Market basket analysis for a Fortune 100 sports merchandise client

TransOrg conducted an exploratory data analysis at multiple levels:

• Popularity of the offered discounts

We analysed the sales in different discount categories and assessed ways to prevent cannibalization. We also analyzed the “discount elasticity” of a product in different price buckets by offering differential discount on the same product. TransOrg merged inventory and sales data for better discount management.

• Sales of product variants

Product variant analysis was done to understand the demand of different variant across products which helps in better positioning of less selling variants and offer differential discounting on different variant of the same products.

• Market basket analysis

TransOrg used the Apriori association method to find frequently bought product item sets and derived association rules to uncover meaningful correlations between different products according to their co-occurrence in a data set. The following measures are used to evaluate the strength of the association:

• Support for the rule indicates its outcome in terms of overall size
• Confidence determines the operational usefulness of a rule
• Lift ratio indicates how efficient is the rule in finding consequences, compared to random selection of transaction

Table 1: Association rule. For Illustration only

The rule suggests that if customers buy short sleeve top, then there is a higher chance of cross selling shoes as well.

TransOrg also performed separate market basket analysis for male vs. female customers and across different factory outlets stores based on location.