The Effectiveness of In-Store Free Samples on Sample Takers
Carrie Heilman, University of Virginia Kyryl Lakishyk, Catholic University of Portugal Sonja Radas, The Institute of Economics, Zagreb
July 2006
Carrie M. Heilman, Assistant Professor of Marketing, McIntire School of Commerce, University of Virginia, P.O. Box 400173, Charlottesville, VA 22904-4173, e-mail: heilman@virginia.edu, phone: 434-243-8738, fax: 434-924-7074. Kyryl Lakishyk, Assistant Professor of Marketing, Faculdade de Ciências Economicas e Empresariais, Universidade Católica Portuguesa, Palma de Cima, 1649-023 Lisboa, Portugal, e-mail: kylaki@fcee.ucp.pt phone: 351-217-214-250, fax: 351-217-270-252. Sonja Radas, The Institute of Economics, Zagreb, Trg J.F. Kennedy 7, 10000 Zagreb, Croatia, e-mail: sradas@eizg.hr, phone: 385-12335700. The authors would like to thank Amar Cheema, David Mick and Jim Burroughs for their helpful comments on this paper.
The Effectiveness of In-Store Free Samples on Sample Takers
Abstract We present an in-store experimental study that examines the impact of in-store free samples on “samplers” (those who accept the free sample). We also provide insights about non-samplers (those who see the sample but do not accept it) by comparing their behavior to that of samplers. We develop a set of hypotheses about the purchase probability of samplers post-sample, many of which apply a benefits congruency framework suggesting that when the benefits of two variables related to in-store sampling are congruent, the likelihood of purchase will be greater. We build a model of purchase probability using a binary logit model that tests our hypotheses and provides some useful insights for manufacturers who invest in in-store free sample promotions.
Key words: product sampling, in-store promotions, purchasing behavior, binary logit
2
1. Introduction Consumers today are bombarded with many different types of promotions. However, few are as effective as free samples for generating trial and actual purchase (e.g., Belch and Belch 1990; Rossiter and Percy 1987; Schulz, Robinson and Petrison 1993, Chapter 10). Studies in the popular press on in-store sampling, or sampling that occurs at the point-of-purchase (POP), have shown that 92% of consumers would rather be presented with a free sample over a cents-off coupon while in the store (Fitzgerald 1996), nearly 70% of shoppers will try an in-store sample if approached, 37% will buy the product once sampled (Lindstedt 1999), and in-store samples can increase sales of the sampled product by as much as 300% on the day of the promotion (Moses, 2005). These benefits of sampling explain why expenditure on sampling programs increased to approximately $2B in 2004, a 50% increase over that spent in 2003 (Zwiebach, 2005). However, in the academic marketing literature, sampling remains one of the most underresearched areas of promotions (Heiman et. al., 2001). The short- and long-term effects of consumer promotions such as coupons and price discounts have been studied extensively (Blattberg and Neslin 1990), however the effects of point-of-purchase1 (POP) samples are not easily inferred from research on other types of sampling programs (e.g., direct mail) due to some distinct differences. For starters, in-store sampling occurs in a public setting and typically involves a face-to-face exchange between the consumer and the person distributing the sample, and possibly in the presence of other consumers. This personal and public exchange may cause in-store samplers to respond differently than they would to promotions experienced in the privacy of their own homes, such as direct-mail samples.
1
We use the terms “point-of-purchase” and “in-store” interchangeable when referring to in-store, POP free samples. 3
Another unique characteristic of in-store samples is the marketers’ lack of control over who receives the promotion. In contrast to direct-mail samples or coupons, where the distribution of the promotion is determined and controlled by the brand manager (typically through the purchase of a mailing list), the distribution of an in-store free sample is influenced by the consumers in the store on that day and their willingness to accept the sample. Assuming consumers are motivated to sample for different reasons, such as their desire to learn more about the product, boredom, or simply hunger, then an understanding of such motives could allow marketers to increase the pool of actual samplers and perhaps the likelihood of purchase once the product is sampled. However, no known research has studied consumers’ motives for sampling or the factors affecting purchase once a product has been sampled. A final distinguishing characteristic of in-store sampling versus other types of promotions is its relevance to retailers as well as manufacturers. Most promotions (e.g., clip coupons, directmail free samples, etc.) are relevant only to the manufacturer. However, the conventional wisdom among store managers and industry experts is that in-store samples increase store traffic and store loyalty by drawing consumers into the store and enhancing their overall shopping experience. An increase in store traffic is obviously beneficial to retailers. And if in-store free samples truly do improve the shopping atmosphere, past research has shown that consumers in a good mood increase their purchasing expenditure at the store level (Donovan et al. 1994, Golden and Zimmer 1986, Heilman, Nakamoto and Rao 2002; Sherman and Smith 1987), leading to another benefit to retailers. However, no study has examined whether free samples do in fact affect consumer behavior at the retail level. Given the importance of POP sampling for brand managers and retailers alike, and due to their uniqueness relative to other types of promotions, it is clear that more research on this topic is needed.
4
In-store free samples can be interpreted very broadly to include everything from a free sample of cereal at the grocery store to an interactive experience and demonstration of a power tool at a home improvement store. Our study focuses on consumable in-store free samples at a grocery store. Our purpose is to investigate the impact of free samples at the individual consumer level, rather than the impact of free samples at the brand level (e.g., lift in sales of the promoted product). We focus on the impact of in-store free samples on those who ultimately accept a free sample, however we do investigate the motives of non-samplers for avoiding free samples and their overall store expenditure relative to samplers. We conduct an in-store experiment where we survey samplers and non-samplers about their experience with the sample promotion on that shopping trip, and their attitudes about, and perceptions of, in-store free samples in general. Our study allows us to investigate whether free samples influence store choice or perceptions of store quality, the reasons why consumers either take or avoid in-store free samples, and to compare the purchase probability between samplers and non-samplers. Then, focusing on samplers only, we develop a set of hypotheses about the likelihood of purchase post-sample. Our hypotheses focus on the impact of the following types of variables on the purchase probability post-sample: 1) product characteristics, specifically whether the product is hedonic or utilitarian in nature, 2) sampling experience characteristics, such as whether information or enjoyment was derived from the sampling experience, and 3) consumer characteristics, such as consumers’ motives for taking a free sample and demographic information. Furthermore, we hypothesize about the interactions between variables in these three categories. Specifically, we suggest that when there is congruency between the characteristics of certain variables above, such as the product characteristics and the motives to sample, the likelihood of purchase will increase (Chandon, Wansink and Laurent, 2000).
5
We build a comprehensive model of purchase incidence to test our hypotheses, which we then estimate using a binary logit model. Our model allows us to investigate the relative impact of different variables on purchase incidence post-sample. We discuss the implication of our results for marketing managers who invest in in-store samples and for retailers who support such programs. Finally, we conclude with some directions for future research. The next section first provides an overview of the previous work on product trial and sampling in the marketing literature. 2. Literature Review on Product Trial and Sampling Previous research on product trial and sampling can be divided into four groups; 1) those that study product trial, or the first usage experience with a product, 2) those that have developed models of the effects of free sampling, 3) those that have empirically studied the impact of direct mail sampling, and 4) those that have empirically study in-store free samples as we do. Our study builds upon these three streams of research and hence we provide an overview of each area below. Studies of Product Trial Most studies of product trial, defined as a consumer’s first usage experience with a brand or product (Kempf and Smith 1998), have been conducted in a laboratory setting and have examined the impact of trial on belief strength and attitude (Marks and Kamins, 1988), affect (Oliver 1992; Westbrook 1987), or perceptions of the brand (Bettinger, Dawson and Wales 1979; Hamm, Perry and Wynn 1969). Some papers have focused on the emotional antecedents of trial on product evaluation (Havlena and Holbrook 1986; Mano and Oliver 1993; Oliver 1992; Westbrook 1987; Westbrook and Oliver 1991), the cognitive antecedents of trial on product evaluation (Marks and Kamins 1988; Smith 1993; Smith and Swinyard 1983, 1988; Wright and
6
Lynch 1995), or both (Kempf 1999; Kempf and Smith 1998). Others have simply tried to reveal the various dimensions of trial experiences and the scales appropriate for measuring them (Batra and Ahtola 1990; Holbrook and Hirschman 1982; Mano and Oliver 1993). Most recently, Nowlis and Shiv (2005) examined the impact of distractions while sampling food on the likelihood of choice in a series of laboratory experiments and found that distractions increased the likelihood of choice. While these studies use laboratory experiments to provide insights into consumers’ cognitive and emotional responses to product trial, their results cannot be used to fully understand the full impact of in-store sampling for the following reasons. First, a consumer accepting an in-store sample may not be experiencing the product for the first time. Also, instore sampling has a self-selection component driven by consumers’ motives to sample, something that studies on product trial do not address. Finally, the studies on product trial conducted in a laboratory setting do not capture the environmental aspects and effects of an instore sampling experience. For these reasons, the results from the studies on product trial cannot be used to make generalizations about the impact of POP free samples on consumers’ shopping behavior. Models of Free Sample Effects Jain, Mahajan and Mullen (1995) use simulation to determine the optimal level of product sampling for a new product. To do so, they modify the Bass model by assuming that the coefficient of innovation is a function of the sampling level. Despite this innovative approach, they note empirical work is required in order to determine how the coefficient of innovation is related to the level of sampling.
