to repeatedly interact with the goods received (Beltramini 1992(Beltramini , 2000Kendrick 1998;McCarthy and Fram 2008). This procedure mirrors a situation in which recipients first encounter a brand and then get the promotional merchandise from the brand directly. In such settings, promotional merchandise may be viewed as a gift that elicits a desire for reciprocity (Cialdini and Goldstein 2004) and when keeping the merchandise in their homes people may learn to like the brand they are repeatedly exposed to (Bonanno and Stillings 1986;Hekkert, Thurgood, and Whitfield 2013).
Existing research on promotional merchandise thus allows conclusions about its effectiveness in situations in which people engage with the merchandise and are conscious of receiving it from the provider it promotes. However, oftentimes, promotional merchandise comes into people's possession in more mysterious and incidental ways and it does not necessarily remain there for long. For example, people often borrow a pen for a single note or use a promotional mug only once when offered a drink at some other person's place. Promotional merchandise often changes hands when used to fulfill its functional purpose. This entails that it is also often used incidentally. Research to date has not yet studied such situations of purely incidental use of merchandise. It is unclear whether promotional merchandise may also affect reactions to the featured brand when used once and incidentally outside of any promotional context. The primary mechanisms thought to make merchandise effective, namely a tendency to reciprocate the brand for the gift received and repeated exposure to this gift (Falk and Fischbacher 2006), are unlikely to work in such settings.
We investigate whether even in such settings, promotional merchandise can help improve consumers' reactions to a brand. To answer this question, we draw on the specifics of incidental merchandise use explicated by the common situation of one person borrowing a promotional pen from another person to take notes. In such a situation, the borrower ends up using promotional merchandise but is unlikely to devote attention to the promoted brand, i.e. the logo on the pen. The person's primary purpose is to take notes and she likely only looks at the pen to the extent necessary for the performance of this task. There is thus also a limited amount of visual exposure to the pen and through it to the logo it features. While the visual exposure is limited, the physical handling of the pen adds another dimension of interaction. Through the use of merchandise, consumers come "in touch" with the promotional item and thus also the logo on it. To deduce whether a promoted brand is likely to benefit from such incidental use of its merchandise we consult theories speaking to both elements of interaction, i.e. scant and incidental visual exposure and indirect touch via the medium of the merchandise.
Visual incidental exposure is a common occurrence even beyond merchandise. An essential part of consumers' decision processes occurs without attention and awareness (for reviews see Bargh 2002;Fitzsimons et al. 2002). According to the so called mere exposure effect even incidental exposure to a stimulus may affect responses to a stimulus (Zajonc 2001). This is because exposure to a stimulus can sensitize the brain to that stimulus and enable its learning, regardless of whether the stimulus is salient or relevant to the target task (Huang and Watanabe 2012;Watanabe, N añez, and Sasaki 2001). Incidental exposure to a stimulus can generate attitudes towards it (e.g. Ferguson and Zayas 2009;Janiszewski 1988;Murphy and Zajonc 1993) and influence behavioral responses to that stimulus (e.g. Strahan, Spencer, and Zanna 2002;Winkielman, Berridge, and Wilbarger 2005). This happens even if people are 'not conscious of the stimulus and thus unaware of having evaluated it' (Ferguson and Zayas 2009, 362).
However, not every incidental exposure improves the evaluation of what people are being exposed to. The literature hints at several contingencies that put into question whether incidental use of promotional merchandise would improve people's evaluation of the brand promoted on the merchandise. Specifically, there are several important differences between incidental exposure to promotional merchandise and the settings commonly found to lead to persuasion via incidental exposure.
Most importantly, the vast majority of research on incidental persuasion documents the necessity for repeated exposure to a stimulus (e.g. Matthes, Schemer, and Wirth 2007). This is the reason why the mere exposure effect, as first observed by Zajonc (1968), has also been called mere repeated exposure phenomenon. (Zajonc 2001). Tellingly, a meta-analysis of 118 studies by Montoya et al. (2017) found that the maximum effect of mere exposure was reached at 62.18 exposures. The power of repeated exposure has also been documented in the context of promotional merchandise (Beltramini 1992(Beltramini , 2000Kendrick 1998;McCarthy and Fram 2008). In contrast, the present manuscript aims to investigate situations in which exposure is limited to a single occasion as described in the example of borrowing a pen.
Although repetition clearly benefits persuasion, there is some evidence that even single exposure might suffice for the mere exposure effect to arise. Monahan, Murphy, and Zajonc (2000) found in their Study 2 that repeated subliminal exposure to Chinese ideographs and polygons led to a more positive evaluation of the presented stimuli than single exposure. However, even a single exposure to the stimuli was enough to improve the respondents' evaluation compared to a control group that was not exposed to the stimuli before. Single exposure thus can persuade. But the situation we study differs from the existing literature in more respects than the amount of exposures present.
When it comes to promotional merchandise, we are less interested in consumers' attitude towards the primary stimulus (e.g. the pen), but rather in their attitude towards a secondary stimulus, i.e. the brand promoted on the merchandise. The literature on incidental exposure commonly asks how exposure to a stimulus affects the evaluation of the primary stimulus. For instance, Janiszewski (1988) found that preconscious exposure to a print ad affects evaluation of the print ad. In addition, some evidence suggests that the effect of visual exposure may generalize to secondary stimuli. Monahan et al. (2000), for instance, showed that incidental exposure to a primary stimulus (e.g. Chinese ideographs) can affect the evaluation of other secondary, even unrelated, stimuli (e.g. polygons). Although incidental use of merchandise thus deviates significantly from settings in which mere exposure effects have commonly been observed (i.e. repeated presentation of a primary stimulus), it is possible that the effect generalizes towards the promoted brand.
