How to use AOV Measurement
A working framework for measuring Average Order Value beyond the sitewide number — the formula, the segmentation cuts that actually drive decisions, and how to benchmark against your category.
AOV Measurement
AOV measurement is the practice of calculating, segmenting, and benchmarking Average Order Value so it becomes a decision-useful diagnostic instead of a vanity number.
Average Order Value is total revenue divided by number of orders over a period. That's the formula. Measurement is the harder part: deciding which revenue counts (gross vs net of discounts, with or without shipping and tax), which orders count (test orders, B2B, subscription renewals), and over which window.
Done well, AOV measurement breaks the sitewide number into segments — channel, device, new vs returning, cohort, SKU mix — so a 4% dip in blended AOV resolves into a specific cause: a discount-heavy paid campaign, a mobile checkout regression, or a shift in product mix toward lower-priced SKUs.
Most stores look at one AOV number in their Shopify dashboard and call it a metric. It isn't — it's an average of averages that hides everything you'd want to act on. A €72 sitewide AOV can mask €110 from returning customers on desktop and €48 from cold paid traffic on mobile, and those two cohorts need opposite interventions.
This guide walks through the four steps that make AOV measurement useful: pin down the formula and its edge cases, choose the segmentation cuts that change decisions, benchmark against the right reference set, and connect the number back to a CRO or merchandising action.
Step 1: Get the formula and its edge cases right
AOV = Total revenue / Number of orders. The arithmetic is trivial; the definitions aren't. Before you trust any AOV number, lock down four choices: gross vs net revenue, whether shipping and tax are included, how refunds and partial cancellations flow through, and which orders are excluded (staff test orders, B2B wholesale, gift cards purchased vs redeemed).
The most common reporting mismatch is gross-of-discount vs net-of-discount. Shopify's default AOV in the admin uses gross merchandise value after discounts but before refunds and including shipping. GA4's purchase event, by default, uses the `value` parameter your dev team configured — which is often pre-tax, post-discount, but inconsistent across themes. If you don't know which one your dashboard shows, your AOV trend is partly noise.
Pick one canonical definition for the business — usually net revenue after discounts, excluding shipping and tax, excluding refunds — document it, and make every dashboard match. Keep a second gross-AOV view if finance needs it, but the CRO and merchandising team should work from one number.
The subscription trap
If you sell subscriptions alongside one-off purchases, blended AOV is almost meaningless. A €25 monthly refill order and a €120 first-time bundle are different products with different economics. Split AOV into one-time vs subscription-initial vs subscription-recurring before you do anything else.
Step 2: Pick the segmentation cuts that change decisions
A sitewide AOV trending sideways can hide two segments moving in opposite directions. The point of segmentation isn't to slice the number a hundred ways — it's to find the cuts where AOV differs enough to drive different actions. For most online retailers, five cuts cover 90% of the diagnostic value, and AOV Segmentation goes deeper on each.
Channel (paid social vs paid search vs organic vs email vs direct) separates audiences with very different intent and basket composition. Device (mobile vs desktop) usually shows a 15-30% AOV gap that reflects both UX friction and browsing context. New vs returning splits customer LTV economics from acquisition economics. Cohort (acquisition month) shows whether AOV is structurally drifting. SKU mix or category isolates merchandising effects from behavioural ones.
Typical AOV variance across segments for a mid-market apparel store
Notice what the blended number hides. A 2.5x spread between the lowest and highest segments means an AOV optimisation aimed at the wrong cohort can move blended AOV by half a euro and look like a failed test. Run AOV-uplift experiments against the segment whose AOV you're trying to move, not against the sitewide average.
Step 3: Benchmark against the right reference set
AOV is one of the most category-dependent metrics in e-commerce. A €40 AOV is excellent for a beauty single-SKU brand and catastrophic for an electronics retailer. Before comparing your number to a published benchmark, check that the reference set matches your category, price point, and channel mix. AOV Benchmarks by Industry breaks this out in detail.
The most useful benchmark is almost always your own historical data — same segment, same season, year over year. External benchmarks are useful for sanity-checking, not target-setting. If your apparel store's mobile-new AOV is €48 and the category benchmark is €55, that's a hypothesis worth testing, not a verdict.
