Reducing Assortment: An Attribute-Based Approach

TLDR

1. Background: "Supermarket categories are cluttered with many similar SKUs, yet prior work shows that cutting up to 54% of low‑selling SKUs can leave perceived variety and sales largely unchanged." 2. Purpose: "The study aims to examine how different types of SKU reductions, defined by changes to available attributes such as brand count, affect purchase behavior." 3. Mechanism: "The authors analyze data from a natural experiment by an online grocer where 94% of categories saw dramatic cuts in SKUs, especially low‑selling ones, and they assess how attribute‑based reductions influence purchase behavior." 4.

Abstract

Most supermarket categories are cluttered with items, or stockkeeping units (SKUs), that differ very little at the attribute level. Previous research has found that reductions (up to 54%) in the number of low-selling SKUs need not affect perceptions of variety and therefore sales, significantly. In this research, the authors analyze data from a natural experiment conducted by an online grocer, in which 94% of the categories experienced dramatic cuts in the number of SKUs offered, particularly low-selling SKUs. Sales were indeed affected dramatically, increasing an average of 11% across the 42 categories examined. Sales rose in more than two-thirds of these categories, nearly half of which experienced an increase of 10% or more; 75% of households increased their overall expenditures after the cut in SKUs. In turn, the authors examine how different types of SKU reductions—defined by how the cuts affect the available attributes or features of a category (e.g., the number of brands)—affected purchase behavior differently. The results indicate that consumers experienced divergent reactions to the reduction in sizes, but they uniformly welcomed the elimination of clutter brought on by the reduction in redundant items. In addition, of households that were loyal to a single brand, size, or brand–size combination that was eliminated, nearly half continued purchasing within the category. Also, contrary to previous research on SKU reductions, the authors find that category sales depend on the total number of SKUs offered. The authors extend the previous research by showing that (1) category sales depend on the availability of key product and category attributes and (2) two particularly important attributes to consumers in an assortment are brand and flavor.