Average order value calculator — and what the mean hides
Last reviewed: 28 July 2026. Every benchmark carries its publisher, scope and read date.
Prefilled with the average Shopify store (Littledata, 2023). Replace with your own numbers.
Average order value
$85.00
total revenue ÷ orders
Below the all-ecommerce average
All ecommerce · Littledata 2023 · USDYou · $85.00
Littledata · 2,800 ecommerce sites, 2023 data · read 2026-07-26
Shopify-only average: US$85 (same source) — context, not a band.
- Revenue per session total revenue ÷ sessions
- $1.19
- Revenue per customer total revenue ÷ unique customers
- Add unique customers
- Orders per customer orders ÷ unique customers
- Add unique customers
If AOV rose and orders held
Incremental revenue
$3,500
over the same period as your inputs · 350 orders × $10.00
A projection, not a forecast. It assumes order volume holds while AOV rises — the thing you must test.
Average order value (AOV) = total revenue ÷ number of orders, over the same period. US$29,750 across 350 orders is US$85. Two warnings before you benchmark: AOV is a mean, and a few large orders drag it — read the median beside it. And a rising AOV is not rising revenue — bundles and thresholds push the average up while conversion can fall. Judge every lever on revenue per visitor — it catches both.
The calculator
Two fields give you the number. Add sessions and unique customers for revenue per session, revenue per customer and orders per customer, each labelled with its formula. A target-AOV slider prices the gap — arithmetic, not a forecast.
Why the mean misleads — twice
A few large orders drag it. Nine US$60 orders and one US$600 order average
US$114 — a figure that describes none of them. The median is US$60. Commerce revenue is
violently skewed — Kohavi's KDD 2014 paper measured revenue-per-customer skewness above 30 at
a commerce site — which is why revenue is the expensive experiment metric
(the skew maths). Read the median beside
the mean — our advice, not a published standard, one =MEDIAN() away on an order
export.
A higher mean is not more money. Revenue = AOV × orders. A bigger ask lifts the average and can cost you orders — the metric improves while revenue drops. Arithmetic, not a study. Revenue per visitor moves only when both multiply out in your favour — AOV is the diagnosis, never the verdict.
What a good average order value is
If you only want your own number, Shopify computes AOV in the admin and benchmarks it in-product against similar stores (changelog.shopify.com, April 2023). What it doesn't hand you is a sourced reference point with scope attached — there are exactly two.
All ecommerce — Littledata, 2023 benchmark of 2,800 sites (USD).
| Metric | Average | Best 20% | Best 10% |
|---|---|---|---|
| Average order value — all ecommerce | US$101 | >US$274 | >US$534 |
| Average order value — Shopify stores | US$85 | — | — |
| Revenue per customer — all ecommerce | US$111 (Shopify: US$92) | — | — |
(littledata.io, read 26 July 2026 — 2023 data, and the date travels with every figure.)
GB market — IRP Commerce, June 2026 (GBP). A live dataset: AOV £127.06, up 2.31% on June 2025's £124.19; average item price £65.94; revenue per session £1.89 (irpcommerce.com, read 26 July 2026 — it re-renders monthly). IRP labels these "Measured — first-party operational data recorded directly through the IRP platform." The same dataset puts GB conversion at 2.03%; conversion benchmarks are their own page. (Direction: Contentsquare's 2026 benchmark, 99 billion sessions, has AOV up 6% year on year — read 26 July 2026.)
The industry-AOV tables circulating in search results carry no source — we left them out. Two currencies, two markets, two vintages: never convert, never average. Prefer your own last quarter.
How to raise it — the honest levers
A free-shipping threshold just above your AOV. Orders under the bar get a reason to grow; every order already above it now ships free at your expense. Where to set it is its own calculation — the free-shipping threshold calculator does that maths. On a store built on Mtrix, the threshold is configurable in checkout, and two thresholds can be A/B tested against each other.
Bundles. Two products in one order instead of one — but a bigger ask can convert less often. Bundles and cart are a testable surface on stores built on Mtrix.
Test the lever, don't install it. On a store built on Mtrix, offers and bundles are A/B testable, and the report reads out in revenue per visitor, conversion rate and add-to-cart rate — an AOV win that costs a point of conversion shows up as the loss it is. The scope, plainly: on a store that stays on Shopify, Mtrix measures — replay, analytics, experiments, error tracking, performance — but does not vary prices or offers. That surface belongs to stores built on Mtrix.