What sells together in your stores — and which pairings are worth a display?
What sells together in your stores — and which pairings are worth a display? Upload order lines (order ID + item) and get pair associations with support, confidence and lift — plus the honesty about sample size that most basket analyses skip.
One row per order line: an order/bill ID column and an item column. Parsed locally in your browser — nothing is uploaded anywhere.
Support = share of baskets containing both items. Confidence = how often B appears when A does. Lift = support ÷ what independence predicts.
Lift above ~1.3 flags a real association worth a display test; below ~0.9 flags pairs customers deliberately don't buy together.
Under ~500 baskets, lift is noisy. Findings are display tests to run — measure attach-rate change over two weeks before rolling out.
Illustrative figures, not a client outcome.
Ideally 1,000+ baskets per store format. Smaller samples still surface hypotheses — just treat anything under 500 as a test to run, not a fact to deploy.
Analyse them separately. E-commerce baskets mix recommendations and delivery thresholds into behaviour; store baskets reflect navigation and adjacency.
Version 1.0 · Reviewed 2026-10 · Methodology is fixed and version-controlled; results depend on your inputs.
Your basket data holds next quarter's assortment plan. Mine it properly.
BLEZZINGS runs this analysis inside a focused engagement — with your data, your market and a plan to act on it.