We propose that this is in fact the case. Notably, we derive this proposition because incidental use of merchandise involves more than incidental visual exposure. Exposure to promotional merchandise also includes haptic interaction with the primary stimulus, i.e. the merchandise. Most prior literature on incidental exposure has focused on purely visual exposure (e.g. Huang and Watanabe 2012;Monahan et al. 2000). Vision is clearly our dominant sense (Krishna 2012). There are several studies that show that when both vision and touch are present, vision strongly or completely dominates touch (e.g. Krishna 2006;Miller 1972). And yet we also know that haptic interaction can improve responses to an object (Brasel and Gips 2014;Peck and Shu 2009;Wolf et al. 2008). Peck and Shu (2009), for instance, found that merely touching an object (i.e. mug, slinky, pencil, play foam) increases feelings of psychological ownership for and valuation of that object. Whereas this research investigated responses to the touched object itself, Brasel and Gips (2014) found a similar effect when consumers did not directly touch the product, but touch a depiction of the product on a touchscreen. Use of promotional merchandise also entails a situation of mediated touch, i.e. the logo of the brand is touched via the handling of the merchandise. The element of touch thus should add to the persuasive powers of promotional merchandise in the case of incidental use.
Although we assume that the incidental use of promotional merchandise can affect consumers even if it happens one time only, we do not assume that this would always be the case. Rather it is likely that the incidental nature of these encounters will limit their persuasive power to contexts in which people are particularly open to persuasion. One factor affecting such openness is the prior familiarity with a stimulus. In current research on promotional merchandise, customers received merchandise directly from a brand. In such settings, brand familiarity is of comparably little importance because customers necessarily get somewhat familiar with a brand prior to receiving a gift from that brand. In contrast, consumers may incidentally use merchandise of any brand including brands they never even heard of. Brand familiarity is thus of potential relevance in the context of incidental merchandise use.
To account for that, we explicitly consider the role of brand familiarity in this research. There is mixed evidence on whether mere exposure works better when a stimulus is familiar or when it is unfamiliar. According to the perceptual-fluency explanation for the mere exposure effect, mere exposure is more powerful for familiar stimuli. This is because they are easier to encode and process which is often attributed to liking of the stimulus (Bornstein and D'Agostino 1994). For the most part, however, incidental exposure has been studied and found as particularly effective in the context of stimuli that were previously unfamiliar to a person. The exposure induced familiarity has been credited with increasing the liking of previously unfamiliar stimuli (Fang, Singh, and Ahluwalia 2007; Monahan et al. 2000;Zajonc 2001). In the absence of prior knowledge and beliefs, consumers draw on whatever evidence they perceive when forming their first impressions. This is what happens when consumers evaluate an unfamiliar brand. In these cases, consumers' judgments often reflect inputs that were garnered without explicit attention and awareness (e.g. Olson and Fazio 2002). This means that input via incidental exposure can form attitudes towards the hitherto unfamiliar (Cacioppo et al. 1992;Crano and Prislin 2006). For familiar brands, however, consumers already hold an attitude, tend to know their preferences and have prior intentions on how to engage with it (Bass and Talarzyk 1972). Marketers aiming to affect responses to familiar brands, thus, need to change rather than form these responses. Attitude change means altering what is there and that is a considerably more effortful process for individuals than the process of attitude formation. Consumers are reluctant to alter existing views (Erdem et al. 1999;Muthukrishnan, Pham, and Mungale 2001) and once formed, attitudes tend to be relatively stable (Solomon, Russell-Bennett, and Previte 2012;. For attitudes to change nonetheless, consumers need to explicitly attend to new information and they need to expand cognitive effort, if change is to happen quickly (Cacioppo et al. 1992;Crano and Prislin 2006). Alternatively, consumers need to be exposed to the new information multiple times for it to gradually shift existing responses (Zajonc 1968).
While this might be the case when consumers repeatedly and deliberately use promotional merchandise, it is not when consumers use promotional merchandise incidentally. Incidental use of promotional merchandise does not cognitively engage consumers with the logo on the merchandise and mostly happens once rather than repeatedly. Incidental use of promotional merchandise is therefore unlikely to produce a change in the way consumers react to promoted familiar brands although it may well alter how they respond to promoted unfamiliar brands. We therefore propose the following hypothesis: H1: A single instance of incidental use of promotional merchandise from an unfamiliar (familiar) brand has a positive (has no) effect on consumer reactions towards the promoted brand.
Note that we have chosen the term reactions because it encapsulates different types of consumer responses such as attitudes, preferences, willingness to pay and purchase intentions. We will explore the effect of incidental use of promotional merchandise on all of these.
Consumers can be incidentally exposed to many marketing and advertising media. As one of these marketing and advertising media (Cooper 2009), promotional merchandise competes with the complete range of advertising techniques and channels. The question thus arises how the incidental use of promotional merchandise performs in comparison to purely visual exposure, which is characteristic of many other advertising tools such as posters, print or display ads. If we just consider the visual advertising element of promotional merchandise, i.e. the logo imprinted on it, we would probably expect other visual advertising tools to be more powerful. For instance, a poster is visually more noticeable and since vision is our dominant sense (Krishna 2012), a poster could be more effective than the logo on a promotional pen. Promotional merchandise, however, adds the element of haptic interaction to the element of visual exposure. The question remains open whether the positive effect of haptic exposure (Brasel and Gips 2014;Wolf et al. 2008) is enough for promotional merchandise to outperform a visually more prominent poster. We assume that it can and therefore suggest the following hypothesis: H2: Compared to incidental visual exposure, incidental use of promotional merchandise has a stronger effect on consumer reactions towards a promoted unfamiliar brand.
To test these hypotheses, we designed two studies, in which we experimentally induced incidental use of promotional merchandise in controlled lab settings. In line with our theorizing, the research designs kept use of promotional merchandise brief and incidental, i.e. participants used the promotional merchandise but were induced to focus their attention away from the merchandise, which was incidental to the focal task. In both studies, we test for the effect of incidental use of promotional merchandise on consumer reactions separately for a promoted unfamiliar and a promoted familiar brand (H1). In Study 2, we additionally test if and how the incidental use of promotional merchandise differs from incidental visual exposure to a brand via a poster (H2).

