RPV

Definition

RPV is revenue divided by the number of visitors (or sessions — say which). In e-commerce, RPV ≈ AOV × CVR if the three share a unit. It is a continuous metric, often full of zeros (non-buyers), hence high variance. That is why CUPED and continuous sample size exist.

A test that improves CVR while breaking AOV can leave RPV flat. Read all three together, or choose RPV as primary if that is the economy you want.

When it matters

When the lever can move both rate and basket (pricing, merchandising, checkout). When you compare arms whose product mix differs. At sizing: RPV mean and σ, not a z on a proportion.

Less relevant on B2B lead gen with no reliable revenue per visitor — there, conversion (and its expected value) is enough.

Pitfalls

Using the binary formula on RPV. Winsorising large baskets without saying so (you changed the metric). Reporting RPV “among buyers” — that is AOV. Forgetting currency and tax across countries in an international test.

Related terms

• AOV
• CVR
• CUPED

Related tool

Mean, standard deviation and MDE for revenue per visitor: Open sample size (continuous).