THE SHORT ANSWER
Average order value is usually included order revenue divided by included orders. You can influence it through product mix, quantity, bundles, cross-sell, upsell, thresholds and pricing—but should evaluate contribution, return behaviour and conversion alongside the average.
Define the average before improving it
State whether revenue includes discounts, tax, shipping and refunds, and whether cancelled or test orders are excluded. AOV can change because customers buy more items, select a higher-priced mix, face a price change or receive a different discount.
Compare like-for-like periods and segments. A seasonal product launch can raise AOV while the underlying basket behaviour remains unchanged.
Evidence & context: Google Analytics Help
Choose a lever that makes the order more useful
| Lever | Customer value | Guardrail |
|---|---|---|
| Bundle | A complete solution or simpler choice | Bundle margin and unwanted items |
| Cross-sell | A complementary product | Relevance, clutter and attachment contribution |
| Upsell | A higher-value alternative | Clear comparison and conversion loss |
| Quantity incentive | Lower unit cost or convenient replenishment | Over-discounting and product waste |
| Threshold | Shipping or benefit unlocked above a basket level | Subsidy cost and artificial basket inflation |
| Recommendation | Useful discovery based on context | Bias, privacy and repetitive suggestions |
| Pricing architecture | Clear good-better-best choice | Cannibalisation and perceived fairness |
Higher AOV can produce lower value
Suppose AOV rises from ₹2,000 to ₹2,300 because a ₹400 discount is introduced above a threshold. Revenue per order increases by ₹300, but discount cost increases by ₹400 before product and fulfilment costs. The visible KPI improved while contribution may have fallen.
Measure incremental gross profit or contribution, order conversion, items per order, return rate and fulfilment cost. If the tactic changes customer mix, review new and returning customers separately.
Recommendation quality depends on the objective
Recommendations can use product relationships, popularity or customer context. The ranking objective matters: a system optimized for clicks may not optimize basket usefulness, margin or long-term satisfaction.
Start with explainable merchandising rules when they solve the problem. Add model-driven personalization only when data quality, evaluation and privacy controls justify the complexity.
Evidence & context: Google Machine Learning education
Test the order, not just the module
- Name the segment and placement.
- Measure attachment and AOV.
- Track conversion, contribution and returns.
- Inspect whether the added item remains in the paid order.
- Check whether the effect persists without excessive discount dependence.
Then connect the result to the growth equation rather than treating AOV as an isolated win.
Sources & further reading
- Ecommerce purchases report
Google Analytics Help. Official definitions for item-level commerce metrics. Revenue fields have different inclusions, so teams must document the field they use.
- Recommendation systems: scoring
Google Machine Learning education. Conceptual explanation of scoring and ranking candidates, not a description of every current commercial recommender.
- Product data specification
Google Merchant Center Help. Official product-data requirements covering identity, variants, price, availability, shipping and returns. It is Google-specific, not a universal commerce schema.
Examples and exercises are illustrative unless attributed to a source. No independent expert review is claimed.
A correction, a counterexample or an experience worth sharing?
Join the conversation ↗