Average order value: formula, calculation
and how to increase it

Average order value is one of the numbers every e-commerce team checks each week, alongside traffic and conversion rate. It takes a single division to calculate, but how you measure it changes a lot about what it tells you: with or without tax, with or without shipping, before or after returns.

This guide covers the formula, a worked example and the most common interpretation mistakes. It then goes through the ways to increase it, including one that rarely makes the list: internal search, which decides what customers find on your site, and therefore what they can add to their order.

What is average order value?

Average order value (AOV) is the average amount spent per order over a given period. It says nothing about the number of customers or how often they buy: it only measures the size of a typical order.

Together with traffic and conversion rate, it forms the basic revenue equation of any online store:

The three factors multiply. A 10% gain in AOV with stable traffic and conversion means 10% more revenue, with no extra acquisition spend. That is why it is watched so closely.

The average order value formula

Decisions to make before you calculate

The math is simple, the conventions less so. Set them once and keep them in every report, or your month-to-month comparisons will mean nothing.

• With or without tax: the tax-inclusive figure is what the customer pays, the tax-exclusive one compares directly with your margin. To manage profitability, use the tax-exclusive figure.
• Shipping: exclude it to measure what customers buy in products. A free-shipping threshold moves the "products" basket and the "paid" basket differently, which is why they should not be mixed.
• Returns and cancellations: calculated on placed orders, AOV looks flattering in categories with frequent returns, such as fashion. Calculate it on net orders as well.
• Period: a sales month or Black Friday cannot be compared with a quiet month. Compare a period with the same period last year instead.

Mean or median?

A handful of large orders (a business customer, a full kit bought in one go) is enough to pull the average up. If your AOV moves for no obvious reason, look at the median. If it stays flat, a few unusual orders made the difference, not the behavior of your customers as a whole.

How to calculate your AOV in practice

Shopify, PrestaShop, WooCommerce and Magento all show an average order value in their sales reports. Just check what it includes: taxes, shipping, cancelled or refunded orders. Definitions vary between tools, which often explains the gap between the back-office figure and the one in Google Analytics.

In GA4, purchase revenue divided by the number of transactions gives the same metric, with one advantage: you can segment it by traffic source, device and on-site behavior. That is what lets you compare the AOV of sessions with and without search. Our guide to analyzing site search queries in GA4 covers the setup.

AOV, revenue per visitor, units per order

Average order value is often confused with related metrics. Each answers a different question:

• Average order value = revenue ÷ orders. It measures the size of an order.
• Revenue per visitor = revenue ÷ visitors. It combines AOV and conversion rate, and sums up a site's commercial performance better than any other single number.
• Units per order = items sold ÷ orders. It shows whether baskets grow through quantity or through unit price.
• Customer lifetime value: what a customer brings in over their whole relationship with the store. An AOV inflated by discounts can push it down.

Tracked on its own, AOV can mislead. A rise can come from lower conversion on small baskets: fewer orders overall, but a higher average. Always read it next to revenue per visitor.

The classic ways to increase average order value

The usual levers are well known and they work, as long as you watch the margin per order rather than just the amount collected.

• Free-shipping threshold. Set a little above your current AOV, it nudges customers to add an item to clear the bar. Too high and it discourages; too low and it gives free shipping to orders that would have paid for it.
• Cross-selling. Offering what goes with the product, on the product page and in the cart: batteries for a toy, a sleeve for a laptop, a care product for leather shoes.
• Upselling. Showing the higher-end version at the moment of choice, with a clear explanation of what it adds.
• Bundles. A coherent bundle at a slightly lower price grows the order without discounting each product on its own.
• Tiered discounts ("10% off over $80"). Effective on the amount, but every dollar of basket gained costs part of the margin.

Internal search, the forgotten AOV lever

All these levers assume customers find the products you want to show them. On most stores, the search bar is the main way in: according to the Baymard Institute (2026), nearly one user in two prefers search to browsing the menus. A search engine that misunderstands queries does more than lose sales, it also shrinks the ones it keeps.

Every zero-result search takes an item out of the basket

A customer has found a tent and now searches for "tent pegs". The engine returns nothing, because your product pages say "stakes". The customer does not necessarily leave: they check out a smaller order and buy the pegs elsewhere. This loss shows up in no conversion report, since the order did happen.

Our article on zero-result pages explains how to spot and reduce them. Search synonyms deal precisely with the gap between the customer's vocabulary and the catalog's.

Autocomplete helps customers discover products

An autocomplete that shows products and categories from the first letters puts in front of customers items they would not have thought to search for. Typing "coffee" and seeing grinders and filters next to the coffee beans is a cross-sell made at the moment intent is strongest. Our guide to search autocomplete covers the settings that matter.

Broad queries call for several products

"Hiking outfit", "gift for a barista", "back-to-school supplies": these queries describe a need, not a product, and the right answer is a selection of several items. A keyword engine only finds product pages containing those exact words. Semantic search understands the need and brings up shoes, backpack and water bottle together, enough to build a full basket from a single search.

Sorting and filters steer customers through the range

A results page sorted by ascending price by default pushes customers toward entry-level products. Relevance ranking, fed by clicks and sales, highlights the products customers actually buy. Well-designed facets and filters (material, use, level) also help customers justify a more expensive product, because they can see what sets it apart.

Searchandising puts the right products first

On your most frequent queries, you can decide what shows up first: a bundle rather than a single product, the premium version, the must-have accessory. That is the idea behind searchandising: applying to the results page the merchandising rules you already apply to your home page.

The assumption is deliberately modest. It leaves out whole orders saved by a search that would otherwise have failed, an effect we quantify in our article on the cost of poor internal search.

Measure the effect of search on your AOV

In GA4, once site search tracking is enabled, create two segments: sessions with at least one search, and sessions without. Compare for each:

• average order value;
• units per order;
• revenue per visitor;
• conversion rate, whose gap is covered in our article on how search users convert.

Within the search segment, then isolate the sessions that hit at least one zero-result page. If their AOV is clearly lower than that of other search sessions, you have found your first project. These numbers are specific to your store: no market average will tell you what search is worth for you.

Where to start

1. Recalculate your AOV excluding tax and shipping, on net orders, and note the median next to it.
2. Compare it between sessions with and without search in GA4.
3. Export last month's zero-result queries and fix first the ones about products you sell.
4. Test your autocomplete on your ten main categories: does it suggest products, or only words?
5. Check your free-shipping threshold against the AOV measured in step 1.