What is Average Order Value (AOV)?

Average order value (AOV) is the average amount of one order in a chosen period. It is counted on orders, not on users: if one customer places three purchases, the denominator is three rows, not one.

Formula:

AOV = Revenue ÷ Number of orders

Revenue and orders must share the same window and the same rule: with or without tax, with or without shipping, before or after returns. Mixing rules breaks week-to-week and channel comparisons. In the e-commerce vertical this is a core merchandise metric next to ROAS and margin after logistics.

What sits in the numerator and denominator

The numerator is the sum of paid (or confirmed) orders. Contested inclusions: shipping, tips, tax, gift cards, partial refunds. The team locks one rule. If shipping is store revenue, it belongs in AOV; if it is a pass-through to the carrier, it does not.

The denominator is the order as a document — not a cart line and not a session. Cancelled COD and full refunds usually leave both sides, or the AOV looks healthy while cash and warehouse do not. A partial return of one SKU cuts revenue while the order may stay in the denominator — that choice must be written down.

AOV is not revenue per user: that is closer to ARPU. It is not LTV: LTV sums all orders and, in subscriptions, lifetime periods, not one transaction.

Link to LTV, CAC, and margin

A rough link for repeat purchase:

LTV ≈ AOV × orders per customer over the lifetime × margin

(or AOV × frequency × lifespan, if that is how the model is built). A higher AOV lifts LTV only if margin and repeat do not fall: a “−30% above cart N” offer raises the receipt in currency and cuts contribution.

Compare CAC with contribution per order (or accumulated margin to payback), not with raw AOV. An 80 € order that loses 40% to logistics and returns does not “cover” a 35 € CAC. In performance marketing the pair is: cost per conversion on one side, AOV and margin as conversion value on the other.

On marketplaces, subtract take-rate before comparing to ads. Otherwise in-account ROAS and P&L unit economics diverge.

Where the metric lives in e-commerce

Online store, D2C, marketplace, social commerce (TikTok Shop, Facebook catalog): AOV moves revenue at the same order count. In COD geos, “ad-account AOV” and AOV after delivery are different numbers: part of the carts are never collected. A buyer who optimizes only form CR can flood cheap small orders and kill margin.

In affiliate, CPS pays a percent of order value. Affiliate revenue ≈ AOV × CR × commission rate (minus cancellations). A low AOV with high CR is not automatically better than a high AOV with mid CR. A tracker sees conversion value only if the advertiser or network passes the amount in a postback or pixel.

Mixing GEOs and currencies without normalization creates a fake rise: add a market with a heavier basket — AOV is up, unit economics may not be.

What raises and lowers AOV

Raises: bundles, checkout upsell, free-shipping threshold, cart cross-sell, mix toward more expensive SKUs, order minimums. Lowers: deep discounts, “first order” coupons, traffic to the cheapest item in the feed, clearance.

Separate mix effect from cart behavior. A new creative that lands on a $15 SKU instead of $40 drops AOV with no change to checkout. The same SKU with the shipping threshold removed drops basket composition.

Optimizing ads “to ROAS” without watching AOV hides the failure mode: the algorithm harvests cheap conversions, the account is green, the warehouse is small. The reverse: bundles push AOV up, CR and COD approval rate fall — revenue per click does not rise.

Common calculation mistakes

Dividing revenue by customers, not orders — that is revenue per buyer in the period, not AOV.
Mixing placed and delivered orders in COD.
Comparing weeks with a different share of B2B invoices or wholesale.
Ignoring returns in one channel and counting them in another.
Confusing AOV with ARPU (all users in the denominator) and with LTV (whole life, not one receipt).

AOV is a snapshot of the transaction. Scale decisions still close on LTV, CAC, and — when the fight is “these orders would have happened without ads” — incrementality.