Skip to content
Faisal Mahmud TusherDigital marketing, SEO & analytics
with Faisal Tusher

E-commerce Marketing › E-commerce Conversion

Average Order Value Levers

How to raise average order value without hurting conversion: free-delivery thresholds, bundles and kits, cross-sells and upsells, volume pricing, and financing.
Guideintermediate12 min readUpdated

Why AOV is the quiet lever

The store growth equation has four terms: traffic, conversion, order value and repeat. The first two get most of the attention because they are visible in the ad account and the analytics dashboard every morning. Order value is quieter. It moves slowly, it rarely has an owner, and nobody gets a report that says the average basket fell. Yet it multiplies everything else. A store that raises what the average basket is worth earns more from the same ad spend, the same conversion rate (CVR) and the same customer base, which is why order value shows up directly in return on ad spend (ROAS) and in the amount a store can afford to pay for a new customer.

The reason it is a lever rather than a target is that it responds to design decisions the store controls completely: what is offered next to the product, how delivery is priced, how quantities are presented, how the basket page behaves. None of those decisions needs a bigger budget or a platform's permission. They do need honesty about side effects, because every order-value tactic has a way of raising the average while lowering the total. A free-delivery threshold set too high reduces orders. A bundle that discounts too deeply raises the basket and lowers the margin. An upsell that interrupts the checkout raises the value of the orders that complete and reduces the number that do. The job is to find the levers whose side effects are smallest for this catalogue and this customer, and to measure the whole, not the average.

The levers at a glance

Each lever suits a different kind of catalogue. The table maps the common ones to where they tend to work and the risk that comes with them.

LeverWhere it worksMain risk
Free-delivery thresholdBaskets that usually sit just below a natural round figure; low-ticket, many-item cataloguesSet too high, it suppresses orders; set too low, it gives away margin
Bundles and kitsProducts that are used together; gifting; starter sets for new customersDiscount erodes margin; the bundle cannibalises full-price sales of the parts
Cross-sellsAccessories and consumables with an obvious fit to the main itemIrrelevant suggestions train shoppers to ignore the block
UpsellsCatalogues with good-better-best tiers or clear upgrade pathsPushy placement interrupts the purchase; buyers feel steered
Volume pricingConsumables, replenishable goods, trade and gifting buyersLower per-unit margin; stock-ups delay the next order
Financing and payment plansHigher-ticket items where the monthly figure is the real decisionFees, returns exposure, and buyers who would have paid in full anyway
Minimum basket for a promotionDiscount codes that must clear a basket valueTrains customers to wait for codes
Gift wrapping and add-on servicesOccasion-driven purchasesSmall gains; operational cost

The levers near the top of the table are the ones most stores should try first because they are reversible and cheap to test. Financing and volume pricing change the economics of the store itself and deserve a margin model before launch. The one-way and two-way door framework explains this split between levers that can be tried and withdrawn and those that need evidence before they are switched on.

Delivery thresholds

Delivery cost is a common reason shoppers abandon a basket, and a free-delivery threshold turns that objection into an incentive: add one more item and the delivery charge disappears. It works because the delivery fee feels like a penalty while the extra product feels like value, even when the extra product costs more than the fee it replaces.

The threshold is a design decision with a quantitative answer. Look at the distribution of order values rather than the average. If a large share of baskets cluster just below a natural figure, a threshold slightly above that cluster gives most shoppers a reachable target. If baskets are spread evenly or sit well below any sensible threshold, the lever does little except reduce margin on the orders that already clear it. Delivery cost per order, average margin and the mix of light and heavy goods all belong in the calculation, and so does the behaviour of returns: a threshold reached by adding a marginal item often produces a marginal return.

Presentation matters as much as the number. A progress bar in the basket and a line on the product page saying how much more is needed for free delivery do the work; a threshold that is only discoverable on the delivery policy page does not. Where the store runs a mix of delivery options, the threshold usually applies to the standard service only, and the checkout optimisation guide covers how to present the options without reintroducing the surprise-cost problem at the last step.