Study 1
In Study 1, we provide a first experimental test for our main proposition that a single incidental use of promotional merchandise can alter consumer responses for an unfamiliar brand promoted on the merchandise. We used pens as promotional merchandise because they are among the most frequent promotional items (Promotional Products Association International 2019) and are often used incidentally. As relevant product category, we chose telecommunication brands. We did so because many unknown brands operate in the European market, the cultural context of our studies. This means that consumers are used to new and unfamiliar brands in this category and it helps to avoid possible confounds and demand effects related to the unexpectedness of the stimulus material. In addition, brand names in the telecommunication category oftentimes hint at the category (e.g. T-Mobile). This facilitated the identification of a suitable unfamiliar brand name, i.e., bmobile, which does not exist in the country of the study.

Pretest
We conducted a pretest (n ¼ 141 within subjects) to assess familiarity of the two target brands, T-Mobile and bmobile. As intended, recognition of the familiar T-Mobile brand was high (pretest recognition: 98.6%) and recognition of the unavailable and thus unfamiliar bmobile brand was low (pretest recognition: 2.8%) and recognition rates significantly differed between brands (X 2 (1) ¼ 258.56, p < .001).

Participants, design and procedure
In return for partial course credit, 128 undergraduate business students (50% female, M age ¼ 21.5) were asked to partake in two presumably unrelated studies. The first study served to manipulate incidental use of the merchandise (see step 1 in Figure 1). Participants were provided with a pen and paper and asked to write a shopping list for Christmas. This decoy task mirrored real-world situations in which consumers focus on something else while incidentally using promotional merchandise. The randomly distributed pen participants wrote with featured either a familiar telecommunications brand (T-Mobile), an unfamiliar telecommunications brand (bmobile) or no brand. There were thus two experimental conditions and one shared control group (no brand) that served as a baseline for both of the experimental conditions (see also Figure 1). The pens were professionally designed by a promotional merchandise company (see Appendix A for a depiction of the stimulus material).
After participants finished writing, they handed back the pen and the shopping list and proceeded to a presumably unrelated online study on telecommunication providers (see step 2 in Figure 1). Participants answered questions on four telecommunication brands. These included questions about T-Mobile, which served as dependent variables for analyses of the familiar brand and questions about bmobile, which served as dependent variables for analyses of the unfamiliar brand. In addition, participants answered questions about two filler brands which served as decoy brands. Subscribing to a holistic definition of the attitude concept (Solomon et al. 2012), we aimed for a comprehensive measure of attitudes and thus adapted brand attitude scales similar to Lee, Li, and Edwards (2012) and . On screen, we assessed brand attitude for each brand using a 5-point bipolar scale consisting of eight pairs of items (low quality/high quality, not reliable/reliable, old fashioned/modern, dislikeable/likeable, cheap/expensive, negative/positive, does not appeal to me/appeals to me, do not like it/like it) (a bmobile ¼ .80; a T-Mobile ¼ .82). Next, participants ranked the four brands according to their preference and indicated how much they would be willing to pay for a pre-specified basic mobile phone contract (unlimited minutes and texts, 3 GB data, download speed at 10 Mbps) with each of the four brands. Before thanking participants, we assessed if respondents recalled and recognized the brand on the pen. First, we asked them if they remembered the brand imprinted on the pen. If they ticked yes, they were asked to enter the brand in an open text field. People that ticked yes and correctly replicated the brand name were coded as recalling the brand. Afterwards all respondents, irrespective of whether they recalled the brand or not, were shown logos of the four telecommunication providers and asked if they remembered whether any of these logos was featured on the pen (recognition).

