How AOV is calculated
Average order value takes the revenue recorded in a period and divides it by the number of orders in the same period. Nothing more complicated is involved, which is why it appears in almost every ecommerce report and why it is so easy to calculate two different ways by accident.
The two inputs both need defining. If the revenue figure includes delivery and tax, the average is higher than one built on product value alone. If refunded and cancelled orders stay in the count, the average is flattered further. Neither approach is wrong, but comparing an average built one way against one built the other will send you chasing a change that never happened.
Why AOV matters
It sets the ceiling on what you can afford to pay for a customer. Advertising costs are charged per click and per impression, not per rupee of basket, so a shop with a high average order value can outbid a shop with a low one for exactly the same traffic. When a competitor seems to occupy every auction position you want, the reason is often basket size rather than budget.
It is also the least painful growth lever available. Winning more visitors costs money and takes time. Persuading the visitors already in the checkout to add one more item costs almost nothing, and every extra rupee of order value flows into return on ad spend immediately.
Where AOV misleads
An average hides its own distribution. A shop selling a cheap accessory alongside an expensive main product does not have a typical order anywhere near the average — it has two clusters and a number sitting in the empty space between them. A handful of unusually large orders can pull the figure up for a month and drop it again the next, with nothing having changed in the business.
Segmentation matters too. New and returning customers, paid and organic traffic, mobile and desktop buyers usually shop differently, so a single site-wide average can rise while the segment you actually care about falls. And a rising average is not automatically good news: it can mean the cheaper products stopped selling, which is a warning rather than a win.
How to act on it
Split it before you act on it. Look at the average by traffic source, by device, by customer type and by product category, and you will usually find that the site-wide number was concealing one segment doing something interesting.
To move it, work at the point of decision rather than in the advertising. A free delivery threshold set a little above the current typical basket, a genuinely relevant related product on the basket page, a bundle that saves the customer a real amount, or a larger pack size all nudge order value upwards without buying another visitor. Test one change at a time and watch conversion rate alongside it — an average pushed up by an offer that scares people out of the checkout has cost you more than it earned. If you would rather work on the checkout than the media, conversion rate optimisation is where that work belongs.