Bundles, kits and cross-sells

A bundle sells several products as one item at a single price; a kit is a bundle designed as a starting point for a new customer; a cross-sell suggests a related item alongside or after the main purchase. All three raise order value by increasing the number of items per order, and all three depend on the same thing: a genuine fit between the products.

Bundles work best where the products are used together and the shopper would otherwise have to assemble the set by hand. A camera body with the lens and card it needs, a skincare routine of three steps, a coffee machine with its filters. The bundle saves effort first and money second, which is why the discount can usually be modest. A deep bundle discount is a sign the fit is weak and the price is doing the persuading. Gift sets are bundles assembled for givers, and the guide to selling gifts online covers what they need beyond the bundle economics. Kits serve new customers in particular: they answer the question of where to start and establish the habit that the repeat purchase and loyalty guide later builds on. Where a shopper's doubt is whether the product suits them at all, the guide to samples and trial sizes covers the smaller step of a paid trial set before the kit, and the trial size product test is a design for checking whether that smaller first order replaces full-size orders or adds to them.

Cross-sells are the lightest-touch lever and the one most often done badly. Automated blocks that surface whatever sells most across the store teach visitors to ignore the block. Curated cross-sells — the case for the phone, the filter for the machine, the matching trousers — convert because they complete the purchase the shopper is already making. On the product page they belong near the add-to-basket action, which the product page optimisation guide treats as part of the page's answer to what else the shopper needs; in the basket they belong below the line items, never between the shopper and the checkout button. Which products are actually bought together, the evidence a cross-sell or bundle should start from, is a query over the purchase items in the analytics export, as the GA4 merchandising analysis guide shows.

Illustrative example — not a real result

A home-coffee store tests a threshold and a kit

A store selling coffee beans and brewing equipment has an average order of £31 and a free-delivery threshold of £50, which only 18% of orders reach. Looking at the distribution, the team sees a large cluster of orders between £24 and £34: two bags of beans. They lower the threshold to £40 and add a basket progress bar. Over the next eight weeks, the share of orders clearing the threshold rises to 41%, average order value rises to £37, and the delivery subsidy per order rises from £1.10 to £2.60. Conversion rate is unchanged within the test's noise. At the same time they launch a £45 "first grinder" kit (an entry grinder plus two bags) with a £6 saving against the parts. The kit becomes the most common first order for new customers, and contribution per visitor rises by 14% across the period, which the team treats as the number that matters rather than the order-value figure alone.

Upsells and volume pricing

An upsell offers a better version of what the shopper has chosen: the larger size, the higher tier, the model with the extra feature. It differs from a cross-sell in that it replaces the item rather than adding to it, which means it changes the item's margin as well as its price. Upsells work when the catalogue has a clear ladder — good, better, best — and when the step up is explained in terms the shopper cares about rather than in specifications. A larger bottle that lasts twice as long per use sells; a line called the premium range does not. Good-better-best tiering for product ranges sets out how to give each rung its own buyer and a clear reason to step up.

Placement decides whether an upsell helps or hurts. On the product page, a tier comparison lets the shopper choose before they commit and costs nothing in conversion. In the basket, a quiet upgrade line can work for low-consideration items. In the checkout, interruptions nearly always cost more orders than they gain in value, and post-purchase upsells — offered on the confirmation page after payment — are the safe place for anything that might otherwise distract.

Volume pricing lowers the per-unit price as quantity rises. It suits consumables and anything a customer uses continuously, and it has a hidden cost beyond the margin given up: a customer who stocks up today does not come back next month. For a store whose economics depend on repeat orders, volume pricing can shift revenue forward without increasing it. Compare the lifetime value (LTV) of customers who buy in volume against those who buy at normal cadence before deciding that the larger basket is a win.

Financing and payment plans

For higher-ticket items, the barrier is often the single payment rather than the price. Instalment options, buy-now-pay-later services and store credit break the price into a monthly figure, and the lever's effect on order value comes from two sources: shoppers who would not have bought at all, and shoppers who choose a higher tier because the monthly difference looks small.