Indicative AOV ranges by category and platform (median range, online retail)
| Category | Shopify median | WooCommerce median | Typical mobile gap |
|---|---|---|---|
| Apparel & accessories | €65–95 | €55–85 | -22% |
| Beauty & personal care | €40–70 | €35–60 | -15% |
| Home & furniture | €120–250 | €110–230 | -28% |
| Electronics & accessories | €85–180 | €80–170 | -18% |
| Food & beverage (DTC) | €35–60 | €30–55 | -10% |
| Health & supplements | €55–95 | €50–90 | -12% |
The mobile gap column is worth staring at. Across categories, mobile AOV runs 10-30% below desktop — partly because mobile sessions skew toward browse-and-bounce traffic, partly because checkout friction kills basket-builders like cross-sells and quantity bumps. If your mobile gap is materially worse than the category norm, the checkout is usually where to look first.
Step 4: Connect AOV to a decision
AOV is a means, not an end. The decisions it should drive sit in three buckets: merchandising (bundling, pricing tiers, free-shipping thresholds), CRO (cart upsells, post-purchase offers, quantity selectors), and acquisition (which channels deserve more spend given the AOV they bring, balanced against CAC). AOV vs ARPU is the relevant distinction when you also need to factor in repeat-purchase rate.
A practical workflow: pick the one segment whose AOV most affects gross profit (usually paid-acquisition new customers, since that's where margin is thinnest), set a target AOV uplift tied to a CAC payback goal, then run experiments — free-shipping threshold tests, bundle offers, PDP cross-sells — against that segment specifically. Measure the segment, not the blended number.
The 5-minute AOV health check
Pull AOV for the last 90 days, split by (a) device, (b) new vs returning, (c) top 3 acquisition channels. If any of those segments differs from your blended AOV by more than 25%, you have a segmentation story that your current reporting is hiding. That gap is where your next AOV experiment lives.
Frequently asked questions
AOV = Total revenue / Number of orders, over a defined period. The decisions that matter are which revenue counts (typically net of discounts, excluding shipping and tax) and which orders count (excluding test orders, refunds, and often B2B). Lock those choices down before comparing periods.
For CRO and merchandising decisions, exclude both — you want to measure what the customer chose to buy, not the logistics overlay. For finance reporting, including them is fine. Keep the two views separate and labelled so nobody confuses the trends.
AOV is revenue per order; ARPU (average revenue per user) is revenue per customer over a period and bakes in repeat-purchase rate. A store can have flat AOV but growing ARPU if customers buy more often. See AOV vs ARPU for when to use which.
It depends entirely on category. Apparel typically sits at €65–95, electronics at €85–180, home goods at €120–250. The more useful benchmark is your own historical data for the same segment year over year — external numbers are sanity checks, not targets.
Almost always a definitional gap. Shopify uses post-discount GMV including shipping; GA4 uses whatever your developer mapped into the `value` parameter, which is often pre-tax and post-discount but excludes shipping. Reconcile the definitions before debugging the data.
Blended AOV: weekly, as a directional health check. Segmented AOV (device, channel, new vs returning): monthly, with at least a 4-week window to smooth promotion noise. Cohort AOV: quarterly, since structural drift takes time to show.
For your canonical operating AOV, yes — net of refunds gives you the truthful revenue per order. For real-time dashboards this is sometimes impractical, in which case use gross AOV in-period and reconcile to net AOV monthly once refund cycles settle.
Start with five cuts: channel, device, new vs returning, acquisition cohort, and SKU category. These cover most diagnostic needs without overwhelming the analysis. AOV Segmentation walks through how to layer them for a specific decision.
Neither in isolation. The metric that pays rent is revenue per visitor (RPV = AOV × conversion rate). A tactic that lifts AOV 10% but drops conversion 12% is a loss. Always check both alongside RPV before declaring a winner.
Setting the threshold 20–30% above your current AOV typically lifts blended AOV 5–15%, with the caveat that customers below the threshold sometimes abandon instead of adding. Test it as an experiment against your paid-new segment, not as a sitewide setting change.
Get an AI expert review of your site
Paste your URL — Metricuno's AI runs the same heuristic checks a senior CRO consultant would, scoring your page and prioritising the fixes that'll move conversion fastest.