7
More recently, Heiman et al. (2001) develop a model that decomposes the sampling effort into the immediate sales and longer-run (goodwill-building) effects. Their model allows them to identify an optimal sampling effort of a firm over time. Although they did not conduct any empirical tests of the model, they note that the conceptual model they developed could provide the basis for an empirical investigation and that one would have to develop "practical measures of goodwill" (p. 544). Direct Mail Sampling The research on free samples sent to and consumed in the home, typically via direct mail, is limited but growing. Some examples include Bawa and Shoemaker (2004) who examine three potential effects of free direct mail samples on sales, (1) an acceleration effect, whereby consumers begin repeat purchasing of the sampled brand earlier than they otherwise would; (2) a cannibalization effect, which reduces the number of paid trial purchases of the brand; and (3) an expansion effect, which induces purchasing by consumers who would not consider buying the brand without a free sample. The authors find that free samples can produce measurable longterm effects on sales that can last as long as 12 months after the promotion, and that the effectiveness of a free sample promotion can vary widely, even between brands in the same product category. McGuinness, Brennan and Gendall (1995) studied the effects of mailed product samples and coupons on product trial among non-users and found that samples and coupons encourage more trial when presented together rather than separately. While the studies above provide external validity for the laboratory studies mentioned earlier, because of the differences between direct mail sampling and in-store sampling, such as elimination of self-selection to the sample and lack of in-store environmental effects, their
8
findings can not be used to predict the impact of in-store samples on store traffic, store perceptions, or purchase incidence of the sampled product. In-Store Sampling The research examining in-store free samples is sparse and in many cases not easily generalized. Lammers (1991) provided samples to consumers in a chocolate store and found the sample increased the immediate sales of chocolate. However, the positive effect was restricted to small purchases of up to $5.00 and to purchases of chocolate varieties other than that sampled. Furthermore, no theoretical explanation was provided to support the findings. Due to the unique characteristics of the sampled product (e.g., impulse/gift purchase) and the retail format (e.g., a specialty shop where consumers patron for a specific purchase with limited product and brand alternatives), it is not clear the findings can be generalized to all in-store samples, and specifically to those in a grocery store setting. In addition, because all consumers who entered the store were given a free sample the study does not probe the reasons why consumers take or avoid free samples and the implications this has on purchasing behavior. Steinberg and Yalch (1978) also examined the effects of in-store free samples on shopping behavior, but they focused solely on the differences in shopping behavior of obese and non-obese consumers. They found that consuming a food sample caused obese consumers to increase their purchasing in the store that day, but again, the parameters of this study are highly specific. Finally, there are some studies that while they do not research the impact of in-store free samples explicitly, in-store samples are included in their study of in-store promotions. For example, Rothschild and Gaidis (1981) used behavioral learning theory in their study of sampling promotions to cast sampling as a positive influence on future purchasing. Gadenk and
9
Neslin (1999) study the impact of retail promotions on purchase event feedback. They develop a logit model and estimate the effect of in-store promotions on brand loyalty. They find that the estimated coefficients in their model were positive for in-store samples of mineral water and negative for price cuts. In summary, given the considerable amount of money used on free samples and the importance marketers seem to place on them it is surprising that more research on this type of promotion has not been published. Our study builds upon the basic findings of the research above and then introduces some unique characteristics and findings to the sampling literature due to our focus on in-store free sampling in its natural environment. For example, we investigate the motives for why consumers accept or avoid in-store free samples, an aspect of sample that has not been investigated in the past. We also develop a set of hypotheses about the likelihood of purchase post-sample that includes the impact of the sampling experience, the motives to sample, and the product characteristics themselves. By investigating these issues we provide a very rich understanding of the effects of in-store free samples for retailers and manufacturers alike. The following section presents our hypotheses about the impact of in-store samples on the purchase probability of sample takers. 3. Hypotheses Development for the Purchase Probability of Sample Takers We consider the impact of three categories of variables on post-sample purchase probability: 1) product characteristics, 2) the nature of the sampling experience, and 3) personal characteristics of the consumer, including demographics. Some of our hypotheses build upon the work of Chandon, Wansink and Laurent (2000) and therefore we review that paper below before discussing our hypotheses.
10
Chandon, Wansink and Laurent (2000) applied a benefit congruency framework to investigate the impact of the congruency between the hedonic/utilitarian characteristics of a promotion and the product being promoted on the effectiveness of the promotion. They began by identifying characteristics of utilitarian (e.g., savings, quality, convenient, etc.) and hedonic (e.g., exploration, entertainment, value expression, etc.) promotions. From this they found that non-monetary promotions (e.g., free gifts, sweepstakes, etc.) were perceived as hedonic in nature, whereas monetary promotions (e.g., coupons, rebates, price reductions, etc.) were perceived as utilitarian. They then tested for the effect of the congruency between the hedonic/utilitarian characteristics of the promotion and the product on the likelihood of purchase. They found that utilitarian promotions were more effective for utilitarian products while hedonic promotions were more effective for hedonic products, hence supporting the benefit congruency theory. Based on the findings of Chandon, Wansink and Laurent (2000), free samples are clearly hedonic in nature in that they are a non-monetary promotion that can provide exploration and entertainment benefits, among other things. We apply their work to hypothesize about the congruency effect between the benefits of the promotion (i.e., hedonic), the characteristics of the product sampled, the sampling experience, and consumers’ motives to sample on the likelihood of purchase. These and other hypotheses are developed below. Product Characteristics – Hedonic/Utilitarian Applying the work of Chandon, Wansink, and Laurent (2000) to free samples, we propose that a free sample, which is hedonic in nature, will be more effective for promoting purchase of a hedonic product than a utilitarian one. H1: Hedonic sampled products are more likely to be purchased than utilitarian sampled products.
11
Although Chandon, Wansink and Laurent (2000) tested the effectiveness of different types of promotions, they did so in a laboratory setting and free samples was not one of the promotions explicitly tested. Many free samples are accompanied by the presentation of a coupon. Chandon, Wansink and Laurent (2000) find “cents-off” coupons to be characteristically utilitarian. Applying a benefits congruency framework we propose that the impact of a cents-off coupon will be greater for utilitarian sampled products than for hedonically sampled products. H2: A cents-off coupon will be more effective for increasing purchase probability for utilitarian sampled products than for hedonic sampled products. Product Characteristics – Perceived Risk In addition to considering the hedonic/utilitarian characteristics of a sampled product (H1 and H2), we also consider the risks related to the product. We examine three types of perceived product risks – the risks associated with buying a new product, the risks associated with buying a non-national brand name product, and the financial risks related to purchasing an expensive product. Given free samples have been found to be one of the most significant risk relievers (Roselius, 1971), we propose a free sample promotion will lower these three types of perceived risk. There is substantial literature reporting on the risk associated with new product adoption (see review paper by Ross, 1975; Dowling and Staelin, 1994; Littler and Melanthiou, 2005). This has prompted research on risk reducing strategies (Michell and Boustani, 1994; Roselius, 1971), including information search as a means to lessen risk (e.g., Dowling and Staelin, 1994). Free sampling provides an opportunity to try and experience a product and therefore should reduce risks related to new products.