Results
On average, participants spent 417 seconds on writing 33 words. There was no difference in the number of words and the duration of writing between conditions (all p's > .607). Using both the number of words and the duration of writing as covariates (all p's > .203) did not change our results.
Due to the fact that our dependent variables were brand-specific, we conducted separate tests for the familiar and unfamiliar brands. To test for the effect of incidental use of merchandise, we contrasted the experimental groups' evaluation of the respective brand with the evaluation of that brand by the shared control group. For instance, brand attitude towards bmobile reported by the group that used the promotional pen imprinted with the bmobile logo (i.e. bmobile condition) was compared to the control group's brand attitude towards bmobile. Likewise, brand attitude towards T-Mobile reported by the group that used the T-Mobile pen (i.e. T-Mobile condition) was compared to the control group's attitude towards T-Mobile. Table 1 includes descriptive statistics of all conditions. We used T-tests to analyze results with regard to the interval-scaled dependent variables, attitude and willingness to pay, and U-tests to analyze results with regard to the ordinal-scaled dependent variable, preference rank. We expected to find more favorable reactions towards the unfamiliar brand bmobile for people who incidentally wrote with a promotional pen featuring this brand but held no such expectations for the familiar brand T-Mobile.

Unfamiliar brand bmobile
Of those writing with the pen branded with the unfamiliar brand bmobile, 90% did not recall the brand on the pen. Even after showing participants the logo, 61% failed to recognize it. This confirms that the pen was in fact used incidentally. Any effects we observed in the unfamiliar brand condition are hence likely to have been obtained outside of conscious awareness. We still found differences between participants using an unbranded pen (control group) and participants using the promotional pen featuring the unfamiliar brand bmobile for all dependent variables.
Specifically, we identified a marginally significant effect of the promotional merchandise on brand attitude. The unfamiliar brand bmobile was evaluated more positively by those who had written with a pen promoting the brand (M ¼ 2.67, SD ¼ 0.50) than those who had written with an unbranded pen (M ¼ 2.46, SD ¼ 0.51; t(78) ¼ 1.96, p ¼ .057). Respondents who had written their shopping list with its promotional pen were also willing to pay a third more for the same contract with the unfamiliar brand bmobile (M pen ¼ 12.75e, SD ¼ 5.74) than those who had written with an unbranded pen (M control ¼ 9.60e, SD ¼ 3.53; t(62.255) ¼ 2.85, p ¼ .006). Finally, participants in the merchandise condition also ranked the unfamiliar brand bmobile significantly better (Mdn control ¼ 4.00; Mdn pen ¼ 3.00; U ¼ 601.00, z ¼ À2.12, p ¼ .034). While 64% of those who had used an unbranded pen ranked the unfamiliar brand bmobile last (higher value), only 42% of those who had used the promotional pen with the bmobile logo on it did so.

Familiar brand T-mobile
Slightly less than half of the respondents (49%) recalled seeing the familiar brand T-Mobile on the pen. After additionally showing them the logo, 60% stated that they had noticed it.
Despite better brand awareness, we neither found a significant effect of the promotional merchandise on brand attitude (M control ¼ 3.68, SD ¼ 0.41; M pen ¼ 3.56, SD ¼ 0.69; t(85) ¼ À0.97, p ¼ .344), nor on WTP (M control ¼ 14.00, SD ¼ 5.60; M pen ¼ 15.48, SD ¼ 6.45; t(77) ¼ 1.07, p ¼ .287) or brand preference (Mdn control ¼ 2.00; Mdn pen ¼ 2.00; U ¼ 815.50, z ¼ À1.13, p ¼ .258). Participants who had written with an unbranded pen assessed the familiar brand T-Mobile similar to participants who had written with a pen featuring this brand. Given that about half of our sample had recalled the familiar brand T-Mobile on the pen, we repeated the analyses by adding recall of the brand on the pen as a predictor. We ran ANOVAs for attitudes and willingness to pay and an ordinal regression for preference rank. Neither of the main effects, nor their interaction approached significance for any of the dependent variables (all p's > .124).

Discussion
As hypothesized, we only found an effect of incidental use of promotional merchandise for the unfamiliar brand bmobile but not for the familiar brand T-Mobile (H1). After a single use of a promotional merchandise, respondents held more favorable attitudes towards the promoted unfamiliar brand bmobile, ranked it higher and were willing to pay more for a product offered by the brand as compared to respondents who had used an unbranded pen. Notably, we found the effect even though most respondents failed to notice the unfamiliar brand on the promotional merchandise. Study 1 delivered a proof of concept but it cannot determine whether as hypothesized the observed effect is specific to the incidental use of merchandise or arose as a simple matter of visual exposure to the stimulus brand. Addressing this question matters theoretically and practically because it allows disentangling incidental visual exposure and incidental use.

