Adding store to web: migration and synergy effects in multi-channel retailing

The purpose of this paper is to investigate the topic of multi-channel retailing. Specifically, the research intends to determine if and to what extent the opening of physical stores by a former web-only retailer reduces or extends overall retail sales, and whether such effects tend to change over time. Empirical analysis focuses on data elaboration from a retailer who has passed from the initial mono-channel model (pure online), to a multi-channel one with the opening of stores.

Through the analysis of an internal data set of a leading consumer electronics retailer applying Probit and Logit estimation techniques, the authors extract information about actual customers’ purchases (or rather retail sales) in three newly opened stores and about online purchases (through an e-commerce web site managed by the same retailer with the same store brand) by people living in the new store service areas before and after the openings.

The paper shows that, for the single customer, the probability of purchasing online is reduced by the store opening in the short term, but tends to increase in the long term. Besides, results indicate that long-term synergy between the two channels depends mainly on indirect influence due to the mere presence of the store brand in the area rather than on the direct experience of shopping in the store.

The study highlights that channel portfolio enlargement from mono- to multi-channel retailing tends to activate a sort of life cycle; while in the early phase of store addition web sales tend to be cannibalized because the two channels are perceived as “substitutes” for each other, in the long run migration turns into a synergy effect; different channels tend to interact with and reinforce each other as customer touch points of the same retailer, in an omni-channel perspective.

The paper herein presents various original elements concerning types of available data (actual sales rather than consumers’ intentions/perceptions and individual level data rather than aggregate level ones), estimation technique used (binary choice model) and research hypotheses (distinguishing between “direct” and “indirect” synergy effects in multi-channel retailing).