Glossary
Product affinity
A measure of how strongly two or more products relate to each other in customer interest or purchase behavior.
Product affinity describes how strongly customers' interest in one product relates to their interest in another, whether that relationship shows up through co-purchase, co-browsing, similar ratings, or shared category membership. It is a general concept for "these products go together" that gets used to power recommendations, merchandising, and segmentation.
Affinity can be derived several ways: co-purchase data analyzed through market basket analysis, collaborative filtering based on what similar customers viewed or bought, or content-based similarity from product attributes like category, brand, or description. Each method produces a somewhat different notion of "related," which is why affinity scores from different systems don't always agree. Affinity differs from simple category grouping in that it's learned from behavior rather than assigned by a merchandising taxonomy, so it can surface non-obvious relationships a catalog structure would miss.
Product affinity matters because it underlies "customers also viewed" and "frequently bought together" features, personalized email recommendations, and assortment planning for what to stock or promote together. The main pitfall is letting affinity scores go stale as catalog and customer behavior change, or over-trusting affinity from a small number of transactions, which can produce recommendations that look data-driven but reflect noise rather than a real pattern.
Last reviewed September 22, 2026