Market Share and Distribution: A Generalization, a Speculation, and Some Implications

Thumbnail Image
Penn collection
Marketing Papers
Degree type
market share
Business Administration, Management, and Operations
Business Analytics
Business and Corporate Communications
Business Intelligence
Management Sciences and Quantitative Methods
Organizational Behavior and Theory
Grant number
Copyright date
Related resources
Reibstein, David J
Farris, Paul W

In this paper we review evidence of a generalized convex cross-sectional relationship between retail distribution and unit market share, i.e., large-share brands have more share points per percentage of distribution than small-share brands. The dynamics and structure of distribution and share can help explain many phenomena in marketing, including this convex shape: (1) market share is both a cause and an effect of distribution, and (2) in the typical convenience goods distribution system there are a few large outlets that stock many brands and numerous smaller outlets that stock the leading brands only. Generally, the observed cross-sectional “curve” relating distribution and share will reflect the retailers' stocking decisions, not the incremental effect of distribution on share. However, a logically consistent model of share based on (1) and (2), when combined with the assumption of low search loyalty, results in customers being willing to switch from preferred to available brands. A further consequence is that the marginal effect of weighted distribution on share is likely to be increasing, i.e., result in convex curves relating distribution and share for a given brand. In some cases, and for some measures of distribution, these convex curves have been observed in time-series data for brands that failed and lost distribution over a relatively short period of time. The implication is that marketers should monitor distribution carefully, as it is the result of combined effects of brand preference, loyalty, and “push” programs. With a better understanding of the market share/distribution relationship, managers should be in a better position to forecast marketplace results for a given level of distribution.

Date Range for Data Collection (Start Date)
Date Range for Data Collection (End Date)
Digital Object Identifier
Series name and number
Publication date
Volume number
Issue number
Publisher DOI
Journal Issue
Recommended citation