The costs are real. Providers charge the merchant a fee per transaction, returns on financed orders involve the provider's process as well as the store's, and some share of financed orders would have been paid in full anyway, which turns the fee into pure cost on those. Financing also attracts regulatory attention in most markets, and the way instalment options are presented — including how clearly fees and the consequences of missed payments are shown — is subject to consumer-credit and advertising rules that vary by country. Treat the provider's documentation and local guidance as the source and keep the presentation plain; this page describes the lever, not the legal position.

The operational test is simple: measure order value, conversion, margin after fees and the returns rate for financed orders separately from everything else. If financed orders are larger but return more often or carry fees that consume the extra margin, the lever is working for the provider rather than the store.

Measuring AOV against margin and conversion

Average order value on its own is a vanity number. It can rise because a tactic worked, because small orders stopped happening, because the product mix shifted towards expensive items, or because a promotion pulled large orders forward. The measure that survives those confusions is contribution per visitor: margin after discounts, delivery subsidy, payment fees and returns, divided by the number of visitors in the period. A lever is working when that number rises, whatever happens to the average. Where the average is read from analytics rather than the shop, check how the tool builds it first: AOV in GA4 explains why Google Analytics' average purchase revenue usually differs from the shop platform's figure.

Three practices keep the measurement honest. First, compare the distribution, not just the mean: a histogram of order values before and after shows whether the change came from the baskets the lever was aimed at. Second, run levers as tests where the platform allows — thresholds, bundles and cross-sell placements can all be split-tested — and read conversion and order value together, because a lever that raises one and lowers the other may net out to nothing. Third, follow cohorts for a few months after the change: volume pricing and deep bundles can show as a gain this quarter and a loss next quarter when the stocked-up customers do not return. For a planned sale, the sale promotion readiness checklist asks for the margin check and the comparison period to be agreed before the discount goes live.

For paid traffic, the ROAS calculator turns the new order value into the return the ad account should expect, and the ROAS glossary entry explains why that figure still needs margin behind it. When the order value comes from a sale, the ROAS calculator walkthrough for discount promotions shows which margin to enter. The e-commerce conversion hub sets order value alongside the other conversion levers so that it is planned with them rather than after them.

How deep a promotion's discount should be is a question the discount depth test answers on gross profit per recipient, with a no-offer arm as the anchor.

The AOV versus revenue per visitor comparison explains why revenue per visitor should lead tests of these levers, with order value as the diagnostic.

When the average falls and nobody changed a lever, the diagnosis page for a dropping average order value separates a shift in customer mix and a change in how value is recorded from a real shrinking of baskets.

Common mistakes

  • Setting the threshold from the average. The average hides the cluster; the distribution shows where a threshold can be reached.
  • Discounting bundles to make them sell. A bundle that needs a deep discount has a weak fit; fix the fit or drop the bundle.
  • Automated cross-sells nobody curates. Irrelevant suggestions train shoppers to skip the block, and the block then fails when it is relevant.
  • Upselling in the checkout. Interruptions at the payment step cost more orders than they gain in value; move the offer to the product page or after purchase.
  • Reading AOV without conversion. A lever that raises one and lowers the other is not a win until profit per visitor says so.
  • Ignoring the next order. Volume pricing and stock-up promotions pull revenue forward; check what the cohort does next quarter.
  • Treating financing as free. Provider fees, returns handling and the presentation rules belong in the margin model before launch.

Frequently asked questions

Does raising average order value always improve profit?

No. Average order value can rise because small orders disappeared or because discounts pulled large orders forward, and both can leave the store worse off. Judge a lever on contribution per visitor after discounts, delivery subsidy, fees and returns, and read it alongside conversion rate rather than on its own.

Where should a free-delivery threshold be set?

Just above the point where a large cluster of baskets already sits, so that most shoppers can reach it by adding one item. The order value distribution, delivery cost per order and average margin decide the figure. The average order value alone is a poor guide because it hides the cluster.

Which AOV lever should a small store try first?

A free-delivery threshold with a basket progress bar and a handful of curated cross-sells are the cheapest to run and the easiest to reverse. Bundles come next once the store knows which products are genuinely bought together. Financing and volume pricing change the store's economics and need a margin model first.

Terms used on this page

Tools to try

Frameworks

Where this page is used

Part of these learning paths

Also related

Related