Study 2
We designed Study 2 to corroborate the evidence for H1 and to test for H2, i.e we investigated how incidental use of promotional merchandise compares to incidental visual exposure. To provide a strict test for this question, we included another frequent advertising tool, i.e. a poster, which we put directly in participants' line of vision. A poster covers larger visual areas and thus entails even more visual exposure but simultaneously lacks the haptic element characteristic of merchandise.
In addition to allowing for a test of all our hypotheses, in Study 2 we aimed to replicate results in a different context. Based on pretests, we chose a different product category, digital cameras. Again, we assume to find an effect of incidental use of merchandise for an unfamiliar brand but not for a familiar brand. Furthermore, we assume that incidental exposure via merchandise will outperform purely visual incidental exposure via a poster.

Pretests
We pretested (n ¼ 160 within subjects) seven product categories (digital cameras, clothing, TVs, smartphones, tablets, insurances, laptops) to identify a product category where participants would not be surprised to come across a brand they are unfamiliar with. We chose digital cameras because people assumed that there would be multiple brands (M ¼ 4.59, SD ¼ 1.05; 1 ¼ few brands, 6 ¼ many brands) and would not expect to know all these brands because their interest in the product category (M ¼ 3.29, SD ¼ 1.57; 1 ¼ very low interest, 6 ¼ very high interest) was not particularly high. Please refer to Appendix B for full results for all pretested product categories.
We next conducted an additional pretest (n ¼ 22 within subjects) to identify suitable brand names. Our aim was to ensure that liking of the brand names was similar but that brand familiarity differed across the unfamiliar and familiar target brands. This was to avoid possible confounds due to differences in brand name liking, an issue we had avoided in Study 1 through the parallel construction of familiar and unfamiliar brand names. Based on the pretest, we chose the widely known, familiar brand 'Panasonic' (pretest recognition: 91.3%) and the fictitious and thus unfamiliar brand 'Aerius' (pretest recognition: 4.3%) to serve as focal brands. While differing in familiarity (X 2 (1) ¼ 10.98, p ¼ .001), both brand names were perceived as similarly likable on a 6-point scale anchored by 1 ¼ not at all likeable and 6 ¼ very likeable (M Panasonic ¼ 4.32, SD ¼ 0.95; M Aerius ¼ 3.73, SD ¼ 1.39; t(21) ¼ À1.44, p ¼ .163). Please refer to Appendix B for full results for all pretested brands.

Participants, design and procedure
Overall, 253 students (59% female, M age ¼ 23.8) participated in return for financial compensation in two presumably unrelated studies. In the first study, we manipulated incidental use of the merchandise via a promotional pen and incidental visual exposure via a poster while participants were engaged in a decoy task (see step 1 in Figure  2). Participants received a bottle of ketchup, entered an individual cubicle and answered 5 open-ended and 21 closed questions about their general consumption behavior of ketchup, attitudes and evaluation of the ketchup bottle in a paper-pen product test. Each cubicle was equipped with a pen and a small poster (A3, i.e. 29.7 Â 42.0 centimeters or 11.7 Â 16.5 inches). We mounted the poster directly in participants' line of vision. This setting ensured actual visual exposure (Van Meurs and Aristoff 2009) and enabled us to keep the duration of purely visual exposure similar to the duration of incidental use of the merchandise. Participants' focal task in all conditions was the evaluation of the ketchup bottle. This drew participants' visual attention away from the promotional merchandise and the poster. Exposure to all stimulus materials was incidental to the evaluation task.
We applied a 2 (brand on merchandise: present vs. absent) x 2 (brand on poster: present vs. absent) design per brand. Depending on condition, pen and/or poster featured either one of the brands or a neutral decoy. The decoy pen featured no brand, the decoy poster announced an upcoming event. All participants were only exposed to one of the brands, either the unfamiliar brand Aerius or the familiar brand Panasonic. Given that the control condition did not feature any brands, neither on the pen nor on the poster, this condition was identical for both brands and served as a shared control group (see also Figure 2). Overall, we thus had seven conditions: three conditions per brand (pen only, poster only, pen and poster) plus an additional shared control condition in which participants wrote with an unbranded pen and saw a decoy poster. The logo for the fictitious brand Aerius, the promotional pens and the posters were professionally designed (see Appendix A for a depiction of the stimulus material).
After manipulating brand exposure in the presumable first study, participants were asked to leave the cubicles to proceed to the second study that took place in an adjourning computer lab (see step 2 in Figure 2). In this study, we assessed the dependent variables in a study on digital camera brands. To keep the study purpose well hidden, we assessed responses to our two focal brands Aerius and Panasonic and three additional decoy brands. Brand attitude for each brand was measured on a 5point bipolar scale and a subset of five items from Study 1 (low quality/high quality, not reliable/reliable, old fashioned/modern, dislikeable/likeable, cheap/expensive) (a Aerius ¼ .84; a Panasonic ¼ .81). Participants also again ranked all brands according to their preference. In addition, we assessed purchase intention for each brand (7-point scale anchored by 1 ¼ not at all likely to buy and 7 ¼ very likely to buy). Before leaving, we assessed whether the participants remembered having seen any of the five evaluated brands on the poster or on the pen used in the presumable previous study.
We used a multiple choice question. Participants could indicate whether they remembered a pen and/or poster of either of the five brands or could not remember.

