Reports & Store Operations
Product Affinity
Find Products commonly bought together using support, confidence, lift, and attachment measures without over-reading small samples.
Overview
- Product Affinity analyses eligible baskets and Product pairs using order support, directional confidence, lift, and attach rate.
- Summary includes Orders, Items per order, Single-item rate, and Modifier attach rate.
- A minimum number of report Orders and pair-support threshold is required before pairs are shown.
- Gift cards, prepaid vouchers, and unmapped items are excluded from eligible pair analysis.
- The full eligible basket is used once per order rather than prorating Product quantities across split payments.
When to use this
- Use affinity for bundle/upsell ideas, menu placement, and staff prompts—then validate with real operational judgement.
Step-by-step
- Choose a sufficiently long period.
- Start with pairs that have meaningful Order support.
- Compare directional confidence A→B and B→A.
- Use Lift to distinguish genuine association from two independently popular Products.
- Test an upsell/bundle before making a permanent menu change.
Common mistakes
- Do not rank a pair from tiny support above a well-supported pattern.
- Do not interpret correlation as proof one Product causes purchase of another.
- Do not include Gift Card/prepaid voucher behaviour in food attachment conclusions.
Troubleshooting
- If no pairs appear, increase the period or reduce the supported minimum pair threshold within the UI's intended range.
- If expected Products are absent, verify they are eligible mapped Product lines.