AOV (Average Order Value)
.webp)
An e-commerce indicator representing the average value of a purchase made by a customer. It is calculated by dividing total sales by the number of orders over a given period:
AOV= Total revenue ÷ Number of orders
Example:
If a site generates €10,000 in sales from 200 orders, the AOV is €50.
In an A/B testing context:
AOV is often used to measure the impact of a variation on the value generated per order (e.g.: testing a cross-sell module, a new promotion or an upsell at the checkout).
However, AOV is a continuous variable sensitive to extreme values (e.g., a few very large baskets can inflate the mean). This makes it statistically unstable for conventional A/B tests based on averages (such as Student's t-tests).
➡️ We therefore recommend using a non-parametric test such as the Mann-Whitney U test, which is more robust to outliers.
CRO recommendation:
It is not advisable to use AOV directly as the main KPI for validating an A/B test, as its distribution is rarely normal.
👉 A more reliable alternative is to transform the KPI into a binary rate, for example :
- % of sessions with basket addition greater than €X
- % of sessions with at least one product from the premium range
- Conversion rate for a specific cross-sell
These metrics are more stable, compatible with conventional A/B testing, and more directly actionable to guide optimization decisions.