Loyalty Register glossary
Basket analysis
The study of which products customers buy together in one transaction, used to shape offers, earning rules and store or site layout.
Basket analysis, also called market basket analysis, is the study of which products customers buy together in a single transaction. It answers questions such as “what else do people buy when they buy pasta?” and “which products bring shoppers into the store, and which ones only ride along?” Loyalty programs make it far more useful, because they tie each basket to a known member over time.
Updated · By Loyalty Register Editorial TeamRelated terms
How does basket analysis work?
It starts with item-level transaction data: one line per product, grouped by receipt. The analysis then looks for products that appear together more often than chance would predict. Three measures are common. Support is how often a combination appears across all baskets. Confidence is how often a basket with product A also contains product B. Lift compares that confidence with how often B appears anyway; a lift above 1 means the two products really are linked rather than both simply being popular.
How do loyalty programs use it?
The results feed offers and program rules. A grocer can pair a bonus on one product with a product that is often bought with it, or target members whose baskets usually include one item but rarely its natural partner. Programs also use it to see which categories bring members in and which ones they add once there, which helps set category bonuses and challenges. With member identity, the analysis can move from single baskets to patterns over weeks, such as the products a household buys every trip.
How can it inform earning rules?
Products in the same basket can carry very different margins. A hypothetical $50 basket might include $10 of low-margin staples and $40 of higher-margin items. A flat earn rate rewards both equally, while category rules or targeted bonuses can put more reward on the items that matter most to the business. Any change should be explained to members simply, since complex earning rules make the program harder to understand.
What should you watch?
Correlation is not cause: two products bought together do not mean that promoting one will sell the other. Test offers built on basket analysis against a control group before rolling them out. Make sure the data is clean, with returns, voids and staff purchases removed, and check that product categories are coded consistently across stores and channels. Finally, review how member data is used against the program’s privacy notice and the consent members gave.