Definition
Basket affinity is a retail analytics concept that measures the likelihood of two or more products being purchased together in the same shopping basket. It highlights statistically significant co-purchase relationships across transaction data to reveal which product combinations naturally bundle in customer behavior.
Why it matters
Basket affinity turns raw transaction data into actionable commercial intelligence. By understanding which products customers consistently buy together, retailers and brands can make smarter decisions across pricing, promotions, store layout, and assortment.
Common applications of basket affinity include:
- Designing multi-buy promotions that don’t feel forced on shoppers.
- Optimizing product placement and category proximity.
- Building personalized cross-sell recommendations.
- Identifying halo effects between categories.
- Strengthening supplier negotiations with co-purchase evidence.
- Reducing cannibalization by understanding substitution vs. complement dynamics.
How it works
Basket affinity analysis scans historical transactional records to detect patterns in which SKUs, categories, or brands appear together more frequently than predicted.
Key characteristics include:
- Analysis is performed at the transaction or basket level.
- Results are directional: product A may lift product B without the reverse being equally true.
- Affinity can be measured at different depths: SKU, subcategory, category, or brand.
- Relationships are dynamic and should be updated periodically.
- Strong affinity signals can indicate complementary relationships, habitual pairing, or occasion-based bundling.
The calculation
Basket affinity is quantified using three core metrics: Support, confidence, and lift.
Support measures how many times a product pair appears across all baskets:
Support (A → B) = Baskets containing both A and B ÷ Total baskets
Confidence measures how often B is purchased when A is already in the basket:
Confidence (A → B) = Baskets containing both A and B ÷ Baskets containing A
Lift is the critical metric, it measures whether the pairing is stronger than chance:
Lift (A → B) = Confidence (A → B) ÷ Support (B)
Interpreting lift:
- Lift > 1: Products are bought together more often than expected, a positive affinity signal
- Lift = 1: No meaningful relationship , co-purchase is random
- Lift < 1: Products are bought together less often than expected, potential substitution effect
Practical example
A grocery retailer runs basket affinity analysis across 12 months of transaction data. The analysis shows a strong relationship between bagged charcoal and aluminum foil; the pair appears in the same basket far more often than the purchase rates of either product would predict individually, with a lift score of 3.4.
Based on this insight, the category team co-locates both products in the seasonal grilling aisle and builds a bundled promotion: buy charcoal and receive a discount on foil.
The result is a higher average basket value, improved foil sell-through during the grilling season, and a shopper experience that feels intuitive rather than manufactured, because it reflects how customers already naturally shop.
