Average order value: 7 Checks Before You Chase a Bigger Basket

Average order value, or AOV, measures the value of the typical order in arithmetic terms: an eligible revenue total divided by the corresponding number of orders. It helps a business understand purchase size, but it does not show customer profit, conversion, or repeat buying on its own.

A larger basket can be useful when it creates sustainable contribution and meets a customer need. It can also result from discounts, price changes, or a few unusually large orders. This article explains the average order value formula, gives worked examples, and sets out seven checks before using AOV as a growth target.

The average order value formula

Average order value = revenue included under the chosen definition ÷ eligible orders for the same scope and period.

The arithmetic is simple; the definition needs care. Decide how the report treats discounts, taxes, shipping, gift cards, returns, cancellations, and later adjustments. Do not assume that every commerce platform includes the same amounts.

Shopify’s sales report documentation describes its own AOV convention. Its analytics field reference provides further metric definitions. Use those as product-specific examples rather than a universal accounting rule.

Keep the numerator and denominator aligned. If the revenue includes one channel but the order count includes several, the result will be misleading even when the formula runs correctly.

A worked example

Consider a hypothetical shop with 20,000 of eligible order revenue across 400 orders in a month. AOV equals 50 in the reporting currency. If revenue rises to 24,000 across 400 comparable orders, AOV becomes 60.

That is a 20% increase in the average, but it does not prove a better commercial result. Prices, discounts, product mix, returns, and fulfillment costs may have changed. The business needs to inspect those drivers.

Now suppose a campaign generates 10,000 across 100 orders, giving AOV of 100, while another generates 12,000 across 200 orders, giving AOV of 60. The higher AOV campaign does not automatically create more revenue or profit. Order volume and cost still matter.

7 checks before acting on average order value

1. Define the revenue boundary

Write down what the revenue figure includes. A product-sales measure excluding shipping and tax differs from a total amount paid by the customer. Both may answer useful questions, but they should not share an unexplained label.

State how discounts and gift cards work. Issuing a gift card and redeeming it can create double-counting risk if the report treats both events as ordinary merchandise sales without a clear policy.

Keep the definition consistent across periods. If the business changes its reporting policy, explain the effect instead of presenting the difference as a basket-size improvement.

2. Choose which orders count

Decide how the report treats canceled, test, fraud, replacement, and zero-value orders. The eligible order count should correspond with the revenue scope. Do not remove low-value orders simply because they make the average less attractive.

Use stable order identifiers. A split shipment can create several fulfillment records while remaining one customer order. Counting each shipment as an order changes the measure.

For service businesses, define the unit carefully. An enquiry, quote, invoice, job, and contract are different records. If the business calls the measure average job value, use a job definition rather than importing an ecommerce label without adaptation.

3. Handle returns and later changes explicitly

AOV at order placement can differ from a later net-sales-per-order measure after refunds and edits. Decide which question the report answers and preserve the distinction.

For example, a promotion may generate large initial baskets that later return at a high rate. A placement-only measure can still describe those orders accurately while failing to show their eventual value. Add a mature outcome view where useful.

Do not mix a current refund total with unrelated current orders without understanding the timing. Returns may refer to purchases from earlier periods. Cohort-based analysis can help when the question concerns the final outcome of a campaign’s orders.

4. Inspect the distribution and mix

An average can rise because a few large orders enter the period. Show the median, order-value bands, or a relevant distribution when those views explain the result. AOV alone does not describe every customer.

Segment by channel, product group, customer type, and location where the data supports a useful decision. A wholesale order and a small consumer purchase may belong in separate operating views.

Also inspect quantity and price. Higher AOV can result from more items, higher prices, or a different mix. Each suggests a different next question and a different risk when the business designs an offer.

5. Review contribution and fulfillment cost

A larger order may carry lower margin, more shipping cost, or more service effort. Estimate contribution after the relevant variable costs and discounts. A revenue increase is not automatically a profit increase.

For bundles, check whether the offer adds useful items or mainly discounts goods the customer would have bought anyway. Keep the cost assumptions visible and test actual redemption behavior.

Also consider operational capacity. An offer that drives bulky or complex orders may create packing, installation, or support demands. The commercial review should include the team responsible for delivery.

6. Balance AOV with conversion and retention

A higher minimum purchase or aggressive upsell can increase AOV among buyers while discouraging other customers from completing an order. Review conversion and total contribution alongside basket size.

Customer experience matters too. An irrelevant add-on may raise one transaction’s value but weaken trust. Offer useful choices and make the price and terms clear. Avoid making the customer navigate unnecessary friction to buy what they need.

