Market Basket Analysis

A data mining technique that discovers products frequently purchased together.

Market Basket Analysis (MBA) is a data mining technique rooted in association rule learning that identifies products frequently purchased together. Originally developed for brick-and-mortar retail (analyzing what's in a shopper's "basket"), it's now a cornerstone of modern e-commerce analytics.

How it works

MBA analyzes your transactional data to find statistically significant product associations. For example, if 68% of customers who buy a "Classic Cap" also buy a "Logo T-Shirt," that's an association rule worth acting on.

Key metrics in MBA

  • Support: How frequently the combination appears in all transactions
  • Confidence: Given product A is purchased, the probability product B is also purchased
  • Lift: How much more likely products are purchased together compared to independently

E-commerce applications

  1. Product bundling: Create bundles from high-confidence associations
  2. Cross-sell recommendations: "Frequently bought together" sections
  3. Store layout optimization: Position associated products near each other
  4. Inventory planning: Stock associated products together to prevent missed sales

The Apriori algorithm is the most common approach to MBA, though newer methods like FP-Growth offer better performance on large datasets. The quality of your MBA depends entirely on the quality and volume of your transactional data — typically you need at least several hundred orders to find statistically meaningful patterns.

See Market Basket Analysis working on your data.

Affinsy turns these analytics concepts into actionable insights for your store, no data-science degree required.