12
Another type of perceived product risk is that related to lesser known brands, such as private labels or regional brands. “Major brands” or national brands are typically viewed as less risky for durable and non-durable products relative to their private label or “minor brand” counterparts (Roselius, 1971; Derbaix, 1983; Batra and Sinha, 2000). Since free samples allow consumers to experience an unfamiliar product and gather information that can reduce perceived risks associate with that product, such a sampling experience should reduce the risks related to regional or private label brands. Finally, the relationship between price and perceived risk have shown that higher priced goods are perceived as having higher financial risks (Jacoby and Kaplan 1972), and the greater the financial risks associated with a product, the lower the likelihood of purchase. A free sample provides information about a product, and specifically whether the product is worth the financial risk. As such, a free sample should lower the perceived risks associated with higher priced goods. In conclusion, if a free sample does in fact reduce the perceived risks associated with a product (Roselius, 1971), then there should be little to no difference in the purchase probability between a new and non-new sampled product, a non-national and nationally branded product, or a higher and lower priced product (within reason). As such, we propose the following: H3: A sampled product that is new to the market should have an equal probability of being purchased as a sampled product that already exists on the market. H4: A sampled non-nationally branded product should have an equal probability of being purchased as a nationally branded sampled product. H5: The negative impact of price on purchase probability should be reduced as a result of sampling a product. Sampling Experience Characteristics – Hedonic/Utilitarian
13
We predict consumers’ impressions of the sample experience will impact their likelihood of purchase. Applying a benefit congruency framework we predict the likelihood of purchasing a utilitarian product will be higher when the sampling experience is perceived as utilitarian in nature (e.g., the experience was informative, it helped make an informed decision, the person distributing the sample provided useful information, etc.). Likewise, we predict that the likelihood of purchasing a hedonic sampled product will be higher when the sampling experience is perceived as being hedonic in nature (e.g., a pleasing experience, the experience puts the consumer in a good mood, the consumer liked the person distributing the sample, etc.). H6: The likelihood of purchasing a hedonic sampled product will be greater for consumers who perceive the sampling experience to be hedonic in nature. H7: The likelihood of purchasing a utilitarian sampled product will be greater for consumers who perceive the sampling experience to be utilitarian in nature. Sampling Experience Characteristics – Persuasion Persuasion (by the sample provider) or the perception of a “hard sell” is a unique aspect of in-store sampling that does not exist with a direct-mail sample or most other consumer promotions. Research has confirmed that consumers have a very negative perception of “hard sell” tactics (Darian, Tucci and Wiman, 2001; Tam and Wong, 2001) and there is ample anecdotal evidence in the business press to support that a “hard sell” is a put-off for customers and that sales people should refrain from it (for example Konopacki, 1997; 1996). As such, we propose: H8: The likelihood of purchasing a sampled product will be lower for consumers who perceive the sampling experience to be persuasive in nature. Consumer Characteristics – Sampling Motives A unique characteristic of in-store sampling is a consumer’s ability to self-select into the promotional experience. Therefore, we are interested in investigating the motives, specifically
14
utilitarian and hedonic, for why people sample, and the implications of such motives on a postsample purchase. The utilitarian and hedonic dimensions of shopping behavior have been documented as having an underlying presence across different consumption experiences (e.g., Batra and Ahtola 1990; Crowley, Spangenberg, and Hughes 1992; Engel, Blackwell, and Miniard 1993). Utilitarian shopping behavior has been described as task-related and rational (Batra and Ahtola 1990; Engel, Blackwell, and Miniard 1993). Consumers driven to shop based on utilitarian motives use shopping as an opportunity to gather information in order to make informed decisions. Consumers who shop based on hedonic goals treat shopping as a form of entertainment (Bellenger, Steinberg and Stanton 1976) and see the experience as one that is fun or emotionally enticing (Holbrook and Hirshman 1982). Hedonic shopping behavior may be motivated by the need for increased arousal, fantasy fulfillment or escapism (Bloch and Richins 1983a; Hirshman 1983), to name a few. Investigating the congruency between the motives for sampling and the benefits of the product we propose that a hedonic sampled product will be more likely to be purchased by a consumer who samples for hedonic reasons (e.g., samples to make the shopping trip more enjoyable) and a utilitarian sampled product will be more likely to be purchased by a consumer who samples for utilitarian reasons (e.g., samples to gather information about the product). H9: The likelihood of purchasing a hedonic sampled product will be greater for consumers with hedonic sampling motives. H10: The likelihood of purchasing a utilitarian sampled product will be greater for consumers with utilitarian sampling motives. Consumer Characteristics – Guilt
15
Next we discuss the impact of guilt on post-purchase sampling probability. In much of our early pre-tests and exploratory research that motivated this study we found that guilt and obligation to purchase were salient emotions associated with free sampling. If consumers view an in-store sample as a “gift”, then it is not surprising this would lead to feelings of, 1) reciprocity to repay the favor of the “gift” (Cialdini 1993, Chapter 2) by making a purchase, and/or, 2) guilt if a purchase is not made (English and Macker 1976), both of which would encourage the likelihood or purchase in the short-run (Adaval 2001; Mowen and Minor 1998; Scott and Yalch 1980). Therefore, we propose: H11: The likelihood of purchasing a sampled product will be greater for consumers who express feelings of guilt about sampling and not making a purchase. The following presents our in-store experiment and modeling approach to test our hypotheses and our findings with their implications. 4. In-Store Experiment We conducted an in-store experiment over a two-month period at different stores within a large Midwestern grocery store chain. With the cooperation of the director of the in-store sampling program and the various store managers, we leveraged the existing free-sample program implemented by the grocery store chain to collect our data. Information about the products sampled in our experiment is shown in Table 1. [Insert Table 1] We intercepted consumers as they exited the store and screened for those that had seen the free sample of interest. Among those who had seen the free sample we were interested in the perceptions and behavior of both samplers (those who took the sample) as well as non-samplers (those who did not take the sample). To compare perceptions of the two groups we asked consumers to fill out a survey about their motives for taking or not taking the free sample, and
16
for those who did sample, their perceptions of the sampling experience. Upon completion of the survey we offered all participants $5.00 and an opportunity to participate in a lottery drawing to win one of three $100 cash prizes in return for their grocery receipt. All survey participants accepted this offer. Of the 380 consumers who had seen the free sample, ninety-three percent (354) agreed to participate in our study. Of those you participated, 259 (73%) had sampled the product of interest and 95 (27%) had seen the free sample but chose not to take it. The distribution of all participants over the sampled products, as well as their demographic make-up, is shown in Table 1. Note there was no statistical difference between the demographics of samplers and nonsamplers. Initial Findings Before presenting the results from our model that formally tested our proposed hypotheses about samplers, we first discuss some initial findings about the differences between samplers and non-samplers. As noted earlier, free samples are as important to retailers as they are to manufacturers. Many store managers we interviewed mentioned their belief that free samples increased store traffic, influenced consumers’ store choice, and in general, made the shopping experience more enjoyable for their consumers, although they had no empirical evidence to support these claims. To investigate these issues we asked all 354 participants (samplers and non-samplers) a series of questions to assess the degree to which free samples influence their store choice and overall store perceptions. Table 2 shows these results. [Insert Table 2]
17