Results
As in Study 1, all analyses were conducted separately per brand and the shared control group's evaluations of each brand served as baseline for the respective experimental conditions. This was necessary because the dependent variables were brand specific and assessed separately for the familiar and unfamiliar brands. For each of our focal brands, we ran 2 (merchandise) x 2 (poster) ANOVAs to predict attitudes and purchase intention and corresponding ordinal regressions (Polytomous Universal Model) with dummies for promotional merchandise, poster, and their interaction to predict an effect on preference rank.

Unfamiliar brand Aerius
Of those exposed to the unfamiliar brand Aerius on either merchandise or poster, only 9% of respondents noticed the brand. Of those exposed to the logo on both poster and merchandise 14% of respondents noticed the brand on the merchandise and 18% noticed the brand on the poster (see Table 2 for all recognition rates). Any effects we find are hence due to incidental exposure and are likely to have occurred beyond awareness.
As expected, we found a main effect of the incidentally used promotional merchandise on brand attitude (F(1,141) ¼ 9.62, p ¼ .002, g 2 ¼ .064). Compared to respondents who had used an unbranded pen (M ¼ 2.79, SD ¼ 0.74), respondents who had used the pen branded with the unfamiliar brand Aerius (M ¼ 3.20, SD ¼ 0.86) evaluated Aerius more positively (see Table 3 for descriptive statistics). As expected the main effect of the purely visual exposure through the poster was weaker and did not turn significant (M control ¼ 3.04, SD ¼ 0.77; M poster ¼ 2.96, SD ¼ 0.88; F(1,141) ¼ 0.61, p ¼ .437, g 2 ¼ .004). However, there was a significant interaction effect between the poster and the promotional merchandise (F(1,141) ¼ 5.10, p ¼ .025, g 2 ¼ .035). The most positive brand attitude was achieved by a combination of incidental use of promotional merchandise and incidental exposure to the corresponding poster (M ¼ 3.29, SD ¼ 0.85).
Furthermore, there was a significant main effect of the merchandise (F(1,141) ¼ 6.04, p ¼ .015, g 2 ¼ .041) on purchase intention. Respondents who had used the promotional pen (M ¼ 2.62, SD ¼ 1.64) were more willing to buy a product of the unfamiliar brand Aerius than respondents who had used an unbranded pen (M ¼ 2.03, SD ¼ 1.22). Again there was no main effect of the poster (F(1,141) ¼ 1.36, p ¼ .245, g 2 ¼ .010) but a marginally significant interaction effect between the poster and the promotional merchandise (F(1,141) ¼ 3.51, p ¼ .063, g 2 ¼ .024) with the combination of incidental use of merchandise and incidental visual exposure to the poster yielding the highest purchase intention (M ¼ 2.97, SD ¼ 1.73). We also found a marginally significant main effect of the incidental use of merchandise on preference for the unfamiliar brand Aerius (b ¼ À1.26, Wald X 2 (1) ¼ 3.26, p ¼ .071). Overall, 80% of the respondents who had used an unbranded pen ranked the unfamiliar brand Aerius last, while this figure shrank to 66% among those using the brand's promotional pen. This was qualified by a significant interaction term between poster and promotional merchandise (b ¼ 1.38, Wald X 2 (1) ¼ 4.41, p ¼ .036). Again, the combination of poster and merchandise resulted in the most favorable ranking with 63% ranking the unfamiliar brand Aerius last. The main effect of the poster did not turn significant (b ¼ À0.23, Wald X 2 (1) ¼ 0.40, p ¼ .525).
Neither incidental use of the merchandise nor incidental visual exposure to the poster nor their combination affected attitude towards the familiar brand Panasonic (all F's(1,140) < 1.37, all p's > .243). There was also no main effect of the promotional pen on purchase intention (F(1,140) ¼ 0.43, p ¼ .512, g 2 ¼ .003). However, we observed a marginal main effect of the poster on purchase intention (F(1,140) ¼ 3.84, p ¼ .052, g 2 ¼ .027). Respondents who had been exposed to the poster of the familiar brand Panasonic (M ¼ 4.84, SD ¼ 1.40) were more likely to buy a Panasonic product than respondents who had not been exposed to the poster (M ¼ 4.32, SD ¼ 1.66). There was no additional interaction effect between the promotional pen and the poster on purchase intention (F(1,140) ¼ 0.02, p ¼ .887, g 2 < .001). Finally, there were also no main effects of the promotional pen or poster or interaction effect on preference (all Wald X 2 (1) < 0.44, all p's > .258).

