The Impact of Household Level Heterogeneity in Reference Price Effects on Optimal Retailer Pricing Policies

The field of marketing has witnessed substantial improvement in modeling household level heterogeneity. However, relatively little has been written about how modeling household heterogeneity translates into better marketing decisions. In this paper, we study the role of incorporating household heterogeneity in reference price effects on a retailer’s pricing policy. In order to study the impact of heterogeneity in reference price effects on retail pricing we use a hierarchical Bayes nested logit model that provides estimates for gain and loss effects at the household level. By using household level estimates and a dynamic programming algorithm, we then develop a normative pricing policy for a retailer maximizing category profit. In the empirical analysis, we find that although at the aggregate level (i.e. no heterogeneity case), the impact of a gain and loss is about the same, for a significant number of households a gain has higher impact than a corresponding loss. Our results indicate that the optimal pricing policy derived from the heterogeneous case is qualitatively different from the case when heterogeneity is ignored. We suggest that when there is household heterogeneity in reference price effects, it is important for the retailer to (a) consider the joint distribution of gains and losses among the households and (b) promote brands that are highly preferred by the “gain seeking” households. Retail pricing policy based upon a model that incorporates household heterogeneity is also shown to be more profitable. INTRODUCTION Over the last ten years, the field of marketing has witnessed substantial improvement in modeling household level heterogeneity. The progression from aggregate (e.g. Guadagni and Little 1981) to latent class (Kamakura and Russell 1989) to hierarchical Bayes (Allenby and Lenk 1994) models reflects the inroads we have made in characterizing household differences. While these methodological advances are commendable, relatively little has been written about the impact of modeling household heterogeneity on marketing decisions. In this paper, we demonstrate the pricing and category profit implications of incorporating heterogeneity in reference price effects for a retailer. We show that for an important marketing problem pertaining to a retailer, the optimal pricing decisions for various brands in a category are inextricably related to household heterogeneity in reference effects and brand preference. A considerable amount of research exists on reference prices (see Kalyanaram and Winer 1995 for a review). Reference prices are certain anchors or standards that households use to compare the observed purchase price of a product against. If the observed price is greater than the reference price it is perceived as a “loss”. On the other hand, if the observed price is less than the reference price it is perceived as a “gain”. Empirical evidence regarding the relative impact of perceived gains and losses on household choice has been quite mixed. For example, Putler (1992) found that, consistent with prospect theory (Kahneman and Tversky 1979), the effect of a loss on demand is greater than that of an equal gain. Greenleaf (1995), on the other hand, finds that the effect of a gain is greater than that of a loss. While investigating the effects of reference price on household utility is of considerable theoretical interest, an equally interesting issue is the normative impact of these effects on a retailer's pricing policy. Some recent papers (Kopalle, Rao, and Assunção 1996, Greenleaf 1995) in marketing have studied the normative implications of reference price effects from the 2 standpoint of a retailer. Kopalle et al.’s (1996) results suggest that when the impact of a price gain is greater than that of a loss, “hi/lo” prices are optimal; on the other hand, when the impact of a loss is greater than that of a gain, constant prices are optimal. This result is consistent with Greenleaf’s (1995) monopoly analysis. A limitation of the above normative models is that they do not take into consideration household level heterogeneity in reference price effects. As noted above, current normative research on reference prices suggests that at an aggregate level, when the impact of a gain is greater than that of a loss, the retailer should promote. However, uncovering heterogeneity is likely to reveal that not all households are the same with respect to the reference price effects (as well as brand preference and price response); for some, the impact of a loss may be larger than that of a gain and for others, the reverse may hold. Further, the magnitude of the gain and loss effects may also vary across households. Under such (perhaps common) circumstances, it is not clear whether a retailer should price promote, and if so, which brands to promote. Therefore normative policies based on a model that does not account for heterogeneity could potentially result in a pricing policy that does not maximize retailer profits. Also, from a methodological standpoint, recent empirical research suggests that it is important to formally consider heterogeneity in reference effects because aggregate models tend to overstate the magnitude of the model estimates. For example, Chang, Siddarth, and Weinberg (1999) use a hierarchical Bayes model to show that upon accounting for the price response heterogeneity, the sticker shock effect gets diminished. Similarly, Bell and Lattin (2000) use a latent class model to show that once the price response heterogeneity is taken into consideration,