We see from the results that consumers look forward to receiving free samples, and to a lesser extent they feel that in-store free samples make the shopping experience more festive. However, based on the lower scores for the last two statements in Table 2 it appears that overall positive feelings about free samples do not necessarily translate into preferences for stores that provide free samples. This contradicts the popular belief of industry experts and may cause retailers to question the benefits of free samples for drawing consumers into the store. Another potential benefit of in-store sampling is its potential to increase consumer spending at the store level, possibly by causing consumers to linger in the store, to browse more, or by drawing them into aisles down which they otherwise might not venture. However, we found that the total expenditure of non-samplers was $74.05 while samplers spent $70.27, a difference that was not statistically significant (t=.66, p > 0.10). Given there was no statistical difference in the demographics of samplers versus non-samplers that would lead us to believe either population has a smaller average basket size, these results suggest that free samples do not seem to benefit the retailer by increasing overall shopping basket expenditure. Comparing planned versus unplanned purchases of samplers and non-samplers we find that 36% and 14% of samplers and non-samplers, respectively, were planning on making a purchase in the category prior to entering the store. This statistical difference (z = 4.05, p < 0.001) suggests that free samples are more likely to attract planned purchasers, something that the retailer would hope otherwise to be true, and even the manufacturer would be disappointed by unless the sample could cause those consumers to switch to their brand. Comparing the “planned to actual” category2 purchase rate we found the rate for nonsamplers to be 1.0. Among samplers we found that 93 samplers planned to make a category
This was the number of consumers who actually made a category purchase (not necessarily for the promoted brand) divided by the number of consumers who planned to purchase in the category prior to entering the store. 18
2
purchase prior to arriving at the store, and 141 actually did purchase, for a “planned to actual” category purchase rate of 1.52. This 52% bump in the “planned to actual” category purchase rate for samplers relative to non-samplers suggest that sampling may have the ability to increase category purchase rates for the promoted category on the day of the sample. This is obviously good news for retailers and should encourage them to find ways to boost the acceptance of free samples. Of greater interest to the manufacturer is the percentage of consumers who bought the promoted product given they made a purchase in the category (e.g., choice share). We found 104 of the 141 samplers who purchased in the category bought the promoted product resulting in a 76% choice share for the promoted product. Of the 13 non-samplers who made a purchase, 7 bought the promoted product resulting in a 54% choice share for the promoted product. The choice share in both cases is higher than the actual market share of any of the promoted products suggesting the free samples drew attention to, and promoted purchase of the promoted product, even among those who did not sample. However, we do find that the sample significantly increased choice share for samplers by 22% (z = 1.53, p < 0.065), suggesting the promotional effort is working for the brand. Finally, we were interested in comparing the percentage of samplers and non-samplers who switched from their preferred brand to the promoted brand given they made a purchase. Of the 104 samplers who purchased the promoted product, 63 (61%) switched from their preferred brand to the promoted one, while only 29% of non-samplers (2 out of 7) did the same. Although the sample size for non-samplers is small, the percentage of brand switchers to the promoted brand is significantly higher for samplers (z= 1.66, p < 0.05) and suggests that not only does the
19
free sample increase choice share among samplers, it increases brand switching as well, again good news for the manager of the promoted brand. Model and Measures We used a binary logit model of purchase incidence probability to test our hypotheses, where the probability of consumer n making a purchase is,
P =
n
exp β 0n + β n x n + e n
(
1 + exp β 0n + β n x n + e n
(
)
)
(1)
where, β 0n is the baseline utility of making a purchase for consumer n, x n is a vector of marketing variables and household specific characteristics explaining purchase incidence for consumer n, β n is a vector of parameters for variables x n , and en is a random error component. To test our hypotheses we needed to categorize the products in our study as either hedonic or utilitarian. Although many studies have categorized product categories as hedonic or utilitarian (Simonson, Carmon and O’Curry 1994, Chandon, Wansink and Laurent 2000, Voss, Spangenberg, and Grohmann 2003), there was very little overlap between the products studied in these papers and the ones we studied here. As such, we asked 52 staff members at a major southern, public university to rate the six product categories of interest on the ten-item hedonic/utilitarian (HED/UT) scale developed by Voss, Spangenberg and Grohmann (2003). For each product we calculated the average of the hedonic and utilitarian measures on the HED/UT 10-point scale and labeled those with an average score of seven or higher for the hedonic (utilitarian) measures as being hedonic (utilitarian) in nature. The average rating on the hedonic and utilitarian measures for each product, as well as the resulting values for the dummy categorical variables Hed_Prod and Util_Prod are shown in Table 3. We note that hedonism and utilitarianism are not necessarily two ends of a one-dimensional scale (Voss, Spangenberg and
20
Grohmann 2003), but rather products can be high or low on both measures simultaneously, and a low measure on one construct does not necessarily mean a high measure on the other (Crowley, Spengenberg and Hughes 1992). [Insert Table 3] To test H2 we created a variable Coupon that was equal to 1 if a coupon was presented with the free sample, 0 otherwise. To test H3 and H4 regarding product risks we created two additional dummy variables. New_Prod was set equal to 1 for the sampled products in our study that were new to the market (bagels, yogurt and sliced bread), and 0 otherwise. NotNat_Prod was set equal to 1 for non-nationally branded sampled products (frozen pizza, ice cream and sliced bread), and 0 otherwise. To capture the financial risks of the product we included the price of the sampled products on the day of the promotion (Price) in our model (see Table 1). To capture consumers’ perceptions of the sampling experience (H6 and H7) we asked consumers a series of questions about the hedonic/utilitarian nature of the sampling experience (e.g., Voss, Spangenberg and Grohmann 2003, Batra and Ahtola 1990; Engel, Blackwell, and Miniard 1993; Holbrook and Hirshman 1982; Bloch and Richins 1983a; Hirshman 1983). These questions are shown in Table 4. For each consumer we calculated the mean of the hedonic and utilitarian measures to create the variables Hed_Exp and Util_Exp, respectively. [Insert Table 4] To test H8 we created the variable Persuade which captured a consumer’s agreement with the statement, “The person distributing the sample tried to persuade me to buy the product,” where 1= strongly disagree and 7= strongly agree. To capture the motives for sampling (H9 and H10) we asked consumers early in the survey to rate (on a 7-point scale) their motives for taking the free sample. The items included
21
were derived from a pre-test that solicited motives for sampling among shoppers. Table 5 presents the top nine motives that emerged from this pretest. We conducted a factor analysis on these nine measures using principal components method with Varimax rotation. A three-factor solution capturing 62% of the variation was maintained based on the fact that the Eigen values for the first three factors were greater than one and those of the remaining factors were less than one. The three-vector solution was also supported by the scree plot. Cronbach’s alpha for this solution was 0.76. The factor loadings are also reported in Table 5 where the loading with the greatest value is highlighted for each variable. [Insert Table 5] (Positive) utilitarian evaluations have been linked to measures such as “usefulness”, “beneficial”, “important” and “valuable”, to name a few, and (positive) hedonic evaluations are believed to be captured with measures such as “in a good mood”, “happy”, “pleased”, “agreeable”, “nice” and “satisfied”, to name a few (Batra and Ahtola 1990; Kempf 1999; Mano and Oliver 1993). Batra and Ahtola (1990) found that sensory attributes such as taste contributed to hedonic components of attitudes. Scott and Yalch (1980) on the other hand treated “taste” as an informational variable. Based on this research and the factor loadings shown in Table 5 we label the first factor “Human Motives” (Hum_Motive) as it captures consumers motivated by human contact with others. The second factor we label “Hedonic Motives” (Hed_Motive) as it captures the measures typically considered hedonic in the literature mentioned above. The third factor we label “Utilitarian Motives” (Util_Motive) as it captures motives identify in the literature as goal oriented and practical in nature.