Discussion
Study 2 replicated and extended the findings of Study 1. Based in a new context and featuring a different task of merchandise use, incidental use of promotional merchandise affected brand reactions for a brand that participants were unfamiliar with (Aerius) but did not do so for a familiar brand (Panasonic) (H1). Again, we observed this effect even though the majority of our respondents failed to realize that they had been using promotional merchandise.
Importantly, Study 2 allows us to disentangle incidental use of promotional merchandise from more general effects of incidental visual exposure. In the case of the unfamiliar brand Aerius, promotional merchandise appears to outperform simple visual exposure (H2) though their combination appears to be most effective in inducing favorable consumer reactions. In the case of the familiar brand Panasonic, no such advantage emerged.
While the familiar brand Panasonic did not benefit from the one time incidental use of its promotional merchandise, it showed some benefit from the presence of a poster, which increased respondents' intention to purchase a product of the brand. That only this particular variable was affected could indicate that visual exposure to the poster served as a reminder. It might have reminded participants that they could purchase a brand they were already familiar with without changing the way they felt about this brand (e.g. Ambler 2000).

General discussion
People often use promotional merchandise incidentally. When, for instance, needing to write something down, people often use the first pen that comes to handoften this may be someone else's promotional pen. The focus in such situations is on the use value, not the aesthetics and message of the merchandise, i.e. the merchandise is used incidentally, and consumers can easily miss the promotional message the merchandise carries. This research is the first to show that promotional merchandise can affect consumer reactions even under these adverse circumstances.
Across two different product contexts, i.e. telecommunication providers and digital cameras, and across two different decoy tasks, i.e. writing a shopping list and evaluating a product, we find that even purely incidental use of a promotional pen can affect the way people react to the brand promoted on the merchandise. Notably, this only holds for unfamiliar brands, but does not extend to familiar brands. We propose that this is because attitudes, preferences and behavioral intentions towards a familiar brand are already established and thus harder to change (e.g. Petty and Brinol 2008). We further propose and find that the effect is specific to the incidental use of merchandise and does not simply reflect the fact that incidental use involves incidental visual exposure. For the unfamiliar brand, results of Study 2 suggest that incidental use of promotional merchandise holds an advantage over purely visual incidental exposure to a poster. Reactions were, however, most favorable when participants both incidentally used promotional merchandise and were visually exposed to a poster of the same brand. For familiar brands, neither one-time incidental encounter was very effective in altering consumer reactions to the brand.

Theoretical implications and future directions
To our knowledge, this paper is the first one to show that promotional merchandise persuades even based on brief, incidental use. This finding goes beyond the main theoretical foci in the area of promotional merchandise, i.e. gift giving and reciprocity (e.g. Beltramini 2000;Bodur and Grohmann 2005), and adds to its theoretical basis. Given that we find that participants were largely unaware of having been exposed to a logo, the results also add to the literature on non-conscious influences on consumer evaluation and behavioral intentions (e.g. Fitzsimons et al. 2002). Specifically, our findings provide several novel insights.
First, most previous research on effects of exposure has documented an effect of incidental exposure to a stimulus on the evaluation of that very same stimulus (e.g. Janiszewski 1988). Our studies show that the effect can also emerge for a brand promoted on that stimulus.
Second, most prior studies on incidental exposure stress the necessity of repeated exposure (e.g. Matthes et al. 2007;Zajonc 2001). We find that even a single incidental encounter suffices to alter consumer reactions.
Third, and that might explain why we find an effect even for single incidental exposure to a secondary stimulus, we find this in a context in which people also haptically interact with the stimulus. Most prior research focused on incidental exposure as a purely visual phenomenon (Bargh 2002). Although vision tends to overrule haptics (Krishna 2006(Krishna , 2012, touch appears to add to the persuasive powers of incidental encounters. This insight also extends prior research on the power of touch (Brasel and Gips 2014;Peck and Shu 2009;Wolf et al. 2008) by showing that haptic interaction cannot only improve the evaluation of the touched object but also attitudes towards an unrelated message on the touched object.
Our results are also in line with the so called elaboration likelihood model of persuasion (Petty and Cacioppo 1986). We find that incidental use of promotional merchandise benefits in particular unfamiliar brands that consumers hold no prior attitudes towards. Results thus also provide further evidence on a somewhat open question in the realm of incidental exposure. Our research aligns well with the dominant view that incidental exposure benefits in particular unfamiliar (Fang et al. 2007;Monahan et al. 2000;Zajonc 2001) rather than familiar stimuli (Bornstein and D'Agostino 1994).
Although we generalized results across contexts, open questions and limitations remain and our results reveal several directions for future research. First, it would be interesting to see how long-lived the observed effects are. While the jury is still out, prior literature suggests that effects would persist. We observed the effect in situations of unfamiliarity where promotional merchandise provided the first impression of a brand. First impressions are known to be particularly powerful among the stream of impressions people draw on when making judgements (Bhargave and Montgomery 2013). In addition, the effect mostly emerged outside of conscious awareness. A majority of participants did not recall seeing the logo. This means that consumers cannot actively question the basis of their reactions, thus making their impressions potentially stick better (Friestad and Wright 1994;Lapate et al. 2014).
Second, though our results are in line with the proposed process of implicit evaluation through haptic and visual exposure, this process appears to happen subconsciously. It thus eludes conventional direct measures. It would be interesting for future research to further study the process via which incidental use persuades. One possibility could be the use of eye tracking studies.
Third, in our studies, attitude towards the familiar brands T-Mobile and Panasonic was moderate (between 3.51 and 3.75 on a 5-point scale). It would be interesting to see whether results change if prior attitudes towards familiar brands are particularly negative or positive such as for popular brands such as Apple or Nike. Since more polarizing attitudes tend to be even stronger (Albert and Merunka 2013) and therefore harder to change (Howe and Krosnick 2017), we doubt that incidental use of promotional merchandise would be more effective in such cases.
Other interesting possibilities and open questions relate to the nature of the promotional merchandise and the way it is used incidentally. We show that incidental haptic exposure to a promotional pen is more effective than incidental visual exposure to a poster. Further research could investigate the role of haptics by manipulating the tactile components of the promotional pen. 1 Moreover, in both studies, we used high quality promotional merchandise. This may be a limitation because it is unclear what would happen if the promotional merchandise were of low quality. Negative quality perceptions of the pen might spillover to attitudes towards the brand, especially if the brand is unfamiliar (Ahluwalia, Unnava, and Burnkrant 2001).
Factors such as the fit between promotional merchandise and brand or the location or size of the logo (Sundar and Noseworthy 2014) might also play a role in determining its effectiveness. Likewise, the nature of the task people fulfill with the promotional item might affect its persuasive power. In our studies, people engaged in agreeable to neutral tasks (writing a Christmas shopping list and evaluating ketchup), it is an open question whether effects would generalize to more disagreeable tasks such as writing an exam or signing bills. Future research is needed to address these and other relevant questions.

