Market Share and Product Distribution: Re-Tested and Extended

Abstract

In Marketing Science’s review of empirical generalizations in marketing, Reibstein and Farris (1995) provide a model for the relationship of market share (MS) and product distribution (D). The relationship is generally convex, with high share brands having more share points per point of distribution. This paper tests this relationship across many categories and across six years from IRI’s Infoscan Reviews database. Most, but not all, categories follow the general pattern. We present evidence on the generality of the model, show how many consumer packaged goods categories follow the general pattern, show how stable the model is when estimated for multiple cross-sections, and assess whether cross-sectional convex models are consistent with changes seen over time for brands with substantial changes in share and/or distribution. We investigate alternative models incorporating both breadth and depth of distribution and show these provide more insight and better prediction.