22
Finally, to test H11 we asked consumers to indicate their agreement with the statement, “I typically feel guilty after sampling a product,” where 1 = strongly disagree and 7= strongly agree. We label this variable Guilt. We also included some control variables to test for individual differences across consumers. These included a dummy variable to indicate whether the consumer was planning a purchase in the category (Planned), a variable Liked which captures (on a scale from 1 to 7) the degree to which the consumer liked the sampled product, both of which we expect to have a positive effect on purchase probability. We also include a set of demographic variables which captured gender, age, education and income level. Table 6 presents the variables discussed above and the following section describes our modeling approach and our findings. [Insert Table 6] We first estimated a binary logit model using the variables discussed above, that would test all of our proposed hypotheses. In addition to this specification we explored some additional specifications that captured more complex relationships of the variables under investigation. For example, we tested whether the effectiveness of a sample for reducing product risk depends on whether the product is hedonic or utilitarian by including interaction terms between the risk variables (New_Prod, NotNat_Prod, Price) and Hed_Prod and Util_Prod. We also explored the existence of a congruency (interaction) effect between the hedonic/utilitarian characteristics of the sampling experience (Hed_Exp and Util_Exp) with the hedonic/utilitarian motives for sampling (Hed_Motive and Util_Motive). All estimated models were compared based on their likelihood statistic and the Schwarz criterion for model selection (Schwarz 1978). The best fitting model was the one with the following specification: Pn = β0 + β1*Hed_Prod + β2*(Coupon*Util_Prod) + β3*(Coupon*Hed_Prod) +
23
β4(New_Prod*Hed_Prod) + β5*(New_Prod*Util_Prod)+ β6*Price + β7*(Hed_Exp*Hed_Prod)+ β8*(Hed_Exp*Util_Prod)+ β9*(Util_Exp*Hed_Prod)+ β10*(Util_Exp*Util_Prod)+ β11*Persuade + β12*(Hed_Motive) + β13*(Util_Motive)+ β14*Hum_Motive + β15*Planned + β16*Liked
The model is normalized on the probability of a consumer purchasing a sampled utilitarian product with no coupon, captured by β0. Parameters β1, β2, and β3 capture the probability of purchase for the three other possible cases -- hedonic product without a coupon,
(2)
utilitarian product with a coupon, and hedonic product with a coupon, respectively (H1 and H2). Parameters β4 and β5 capture the impact of the sampled product being new, and whether this is different for hedonic or utilitarian products, respectively (H3). β6 captures the impact of the price of the sampled product on purchase probability and parameters β7 - β10 capture the impact of the congruency between the hedonic/utilitarian perceptions of the sampling experience and the hedonic/utilitarian characteristic of the product (H6 and H7). The remaining parameter estimates are self-explanatory. We note that in testing different models we found the interaction between the hedonic/utilitarian motives with hedonic/utilitarian products (H9 and H10) to be insignificant and therefore were dropped these terms from our model. However, the main effects of the sampling motives were significant and hence were kept as show in Equation 2. Also, the impact of a non-national brand (H4) and the demographic variables tested proved not to be significant and therefore were not included in our final specification. The results for Equation 2 are presented in Table 7 and discussed below. [Insert Table 7]
Results
Looking at the four promotional cases (utilitarian sampled product with and without a coupon and hedonic sampled product with and without a coupon), we see consumers were most
24
likely to make a purchase for a utilitarian product accompanied by a coupon, based on the positive and significant sign on β2 (the only estimate that was positive and significant β0–β3). In fact consumers were least likely to buy a utilitarian sampled product without a coupon, based on the negative and significant sign for β0. Neither estimate for β1 and β2 were significantly different from zero, suggesting the probability of purchasing a hedonic sampled product was approximately the same with or without a coupon and was greater than that of a utilitarian product without a coupon but less that than of a utilitarian product with a coupon. These results support our hypotheses regarding the congruency between the product and the promotion in H1 and H2. Specifically, they support that hedonic sampled products are more likely to be purchased relative to a utilitarian sampled products (H1), and that a cents-off coupon provided with the free sample will be more effective for boosting purchase for a utilitarian product than a hedonic one (H2). These findings highlight the effectiveness of investing in free samples for hedonic products and the importance of providing a coupon with a utilitarian sampled product to increase purchase incidence. Because we do not have a condition for “coupon only”, it is not clear if the combination of “sample and coupon” would perform better than just a coupon for either product type. However, we at least have shown the combination of the sample-coupon combination has greater impact than the sample alone for utilitarian products. Hedonic products on the other hand do not seem to benefit from a coupon -- the free sample is sufficient. Next, we examine the impact of free samples on risky products, specifically new products and non-nationally branded products, as discussed in H3 and H4. As shown in Equation 3, our best fitting model did not include an effect of non-national brands as this term was not significant. This suggested there is no statistical difference between the post-sample purchase probability of a sampled national and non-national brand, supporting H4. However, our best
25
fitting model did identify an effect of new products, just not as expected. Decomposing the effect of hedonic and utilitarian new products we found that while new utilitarian products were just as likely to be purchased post-sample as non-new utilitarian products (β3 was insignificant supporting H3), new hedonic products were actually more likely to be purchased than their nonnew hedonic counterparts, based on the positive and significant estimate for β4. Perhaps this could be driven by the fact that when consumers sample new products their purpose is to decide whether or not to buy the product, and if they like it they buy. Whereas, with a non-new product, if the consumer has already had an experience with the product, the sample should have less of an impact on whether they buy the product and the choice may be more tied to their inventory at home for the product. In conclusion we find support for H3 and H4 suggesting free samples reduce the risks associated with certain risky products, specifically new and non-national brands. Examining the financial risk of purchasing a product, and the impact a free sample can have on reducing that risk (H5), we see that the parameter estimate for price is actually positive and significant, showing that within the range of prices we studied ($0.50 - $3.33), higher priced sampled products were more likely to be purchased relative to less expensive sampled products. This interesting support for H5 could suggest that a free sample not only reduces the financial risk of a sampled product, but it also encourages consumers to buy something they otherwise may not have given its higher price. Whatever the explanation, the free sample seems to mitigate the impact of price and even encourage consumers to purchase more expensive products. Looking at the impact of the congruency between a hedonic/utilitarian sampling experience with a hedonic/utilitarian product (H6 and H7) we see that the parameter estimates for β7 –β9 were all insignificant and the estimate for β10 was significant and positive. This suggests that a hedonic sampling experience does not increase the likelihood of purchase,
26
whether the product sampled is hedonic or utilitarian and that a utilitarian sampling experience does not increase the likelihood of purchase for a hedonic product (no support for H6). However, we do see that consumers are more likely to purchase a utilitarian product if the sampling experience is utilitarian in nature (e.g., the person distributing the sample provides useful information and the experience is informative in nature), supporting a congruency effect between the sampling experience and the product characteristics for utilitarian products (H7). The findings highlight the importance of training sample providers on the factual details of utilitarian products and the importance of providing consumers with that rational information. They also suggest that efforts to create a festive and pleasing sampling experience will have little to no impact on purchase whether the product is utilitarian or hedonic. H8, which predicted that consumers who feel persuaded to buy a sampled product will be less likely to buy the product was supported based on the negative and significant estimate on β11. This should encourage marketers to emphasize to sample distributors the importance of not creating a “hard sell” experience for the consumer. Next, there was no evidence to support a congruency effect between the motives to sample the product and the product characteristics (H9 and H10). However, we did find evidence of a main effect of sampling motives. Specifically, based on the significant and negative sign on β12 and the insignificant estimates for β13 and β14 we discovered that consumers with hedonic motives to sample were less likely to purchase the sampled product than those who sampled for utilitarian reasons or in order to interact with the sample provider or others sampling. This suggests that marketers can benefit from encouraging consumers to sample for utilitarian reasons, perhaps by in-store signage to emphasize the informative and rational benefits of sampling the products in the store that day.
27
Finally, we note that feelings of guilt as a result of sampling had no impact on consumers’ decision to buy the product, hence the variable Guilt was dropped from our model and H11 was not supported. This was surprising given that in our exploratory research to motivate this study many consumers expressed feelings of guilt as a result of sampling. In fact, the mean score on the question, “I typically feel guilty after sampling a product,” was close to two (on a seven point scale where 1=strongly disagree) for both samplers and non-samplers suggesting guilt may not be as prominent an emotion as earlier research indicated. Examining the control variables we see that consumers who planned to purchase in the promoted category prior to arriving at the store and those who expressed liking the taste of the sample were more likely to buy the sampled product as expected. This is not surprising and simply emphasizes some of the variables over which the marketer has no control in influencing purchase of the sampled product. Finally, the lack of support for the inclusion of demographic variables (gender, age, education, or income) in our model suggests that any efforts to encourage certain consumers to sample a product based on their demographic profile should not increase the sale of the product that day.
Sensitivity Analysis
We conducted some sensitivity analyses to better understand the change in purchase probability post-sample given different values of the significant variables in our model. Holding all continuous variables at their means and all dummy variables at their mode, we calculated the difference in purchase probability when varying a particular variable in the model. For example, when comparing the purchase probability of a hedonic product that was new to the market versus one that was not we see the purchase probability increases 47%, from 0.36 to 0.83, when the
28
hedonic product is new versus when it already exists in the market place. We also see that a utilitarian product is almost 75% more likely to be purchased with a coupon (0.54 choice probability) versus without a coupon (0.31 choice probability). Examining the impact of the sampling experience we see that choice probability drops approximately 68% (from .52 to .31) if the consumer perceives the experience to be persuasive in nature. However, increasing consumers’ perceptions of the sampling experience as being “utilitarian” (e.g., providing useful information about the product, being impartial to the brand, etc.) by only one point on a sevenpoint scale can increase the likelihood of purchase of a utilitarian product by 7%. These sensitivities illustrate how decisions of the type of product to promote with a free sample, and the way in which that promotion is conducted, can have dramatic impact on the likelihood that a sampler will buy the product once sampled.
