This paper (Chapter 3 of a dissertation on online shopping behavior) evaluates the role of reference price on consumer behavior at the online channel, focusing on whether it plays a relevant role, how it is formed, the importance of two broad types of reference prices, internal reference price (IRP) and external reference price (ERP), and potential differences to those effects observed offline. Using a unique data set of online and offline grocery purchases from the same sample of consumers of a large multichannel grocery chain, we analyze purchases and unplanned no-purchases in the orange juice category for consumers who face equal prices across online and offline channels, and use multinomial logit models which account for IRP and ERP effects. To avoid that IRP and ERP drop out of our multinomial estimations, we use item-specific measures for both of them. IRP is measured as the price observed for each alternative by consumers on their last visit to the store, whereas ERP is measured according to four different formulations: ERP (mean), ERP (low), ERP (high) and Rank. In our data we find that ERP (mean) performs better than any other formulation for ERP. One of the main contributions of the chapter is that reference price effects on consumer behavior differ across online and offline channels. Regarding reference price in general, we find that the impact of reference price is greater offline. Regarding IRP and ERP, we find that at both channels, the impact of IRP is greater than that of ERP, but that the relative impact of IRP with regard to ERP is greater online than offline. Besides, results show that whereas the impact of IRP is similar online and offline, the impact of ERP is significantly greater offline; in fact ERP is not significant online. Additionally, the influence of purchase frequency on the impact of IRP and ERP on brand choice across channels is evaluated. For frequent shoppers, the effects of IRP and ERP coincide with those observed for the general population. However, they are slightly different for infrequent shoppers. Results indicate that the role of ERP increases considerably for infrequent shoppers compared to frequent shoppers, and that it becomes significant in the online channel for infrequent shoppers. These results are of remarkable interest for retailers because they would help them to organize correctly the timing of promotion and price changes online. Moreover, these findings become especially relevant for multi-channel retailers operating both online and offline. As shown by a simulation detailed in the chapter, a change in price impacts IRPs and ERPs, and these changes in IRPs and ERPs influence alternatives’ choice probabilities in the category. This change in alternative’s choice probabilities translate into differences in sales’ distribution in the category and, consequently, on the revenues obtained by the retailer. If retailers do not consider that the impact of reference prices differs across channels, they may follow inadequate pricing strategies for the online and offline channels. In fact, for the category of orange juice, we find that the grocery retailer could increase its revenues by 3.43% if it discriminated prices across shopping channels
