AOV = UPT × Average Selling Price — UPT splits AOV into its unit-count and unit-price factors.
UPT is computed over a defined window. It is an order-level basket-count metric, not a customer-level one — the denominator is orders, so the number is unaffected by how those orders are distributed across buyers. Most retail and POS tools report the same ratio under different names: Items Per Order (IPO), average units per order, or “basket size.” A dashboard labeled IPO and one labeled UPT are computing the same thing.
The operator value of UPT is that it splits AOV into its two factors. By construction, AOV = UPT × average unit price, so any change in AOV is mechanically attributable to a change in unit count, a change in unit price, or both. A brand can grow AOV without UPT moving at all by raising prices or shifting mix toward premium SKUs; conversely, a brand can grow UPT via a bundle or BOGO without AOV moving proportionally if the added units are lower-priced or free. Reading UPT alongside average selling price tells the operator which lever actually moved.
That decomposition is why UPT is the metric merchandisers and retention marketers track behind cart-builder mechanics: free-shipping thresholds, “add one more to unlock a gift,” BOGO promotions, cross-sell modules, and bundle pricing. The goal of those mechanics is unit lift, not price lift, so UPT is the honest read for whether they worked — AOV will move too, but it conflates the unit signal with any incidental price-mix change.
UPT is distinct from attach rate: UPT counts units regardless of basket composition, while attach rate names a specific pair (A given B). A 2.1 UPT describes basket size; a 38% accessory-to-anchor attach rate describes a pairing. Both are unit-lift metrics, but they answer different questions — how full is the cart, and which pairings are pulling.