6. Discussion
This paper focuses on the impact of in-store free samples, specifically consumable instore free samples at a grocery store, on “sample takers” (those who accept the free sample). We also compare the behavior and attitudes of non-samplers to sample takers to provide insights there as well. Focusing on sample takers, we develop a set of hypotheses about the probability of purchase post-sample, many of which are based on a benefit congruency framework that would suggest when the characteristics of the product, the sampling experience, and consumers’ sampling motives are congruent, the likelihood of purchase will be greater. We conduct an in-store experiment that leverages the sampling program of a mid-western grocery store chain. Our study allows us to investigate whether free samples influence store choice or perceptions of store quality, the reasons why consumers either take or avoid in-store free samples, and to compare the purchase probability between sample takers and non-samplers.
29
We build a model of samplers’ post-sample purchase probability using a binary logit model that tests our hypotheses and provide some useful insights for manufacturers who invest in in-store free sample promotions. In general we find that while consumers state they look forward to free samples and that free samples make a store more festive, they do not report choosing a store based on the fact that it provides free samples or conducting their shopping when they know free samples will be available. Of course, this conclusion was based on stated preferences and not actual behavior and therefore the impact of free samples on store traffic probably deserves more investigation. Comparing the behavior and attitudes of samplers to non-samplers we find there is no statistical difference between the shopping basket expenditure of the two populations, nor was there evidence to support that free samples have the power to draw consumers into the category being promoted if they were not planning on making a purchase in that category. These results are most likely disappointing to store managers who tend to believe that free samples increase store traffic and spur unplanned category purchases. On a brighter note however, we find that sampling, 1) increases the “planned to actual” category purchase rate for samplers relative to non-samplers, 2) increases choice share of the promoted brand among category purchasers that day, and 3) increases brand switching to the promoted brand on the day of the sample, all good news for retailers and manufacturers alike. Focusing on sample-takers, we find that hedonic sampled products are more likely to be purchased relative to utilitarian sampled products, and that a cents-off coupon provided with the free sample are effective for boosting purchase for a utilitarian product than a hedonic one. These findings highlight the effectiveness of free samples for hedonic products and the
30
importance of providing a coupon with a utilitarian sampled product to increase purchase incidence. Next, we find that consumers are more likely to purchase a hedonic sampled product if it is new, but that a utilitarian product was no more likely to be purchased if it was a new versus existing on the market. We also found that the more expensive the sampled product, the more likely it was to be purchased. These results suggest that free samples are an effective tool for boosting purchase incidence of new hedonic products and more expensive products as well. Next, we found that consumers are more likely to purchase a utilitarian product if the sampling experience is utilitarian in nature (e.g., the person distributing the sample provides useful information and the experience is informative in nature). However, a hedonic experience had no positive impact on purchase probability, whether the product was hedonic or utilitarian. These findings highlight the importance of training sample providers on the factual details of a utilitarian product and the importance of providing consumers with that rational information. They also suggest that efforts to create a festive and pleasing sampling experience will have little to no impact on purchase. The results that examined the impact of a persuasive sample distributor showed that consumers who felt persuaded were less likely to purchase the sampled product than those who did not report such an experience. This should encourage marketers to emphasize to sample distributors the importance of not creating a “hard sell” experience for the consumer. Prior to estimating our model we identified three main consumer motives for sampling: 1) the desire for human contact, 2) the desire for a pleasant experience, and 3) the desire for an informative experience. While the first and third motives were insignificant in predicting purchase, we found that consumers who sampled for the pleasant experience that sampling could
31
provide and that could make their shopping trip more enjoyable were less likely to purchase than other consumers. This provides more support for the benefits of creating a utilitarian sampling experience, or at least the ineffectiveness of a hedonic one for promoting purchase. Encouraging consumers to sample for utilitarian reasons could be accomplished, for example, by using instore signage at the front of the store emphasizing the informative and rational benefits of the samples in the store that day. Finally, the control variable in our model revealed, not surprisingly, that consumers who planned to purchase in the promoted category prior to arriving at the store and those who expressed liking the taste of the sample were more likely to buy the sampled product as expected. We also found no evidence to support that any demographic segment was more likely to purchase post-sample than another. While this paper provides some wonderful insights into the power and limitations of free samples for issues important to retailer and manufacturers, it only begins to touch upon the many complex aspects of this type of promotion and leaves many directions for future research. To begin, while this study provides real world validity to many of the experimental studies mentioned earlier, it might be advantageous to test the theory developed in this paper in a mockgrocery setting where more variables could be controlled to study the causes of the behavior we were only able to observe but for which we were not able to explicitly test the drivers. For example, it would be interesting to more systematically control and test for the impact of the demeanor of the sample distributor and the information he/she provides. In fact, it would be interesting to compare the impact of a free sample for two conditions – one where a sample distributor was present and one where he/she was not.
32
Another direction would be to study the impact of different types of free samples (e.g., consumable grocery store products vs. product interactions and demonstrations for durables) across different types of retail formats to see if the results from this paper may be generalized across different types of samples and retail outlets.
33
Table 1 Descriptive Data on Sampled Products and Survey Participants
Product Category Sampled Ground Turkey Bagels Frozen Pizza Yogurt Ice Cream Bread Demographics of Participants % males % older than 55 % with college degree % employed full-time Average income Average household size Was the Sampled Product a New Product? No Yes No Yes No Yes 19% 43% 56% 49% $51,000 2.8 people Was the Sampled Product a National Brand? Yes Yes No Yes No No Average Price of Sampled Product $1.59 $2.49 $3.33 $0.50 $2.50 $2.00 % of Observations for Each Product 10% 9% 5% 31% 17% 28%
34
Table 2 The Impact of Free Samples on Store Traffic and Store Perceptions
“How much do you agree with the following statements about in-store free samples?” (1=strongly disagree; 7=strongly agree)
Mean (SD)1 5.20 (1.99) 4.64
a,b b a
I look forward to receiving free samples when I do my grocery shopping Store that provide free samples are more festive Store that provide free samples are of better quality I choose my grocery store based on the free samples they provide I do my grocery shopping when I know there will be free samples
1
(2.14)
3.52 (2.19) 2.77 (1.98) 2.03 (1.51)
c c
N=354; a,b,cValues indicated with the same letter a,b or c were not statistically different.
Table 3 Mean Scores on HED/UT Scale Measures and Resulting Values for the variables Hed_Prod and Util_Proda
Ground Turkey 2.51 5.75 Bagels 6.17 7.43 Frozen Pizza 5.85 6.01 Ice Cream 9.04 6.44 1 0 Yogurt 7.17 7.19 1 1 Bread 4.51 8.45 0 1
Average on Hedonicb Measures Average on Utilitarianc Measures
Value for Dummy Hedonic 0 0 0 Variable Hed_Prod Value for Dummy Utilitarian 0 1 0 Variable Util_Prod a All measures were captured using a 10-point scale b Hedonic Measures = Fun, Enjoyable, Thrilling, Delightful, Exciting c Utilitarian Measures = Helpful, Functional, Practical, Necessary, Effective
35
Table 4 Statements Capturing Hedonic/Utilitarian Perceptions of the Sampling Experience
Hedonic Statementa The person distributing the sample was nice. The person distributing the sample put me in a good mood. I liked the person distributing the sample. The person distributing the sample made me feel comfortable. The overall sampling experience put me in a good mood. The overall sampling experience was pleasing. The person distributing the sample provided use information about the product. The person distributing the sample brought important attributes to my attention. The person distributing the sample was impartial to the brand. The overall sampling experience was informative. The overall sampling experience was helpful in making an informed purchasing decision.
Utilitarian
a
All statements were measured on a 7-point scale where 1= strong disagree, 7= strongly agree.
36
Table 5 Factor Loadings for Responses to the Question, “How Well Do the Following Statements Describe Why You Take Free Samples?”
