Definition
Average order value is total revenue divided by the number of orders over the same period. "Revenue" needs a convention: gross merchandise value before discounts, or net of discounts, and with or without shipping charges and tax. Most commerce platforms and analytics tools report it net of discounts and excluding tax, but not all, and refunds are usually not deducted. The order is the unit, not the customer, so one person placing three orders contributes three times to the denominator. Customer-level averages (revenue per customer) are a different metric with a different use.
Because AOV is a mean, a small number of very large orders pulls it up. A business selling mostly small items with occasional bulk orders can have an AOV nobody actually pays. The median order and the distribution by price band say more about typical behaviour; the mean is still the figure that reconciles to revenue.
Revenue ÷ Orders(currency)
How to interpret it
Direction is "higher is better", holding order volume constant. That condition does the work. Revenue is orders multiplied by AOV, and most things that raise AOV also lower the number of orders: a higher free-delivery threshold, fewer discounts, bundling instead of single items. The right reading is revenue per visitor, which is CVRShare of visits or clicks that convert. multiplied by AOV, and the useful question is whether a change raised that product, not whether it raised AOV alone.
AOV is the reason the same acquisition cost can be profitable for one retailer and ruinous for another. Return on ad spend is revenue divided by spend, and revenue is AOV times orders; at a fixed cost per order, a higher AOV is a higher ROASRevenue generated per unit of ad spend. with no change in bidding, which the ROAS page spells out. Lifetime value takes AOV as its first factor, so a higher basket feeds through to how much the business can afford to pay for a customer. The conversion section of the ecommerce hub covers where AOV sits among the on-site metrics.
What moves it
- Range and price architecture. What is sold, and at what price points, sets the floor and ceiling.
- Bundles, multi-buy offers and delivery thresholds, which nudge the basket up; the AOV levers guide covers these in detail and this page does not repeat them.
- Cross-sell and upsell placement in product pages and checkout.
- Discounting. Sitewide discounts lower AOV directly even when they raise order volume.
- Customer mix. Returning customers and business buyers typically place larger orders than first-time consumers.
- Channel and campaign. Traffic sent to a sale landing page produces smaller orders than traffic sent to a premium range.
- Seasonality and gifting, which change basket composition more than they change conversion rate.
Most of these are deliberate choices; the last two are reasons an AOV can move with nobody on the team having done anything.
Reading it in context
| Pattern | Likely reading | Check |
|---|---|---|
| AOV up, orders down, revenue flat | A threshold or pricing change shifted the mix | Revenue per visitor |
| AOV up, ROAS up, spend flat | Basket growth doing the work of efficiency | Which products drove it |
| AOV down, CVR up | A promotion widened the audience with smaller baskets | Margin per order |
| AOV volatile month to month | Few large orders dominate the mean | Median and distribution |
| AOV differs sharply by channel | Campaigns land on different parts of the range | Landing page mix |
Raising the basket instead of the bid
A homeware store spends £10,000 a month on paid media and gets 500 orders at an AOV of £60, so revenue is £30,000 and ROAS is 3.0. The team introduces a free-delivery threshold at £75 and a "complete the set" prompt in the basket. Orders fall slightly to 480 as some shoppers balk, but AOV rises to £74, revenue to £35,520 and ROAS to 3.55, on the same spend and the same CPC. Revenue per visitor rises from £1.50 to £1.78 across 20,000 visits. The median order, previously £42, moves to £51, so the gain is broad rather than driven by a handful of large baskets.
Related metrics
CVRShare of visits or clicks that convert. is the metric AOV trades against, and the two are only read together as revenue per visitor; a tactic that raises one and lowers the other has to be judged on the product. ROASRevenue generated per unit of ad spend. is where a change in AOV shows up first in paid media reports, since revenue moves with the basket while spend does not, and a retailer that cannot lower its click costs can still raise its return this way. LTVTotal value a customer brings over time. carries AOV forward across the customer's lifetime, so a merchandising change that raises the basket compounds across repeat purchases. Units per order and average selling price are the two factors AOV decomposes into, and they tell you whether a bigger basket came from more items or dearer ones. Read AOV with all three beside it, never as a headline on its own.
Common mistakes
- Celebrating AOV without checking orders, when a threshold change has cost more revenue than it gained.
- Using the mean to describe the typical customer in a long-tailed distribution where most baskets are far smaller.
- Comparing AOV across tools that treat discounts, shipping and tax differently, then chasing a gap that is only a definition.
- Attributing an AOV change to the site when the channel mix or a sale changed who was buying.
- Treating AOV as a merchandising metric only, when it is the lever that moves paid-media efficiency with no bidding change.
- Pushing the basket up with discounts, which raises revenue per order while lowering margin per order.
- Reporting AOV without the order count beside it, which hides whether a change in the average came from the basket or from who was buying.