Managerial implications
Given the prevalence of incidental exposure to promotional merchandise in the real world, our results are of high practical relevance. Prior research (Beltramini 1992(Beltramini , 2000Kendrick 1998;McCarthy and Fram 2008) has already demonstrated a positive effect of promotional merchandise on consumer reactions towards the promoted brand. However, this pertains to settings in which consumers pay attention to a brand they already encountered. In a nutshell, existing research primarily suggests that promotional merchandise may be a great tool if the purpose is to stay in people's mind or to entice them via a gift. Our research treads novel ground beyond such settings and shows that the effectiveness of promotional merchandise might have still been underestimated. Specifically, we study an incidental use setting in which promotional merchandise cannot benefit from its known strong points and in which consumers may encounter a brand's merchandise without having ever encountered the brand itself. In the setting we studied, promotional merchandise was not received as a gift and exposure was limited to a single interaction. Still, we find that such incidental onetime use of the merchandise was able to dispose consumers more favorably. Notably, this happened only for brands consumers were unfamiliar with. Unfamiliar brands or brands entering a new market are hence particularly well advised to include promotional merchandise in their communication portfolio. In addition, these insights hold implications for the choice of merchandise. Unfamiliar brands are well advised to use merchandise that is durable and likely to switch hands often, such as pens. The fact that purely incidental use can be effective also holds implications for the relevant target group and merchandise distribution. People that lend certain items often such as clerks or waiters appear particularly promising targets for the distribution of merchandise. Similarly, hubs of engagement such as community centers or common rooms might serve the purpose of a multiplication of incidental exposures well.
As for many other advertising tools 'the upward potential for familiar brands is limited' (Verhellen, Dens, and De Pelsmacker 2016, 469); at least when it comes to the effects promotional merchandise unfolds upon incidental use. Prior research, however, suggests that repetition may achieve what one time exposure does not (Zajonc 1968). It is likely that familiar brands benefit from repeated incidental use of promotional merchandise over time (e.g. Kendrick 1998;McCarthy and Fram 2008). Moreover, the effects obtainable via receiving the merchandise as a gift (Cialdini and Goldstein 2004) are likely to extend to all brands.
In addition, our findings provide recommendations for a brand's media mix. In the case of unfamiliar brands, a combination of traditional advertising, such as posters, and promotional merchandise yielded the most promising results. Visual exposure to the poster alone had no effect. This is practically relevant because it suggests that financially stretched unfamiliar brands should prioritize investing into merchandise and its distribution. It also suggests that unfamiliar brands that aim to invest into multiple tools should concentrate rather than spread the timing of the respective efforts in their media schedule.
To conclude, our results suggest that brands may benefit from promotional merchandise to a larger extent than previously thought. Even if consumers fail to notice the logo, purely incidental, one-time use of a merchandise may suffice to favorably dispose them towards the promoted unfamiliar brand. For unfamiliar brands, promotional merchandise appears to be a particularly effective silent persuader that affects even those that incidentally use it. Note 1. We thank an anonymous reviewer for these suggestions.

Disclosure statement
The design and analyses of all studies were entirely that of the authors. Funding for Study 2 was obtained from the Association of Austrian promotional merchandise companies (V € OW).