AOV calculator

Revenue is what customers paid over the period, in one currency. Decide once whether returns and cancellations are netted off and hold to it. The gross figure describes what people chose to buy; the net figure describes what the business kept, and only the second one should ever be used to set an acquisition ceiling.

Orders is the count of transactions in the same period, on the same definition. A subscription renewal, an exchange and a split shipment are each either an order or not, and the answer only has to be consistent.

Items sold is optional and worth supplying. It turns one number into three, and the two extra figures are the ones that say whether the order value is driven by price or by basket size — which decides what can be done about it.

An average order value on its own does not distinguish between two very different businesses. One sells a single expensive item to each customer. The other sells several inexpensive ones. They can report the same average and they have almost nothing in common.

Items per order says which one you are. At 1.6 items an order, most customers are buying one thing and a minority are adding a second, so the levers are bundles, cross-sells, thresholds and anything that makes a second item obvious. At four items an order the basket is already being built, and the lever is more likely to be the price and the mix of what goes into it.

Revenue per item is the other half of the same reading. An average order value that rose while revenue per item fell means people are buying more, cheaper things — good for volume, usually worse for margin. One that rose while items per order held means prices or the mix moved. The two together turn a single figure into a diagnosis.

Average order value multiplied by the gross margin is the gross profit one conversion leaves. That figure is the ceiling on what a conversion may cost, and dividing it by the rate at which clicks convert gives the ceiling on what a click may cost. Every acquisition ceiling on this site ultimately traces back here.

Which is why the number deserves as much attention as the auction, and rarely gets it. A ten per cent lift in order value raises the affordable cost per click by ten per cent. Finding the same ten per cent in the auction means outbidding everybody who wants the same person — a harder, slower and more expensive fight than adding a free-shipping threshold just above the current average.

It is also the figure that explains most sudden changes in reported return on ad spend. A campaign whose return fell while its conversion count rose has usually shifted the mix toward cheaper products. That is a merchandising event, and it arrives on the media team's desk labelled as a performance problem.

The tail. A few very large orders drag the mean above nearly every real order. If the distribution is skewed, the median is the number that describes a typical customer, and an acquisition ceiling set from an inflated mean is a ceiling most orders cannot pay for.

Mixed periods. A sale week and a full-price week have different order values, and an average across both describes neither. The same is true of a period that spans a product launch.

Mixed channels. Customers acquired from search, from a feed and from a newsletter buy differently. A blended average hides a channel that is bringing in cheap orders and one that is bringing in expensive ones, and the media decisions for the two are opposite.

New against returning. Returning customers usually spend more. An average that includes them overstates what a newly acquired customer is worth, which is precisely the figure an acquisition ceiling needs.

Currency. Two markets averaged without a stated exchange rate produce a figure nobody can check, and this page converts nothing.

One period. A single month's average is a sample, and a small business's month is a small sample. A figure that moved five per cent between two months has usually not moved at all; a figure that has trended in one direction for a quarter has. Reading a monthly wobble as a merchandising result is how perfectly good product pages get rewritten for nothing.

Thresholds, most reliably: free shipping or a gift just above the current average moves a measurable share of orders up to it. Bundles and multi-buys, which raise items per order directly. Better-chosen products in the ads, which changes what people arrive intending to buy. Post-purchase offers, which add to an order after the decision to buy has already been made. And price, which moves the number immediately and has to be argued on its own terms rather than as a media tactic.

Each of those is worth measuring against gross profit rather than against this number alone. An order value lifted by discounting a high-margin product into a bundle can raise the average and lower the profit, which is the one outcome this page cannot see on its own.