Review later returns, complaints, and repeat behavior. AOV is a transaction measure, while customer value develops over time. Do not optimize the first order at the expense of the relationship without understanding the tradeoff.

7. Test a specific growth hypothesis

Choose one change with a clear reason. Examples include a relevant accessory suggestion, a practical bundle, or clearer delivery options. Define the outcome and comparison before launching the test.

Measure the complete result: AOV, order count, conversion, contribution, returns, and service issues where relevant. A favorable movement in one number should not settle the decision.

Record other changes such as price revisions, seasonality, or channel mix. If several things change at once, explain the uncertainty rather than assigning the whole outcome to the experiment.

A bundle example with a margin check

Imagine a hypothetical retailer selling a main item for 80 with variable cost of 50. A proposed bundle adds an accessory and sells for 95, with total variable cost of 62. The contribution rises from 30 to 33 per bundle sale under this simplified boundary.

The larger order creates only three additional units of contribution, not the full 15 increase in revenue. If the bundle also increases fulfillment cost or reduces conversion, the overall result may be weaker than expected.

These figures illustrate the review method. Use the business’s actual costs and a sensible test before adopting the offer. A higher AOV is evidence of a larger order, not proof that the offer pays for itself.

Use customer and order cohorts carefully

Separate first orders from repeat orders when that distinction matters. New customers may respond differently to an offer from established customers. A pooled average can hide both patterns.

Keep identity rules clear. A guest checkout, changed email, or shared household address can complicate customer matching. Avoid assuming that every new record represents a genuinely new customer.

For campaign evaluation, retain the original audience or order cohort and follow its later outcomes. This helps distinguish an initial transaction effect from the final value after returns or adjustments.

Build a report that shows the arithmetic

Display eligible revenue, eligible order count, AOV, reporting period, and key exclusions. Add a short breakdown for material mix changes. A reader should be able to reproduce the main result from the source data.

Anomoz’s article on data literacy for reporting explains why definitions and context matter. A chart labeled “order value” should not force users to guess which revenue version it includes.

Also review currency and time-zone settings. A cross-border store may need a reporting-currency convention, while a late-night business may define its operating day differently from midnight. Keep the basis consistent.

Data checks before an automated dashboard

Test a known set containing a normal order, discount, cancellation, partial refund, gift-card transaction, and split fulfillment. Confirm the expected revenue and count under the chosen definition.

Check duplicate events and later updates. An integration retry should not create a second order, and a corrected transaction should follow the report’s adjustment rules. Keep enough history to explain a restated period.

Finally, assign an owner for the metric. When finance, marketing, and operations use different versions for different purposes, label them distinctly. Agreement on the question is more useful than forcing every team into one ambiguous number.

Apply AOV differently in a service business

A contractor may use average accepted quote value or average completed job value. Those measures can support planning, but each needs its own status and time boundary. Quotes that never become work should not inflate a completed-job measure.

Where customer, quote, and job records are disconnected, custom crm development may help connect the workflow. Define the commercial events before building the report.

Keep scope and profitability visible. A larger contract can create more risk and delivery effort as well as more revenue. Average value should inform a decision, not replace the assessment of the work.

Common mistakes

• Changing revenue definitions between periods.
• Counting shipments or invoices as orders without explanation.
• Ignoring returns and later adjustments.
• Letting a few large orders define the whole story.
• Celebrating AOV while conversion or margin falls.
• Assuming an upsell is useful because it increases revenue.

Also avoid borrowing a target from an unrelated business. Product price, purchase frequency, channel, and customer mix can make the comparison unhelpful. Your own consistent history and controlled tests provide a stronger starting point.

Check the customer choice behind a larger order

A useful basket-growth experiment should help the customer make a relevant choice. Review whether the suggested accessory, bundle, or service genuinely fits the main purchase. An offer can increase order value while creating confusion or unnecessary returns.

Ask staff to record common questions during the test. Customers may misunderstand compatibility, quantities, or delivery conditions. Those observations can explain an outcome that the transaction data alone cannot show.

Also inspect the distribution of the benefit. If only a few unusually large orders drive the change, the result may not support a broad rollout. Review comparable groups and avoid assuming that every customer responded in the same way.

Retain the offer terms, audience, dates, cost assumptions, and outcome. This creates a practical learning record and makes later comparisons more credible than relying on a campaign name and one average.

Grow useful order value, not just the average

Average order value helps when the revenue and order definitions are clear. Inspect mix, costs, conversion, and later outcomes before acting. Then test a relevant offer and retain it only when the complete result supports the change.