Factor Labels and Loadings Factor 1: Factor 2: “Human “Hedonic Reason for Taking a Free Sample Motives” Motives” I wanted to see how the product tasted (6.19a) -0.031 0.126 I enjoy sampling new products (5.98) 0.063 0.637 The sample was free (5.60) 0.112 0.782 I wanted to learn more about the product (5.51) 0.212 -0.056 The person distributing the sample invited me to taste it (5.33) 0.833 -0.057 I saw other people taking the free sample (5.11) 0.727 0.198 I was hungry (4.83) 0.103 0.656 The person distributing the sample seemed nice (3.53) 0.834 0.186 Samples make the shopping experience more enjoyable (3.48) 0.497 0.567 % Variation Explained by Each Factor .33 .17 a Mean on a 7-pt scale where 1= strongly disagree, 7= strongly agree Factor 3: “Utilitarian Motives” 0.820 0.440 0.095 0.701 0.244 -0.040 -0.415 0.004 0.175 .15
37
Table 6 Variables Definitions Variable Label Variable Definition = 1 if promoted product was hedonic, 0 otherwise = 1 if promoted product was utilitarian, 0 otherwise = 1 if a coupon was presented with the promoted product, 0 otherwise = 1 if promoted product was a new product, 0 otherwise = 1 if promoted product was not a national product, 0 otherwise The price of the promoted product The average of a consumer’s agreement with six statements capturing his/her hedonic impressions of the sampling experience (1=strongly disagree and 7= strongly agree) The average of a consumer’s agreement with five statements capturing his/her utilitarian impressions of the sampling experience (1=strongly disagree and 7= strongly agree) Agreement with the statement, “The person distributing the sample tried to persuade me to buy it.” (1=strongly disagree and 7= strongly agree) Factor score indicating consumer sampling motives related to the desire for human contact Factor score indicating consumer sampling motives related to the hedonic benefits of sampling. Factor score indicating consumer sampling motives related to the utilitarian benefits of sampling. Agreement with the statement, “I typically feel guilty after sampling a product.” (1 = strongly disagree and 7= strongly agree) =1 if the consumer was planning to make a purchase in the category of the promoted product, 0 otherwise Agreement with the statement, “I liked how the product tasted.” (1 = strongly disagree and 7= strongly agree) =1 if the consumer was male, 0 otherwise Consumer’s age =1 if the consumer had a college degree, 0 otherwise Household income
Hed_Prod Util_Prod Coupon New_Prod NotNat_Prod Price Hed_Exp Util_Exp Persuade Hum_Motive Hed_Motive Util_Motive Guilt Planned Liking
Gender Age Educa Income
38
Table 7 Results for the Binary Logit Model of Purchase Incidence Among Samplers Parameter Variable Label Hypotheses (Expected Sign) Maximum Likelihood Estimate -7.61*** 0.84 0.94** 1.13 2.15** -0.63 0.73** -0.74 0.02 -0.02 0.27** -0.86** -0.29* -0.21 -0.10 1.44*** 0.60***
*** **
t-statistic -4.00 0.59 1.65 0.93 1.84 -0.99 2.06 -0.33 0.09 -0.08 1.66 -2.13 -1.61 -0.97 -0.49 3.94 2.27
β0 β1 β2 β3 β4 β5 β6 β7 β8 β9 β10 β11 β12 β13 β14 β15 β16
Util_Proda Hed_Prod Coupon*Util_Prod Coupon*Hed_Prod New_Prod*Hed_Prod New_Prod*Util_Prod Price Hed_Exp*Hed_Prod Hed_Exp*Util_Prod Util_Exp*Hed_Prod Util_Exp*Util_Prod Persuade Hed_Motive Util_Motive Hum_Motive Planned Liking
H1 (n/a) H1 (+) H2 (+) H2 (n/a) H3 (ns) H3 (ns) H5 (ns or -) H6 (+) H6 (ns) H7 (ns) H7 (+) H8 (-) n/a n/a n/a n/a (+) n/a (+)
N=259 Log-Likelihood = -149.47 U2 = 1-LL / LL(C)= .195b
a b
p < 0.01 p < 0.05 * p < 0.10
Model is calibrated on the base case of a utilitarian sampled product without a coupon. The U2 statistic (Hauser 1978) is similar to the R2 statistic used in regression, but is appropriate for probabilistic models of choice, where LL(C) is the Log-Likelihood statistic for a model with intercept only.
39
Adaval, Rashmi (2001), “Sometimes It Just Feels Right: The Differential Weighting of Affect-Consistent and Affect-Inconsistent Product Information,” Journal of Consumer Research, 28 (June), 1-17.
Batra, R. and O.T. Ahtola (1990), “Measuring the Hedonic and Utilitarian Sources of Consumer Attitudes,” Marketing Letters, 2, 159-170.
Batra, R. and I. Sinha (2000), “Consumer-Level Factors Moderating the Success of Private Label Brands,” Journal of Retailing, 76(2), 175-191.
Bawa, Kapil and Robert Shoemaker (2004), “The Effects of Free Sample Promotions on Incremental Brand Sales,” Marketing Science, 23 (3), 345-363.
Belch, G.E. and M.A. Belch (1990), Introduction to Advertising and Promotion Management, Homewood, IL: Richard D. Irwin.
Bellenger, Danny N., Earle Steinberg, and Wilbur W. Stanton (1976), “The Congruence of Store Image and Self Image,” Journal of Retailing, 52 (Spring), 17-32.
Bettinger, Charles O., Lydon E. Dawson, Jr. and Hugh G. Wales (1979), “The Impact of Free-Sample Advertising,” Journal of Advertising Research, 19 (3), 35-40.
40
Blattberg, Robert C. and Scott A. Neslin (1990), Sales Promotion: Concepts, Methods, and Strategies, Englewood Cliffs, NJ: Prentice Hall.
Bloch, Peter H. and Marsha L. Richins (1983a), “Shopping Without Purchase: An Investigation of Consumer Browsing Behavior,” in Advances in Consumer Research, 10, ed. Richard P. Bagozzi and Alice M. Tybout, Ann Arbor, MI: Association of Consumer Research, 389-393.
Chandon, Pierre, Brian Wansink, and Gilles Laurent (2000), “A Benefit Congruency Framework of Sales Promotion Effectiveness,” Journal of Marketing, 64 (October), 6581.
Cialdini, Robert B., (1993) Influence: Science and Practice, 3rd ed. New York, NY: HarperCollins College Publishers.
Crowley, Ayn E., Eric R. Spangenberg, and Keven R. Hughes (1992), “Measuring the Hedonic and Utilitarian Dimensions of Attitudes Toward Product Categories,” Marketing Letters, 3 (July), 239-249.
Darian, J.C., L. A. Tucci, and A. R. Wiman (2001), “Perceived Salesperson Service Attributes and Retail Patronage Intentions,” International Journal of Retail & Distribution Management, 29 (5), 205-213.
41
Derbaix, C. (1983), “Perceived Risk and Risk Relievers: An Empirical Investigation,” Journal of Economic Psychology, 3 (1), 19-38
Donovan, Robert J., John R. Rossiter, Gilian Marcoolyn, and Andrew Nesdale (1994), “Store Atmosphere and Purchasing Behavior,” Journal of Retailing, 70, 283-294. Dowling, G.R., Staelin, R., (1994). “A Model of Perceived Risk and Intended Risk-Handling Activity”. Journal of Consumer Research 21, 119–134
Engel, James F., Roger D. Blackwell, and Paul W. Miniard (1993), Consumer Behavior, Chicago: Dryden.
English, H.B. and A.C. Macker (1976), A Comprehensive Dictionary of Psychology and Psychoanalytical Terms, David McKay Company, New York, NY.
Fitzgerald, Kate (1996), “Survey: Consumers Prefer Sampling over Coupons,” Advertising Age, January 29, 1996, pg. 9
Gadenk, Karen and Scott A. Neslin (1999), “The Role of Retail Promotion in Determining Future Brand Loyalty: Its Effect on Purchase Even Feedback,” Journal of Retailing, 75 (4), 43-59.
