LAUSR

In competitive markets, brands predictably share more customers with bigger competitors and fewer with smaller ones. Yet this robust duplication of purchase pattern occasionally has deviations. Where two or more brands share excess additional customers with each other (within) and fewer than expected with the rest of the market (between) – there is a partition. If brands have excess sharing (within) but sharing as expected with the rest of the market (between) – there is a grouping. Partitions generally occur because of functional difference from other brands. While partitions observations occur in cross-sectional data, typically of a year or less, no empirical evidence exists showing whether they persist over extended periods. Examining expected partitions in the United Kingdom for ten consumer goods categories, the authors show that two in five expected partitions are groupings, and an overwhelming majority persists over the medium term of three years. The findings contribute to the theoretical knowledge around market stability and Dirichlet assumption and allow marketers to confidently add or remove products knowing the partitions/groupings are likely stable.