Average Order Value Explained

Average order value (AOV) is the typical size of an order.

AOV = revenue / orders

Use the average order value calculator. Agree whether revenue is:

• Merchandise value before shipping,
• Including shipping charged to the customer,
• Net of discounts,
• Net of returns.

Any of those can be valid. Mixing them week to week is how AOV “grows” after an accounting change.

AOV is an ingredient in revenue:

Revenue ≈ sessions × conversion rate × AOV

Check the identity with the revenue calculator. How to Improve E-Commerce Conversion Rate is the CVR side. This article is the basket side.

Why AOV moves

AOV is not a personality of the brand. It moves when:

• Mix shifts (more high-priced SKUs, or more accessories).
• People add units (quantity, bundles, “complete the set”).
• People add shipping or paid gift-wrap (if you include those in revenue).
• Discounts change the numerator (discount calculator).
• Bots or staff test orders pollute the denominator (filter them).

Currency and tax display (inclusive vs exclusive) should match the finance definition.

The conversion tradeoff

Tactics that raise AOV often add friction or cut margin.

Tactic	Typical AOV effect	Typical CVR / margin risk
Free shipping over a threshold	Up (people add items)	Down if threshold is far above current AOV; shipping cost hits contribution
Bundle “save 15%”	Up	Margin depends on bundle cost; profit margin
Upsell at checkout	Up	Down if the upsell is slow or irrelevant
Sitewide 25% off	Down (numerator)	Up in CVR; often down in contribution
Remove cheap SKUs	Up	Down in conversion and SEO if those SKUs were entry products

There is no universal “raise AOV 20%” playbook. If the threshold is $120 and AOV is $42, you are training people to bounce, not to add three extra items.

Bundles and thresholds that stay honest

Thresholds. Set them just above current median order value, not at a round number you liked in a workshop. Show the remaining amount in the cart (“$12 away from free shipping”) only if you will honor it. Absorbing shipping is a cost: model it like a discount.

Bundles. Pre-built bundles work when the second item has high contribution and low return risk (consumable refill, complementary accessory). Bundling two high-return fashion SKUs can raise AOV and wreck net.

Quantity breaks. Useful in wholesale-like B2C (packs of 6). Publish the unit price so the customer is not doing mental math against a competitor’s single unit.

Post-add upsells. One relevant offer beats a modal tour. If INP and CLS suffer, you will lose the original conversion—see Core Web Vitals.

AOV vs profit

A $200 order of a 10% contribution SKU with 20% returns is not “better” than a $60 order of a 45% contribution SKU that stays sold. Put AOV next to contribution and return rate on the metrics dashboard.

CAC also interacts: higher AOV can support higher CAC if contribution rose too. If AOV rose only because of a coupon, CAC payback got worse.

How to report it

• Weekly AOV with the same revenue definition.
• Split prepaid vs COD if COD orders are larger and then rejected.
• Split new vs returning (returning AOV is often different; do not hide it in a blend).
• Do not use AOV as a proxy for LTV. Frequency is the other half.

Experiments that do not fool the average

Test one lever at a time and watch contribution dollars, not only AOV:

• Threshold: move free-shipping by a small step toward median order value. If conversion falls more than contribution from larger baskets, you overshot.
• Bundle: a two-SKU bundle vs the same SKUs sold separately at the same total. If people only wanted item A, you hid the single-item path.
• Quantity break: show unit price. Wholesale-style packs fail when the customer cannot compare to a competitor’s each-price.
• Checkout upsell: one SKU, relevant, fast. If INP spikes, you traded AOV for lost paid checkouts—see Core Web Vitals and payment recovery.

Exclude staff orders, wholesale test invoices, and fully refunded orders from the experiment’s AOV if those are not the customer behavior you meant to change.

Mix vs behavior

AOV can rise because you sold more high-ticket SKUs (mix) with no change in basket-building. That is not a “threshold win.” Split AOV by category or by new vs returning before you celebrate a merchandising tactic.

Gift-with-purchase that adds $0 revenue but a unit can lower AOV if you count units in a different report. Keep revenue-based AOV as the default.

Currency and tax-inclusive display should match finance. If the store shows GST-inclusive prices and the dashboard strips tax, AOV will jump on a reporting change that was not a merchandising win. Document the convention next to the chart.

Wholesale or B2B invoices mixed into the same AOV as D2C will inflate the average and hide a weak consumer basket. Split channels.

Gift cards sold as products inflate AOV and confuse merchandise margin. Report them separately if they are material.

The e-commerce analytics guide and the conversion optimization guide cover experiments. Raise AOV when the extra items are wanted, in stock, and still contribute. Otherwise you are decorating the average.