Golden, Linda L. and Mary R. Zimmer (1986), “Relationship Between Affect, Patronage, Frequency and Amount of Money Spent With a Comment on Affect Scaling and
42
Measurement,” In: R.J. Lutz (ed.), Advances in Consumer Research, Vol. 13, Provo, UT: Association for Consumer Research.
Hamm, B. Curtis, Michael Perry and Hugh F. Wynn (1969), “The Effect of a Free Sample on Image and Attitude,” Journal of Advertising Research, 9 (4), 35-38.
Havlena, William J. and Morris B. Holbrook (1986), “The Varieties of Consumption Experience: Comparing Two Typologies of Emotion in Consumer Behavior,” Journal of Consumer Research, 13 (December), 394-404.
Heilman, Carrie M., Kent Nakamoto and Ambar Rao (2002), “Pleasant Surprises: Consumer Response to Unexpected In-Store Promotions,” Journal of Marketing Research, 34 (May), 242-252.
Heiman, Amir, Bruce McWilliams, Zhihua Shen and David Zilberman (2001), “Learning and Forgetting: Modeling Optimal Product Sampling Over Time,” Management Science, 47 (4), 532-546.
Hirschman, Elizabeth C. (1983), “Predictors of Self-Projection, Fantasy Fulfillment, and Escapism,” Journal of Social Psychology, 120 (June), 63-76.
43
Holbrook, Morris B., and Elizabeth C. Hirschman (1982), “The Experiential Aspects of Consumption: Consumer Fantasies, Feelings and Fun,” Journal of Consumer Research, 9 (September), 132-140.
Jacoby, Jacob and Leon B. Kaplan (1972), “The Components of Perceived Risk,” in M. Venkatesan, ed., Proceedings of the Third Annual Conference of the Association of Consumer Research, College Park, MD: Association for Consumer Research.
Jain, Dipak, Vijay Mahajan, and Eitan Mullen (1995), “An Approach for Determining Optimal Product Sampling for the Diffusion of a New Product,” Journal of Product Innovation Management, 12, 124-135.
Kempf, DeAnna S. (1999), “Attitude Formation from Product Trial: Distinct Roles of Cognition and Affect for Hedonic and Functional Products,” Psychology and Marketing, 16 (1) 35-50.
---- and Robert E. Smith (1998), “Consumer Processing of Product Trial and the Influence of Prior Advertising: A Structural Modeling Approach,” Journal of Marketing Research, 35 (August), 325-338.
Konopacki, Allen (1996), “Successfully Selling to Most Women,” Agency Sales Magazine, 26(1), 45-46.
44
Konopacki, Allen (1997), “Tips for Successful Exhibiting at Trade Shows,” Agency Sales Magazine, 27 (6), 19-21.
Lammers, Bruce H. (1991), “The Effect of Free Samples on Immediate Consumer Purchase,” Journal of Consumer Marketing, 8 (2), 31-7.
Lindstedt, Sharon (May 24, 1999), “Tops Supermarkets in Western New York Entice Shoppers With Free Food Samples,” Buffalo News, B3.
Littler, Dale A. and D. Melanthiou (2005), “The Uncertainty of Risk or the Risk of Uncertainty,” International Congress on Marketing Trends, Paris, France.
Mano, Haim and Richard Oliver (1993), “Assessing the Dimensionality and Structure of Consumption Experience: Evaluation, Feeling and Satisfaction.” Journal of Consumer Research, 20 (December), 451-466.
Marks, Lawrence J. and Michael A. Kamins (1988), “The Use of Product Sampling and Advertising: Effects of Sequence of Exposure and Degree of Advertising Claim Exaggeration on Consumers’ Belief Strength, Belief Confidence, and Attitudes,” Journal of Marketing Research, 25 (August), 266-81.
McGuinness, Dalton, Mike Brennan and Philip Gendall (1995), “An Empirical Test of Product Sampling and Couponing,” Journal of the Market Research Society, 37, 2 159170.
45
Mitchell, V.W., and Boustani, P. (1994), “A Preliminary Investigation Into Pre- and PostPurchase Risk Perception and Reduction,“ European Journal of Marketing 28 (1), 56–71.
Moses, Lucia, “Food City Expands Sampling Program,” Supermarket News, April 4, 2005, pg. 43.
Mowen, John C. and Michael Minor (1998), Consumer Behavior, 5th Ed., Chapter 14, Upper Saddle River, NJ: Prentice Hall.
Nowlis, Stephen M. and Baby Shiv (2005), “The Influence of Consumer Distractions on the Effectiveness of Food-Sampling Programs,” Journal of Marketing Research, 42 (May), 157-168.
Oliver, Richard L. (1992), “An Investigation of the Attribute Basis of Emotion and Related Affects in Consumption: Suggestions for a Stage-specific Satisfaction Framework,” in Advances in Consumer Research, 19, ed. John F. Sherry, Jr. and Brian Sternthal, Provo, UT: Association for Consumer Research, 237-244.
Roselius, T. (1971), “Consumer Rankings of Risk Reduction Methods,“ Journal of Marketing, 35 (1), 56–61.
46
Ross, I. (1975) Perceived Risk and Consumer Behavior: A Critical Review, Advances in Consumer Behavior, 2(1), 1-19
Rossiter, J.R. and L. Percy (1987). Advertising and Promotion Mangement. New York: McGraw-Hill.
Rothschild, Michael L. and William C. Gaidis (1981), “Behavioral Learning Theory: Its Relevance to Marketing and Promotions,” Journal of Marketing, 45 (Spring), 70-78.
Schultz, Don E., William A. Robinson and Lisa A. Petrison (1993), Sales Promotion Essentials, Lincolnwood, IL: NTC Business Books.
Schwarz, Gideon (1978), “Estimating the Dimension of a Model,” Annals of Statistics, 6 (2), 461-464.
Scott, Carol A. and Richard F. Yalch (1980), “Consumer Response to Initial Product Trial: A Bayesian Analysis,” Journal of Consumer Research, 7 (1), 32-41.
Sherman, Elaine and Ruth B. Smith (1987), “Mood States of Shoppers and Store Image: Promising Interactions and Possible Behavioral Effects,” In: M Wallendorf and P. Anderson (eds.), Advances in Consumers Research, Vol. 14, Provo, UT: Association for Consumer Research.
47
Simonson, Itmar, Ziv Carmon and Suzanne O’Curry (1994), “Experimental Evidence on the Negative Effect of Product Features and Sales Promotions,” Marketing Science, 13 (1), 23-40.
Smith, R.E. (1993), “Integrating Information From Advertising and Trial: Processes and Effects on Consumer Response to Product Information,” Journal of Marketing Research, 30, 204-219.
---- and W. R. Swinyard (1983), “Attitude-behavior Consistency: The Impact of Product Trial Versus Advertising,” Journal of Marketing Research, 20, 257-267.
---- and ---- (1988), “Cognitive Response to Advertising and Trial: Belief Strength, Belief Confidence and Product Curiosity.” Journal of Advertising, 17 (3), 3-14.
Steingberg, Sandon A., and Richard F. Yalch (1978), “When Eating Begets Buying: The Effects of Food Samples on Obese and Non-obese shoppers,” Journal of Consumer Research, 4 (March), 243-246.
Tam, J. L.M. and Y.H. Wong (2001), “Interactive Selling: A Dynamic Framework for Services,” Journal of Services Marketing, 15 (5), 379-396.
48
Voss, Kevin E., Eric R. Spangenberg, and Bianca Grohmann (2003), “Measuring the Hedonic and Utilitarian Dimensions of Consumer Attitude,” Journal of Marketing Research, 40 (30), 310-320.
Westbrook, Robert A. (1987), “Product/Consumption-based Affective Responses and Postpurchase Processes,” Journal of Marketing Research, 24 (August), 258-270.
---- and Richard L. Oliver (1991), “The Dimensionality of Consumptions Emotion Patterns and Consumer Satisfaction,” Journal of Consumer Research, 18 (June), 84-91.
Wright, Alice Ann and John G. Lynch Jr. (1995), “Communication Effects of Advertising Versus Direct Experience when Both Search and Experience Attributes Are Present,” Journal of Consumer Research, 21 (March), 708-718.
Zwiebach, Elliot, “Marsh Increases In-Store Sampling,” Supermarket News, February 28, 2005, pg. 24.
49