Expert analysis · reviewed 6 Oct 2026
Cross-selling: measuring the margin added, not the purchases that would have happened anyway.
Key distinctions
Three conditions before calculating.
A purchase rate is not an effect
Some customers would have added the product without a recommendation. Only the gap with a control group measures what the offer changes.
The discount applies to every buyer
A cross-sell discount also benefits customers who would have paid full price. Its real cost includes this subsidy, which often exceeds the gain.
An added purchase can replace another
The product offered can substitute for an item in the basket or for a future purchase. The net margin must account for these displaced sales.
Method
Measuring a cross-sell offer in four steps
- 01Define the offer and the moment
Specify the main product, the complementary offer, the touchpoint and any discount before measuring anything.
- 02Hold out a control group
Do not show the offer to a random share of eligible customers, so as to know the purchase rate without it.
- 03Calculate the incremental margin
Multiply the gap in purchase rate by the margin after discount, then subtract the discount given to buyers who would have bought anyway.
- 04Check for displaced sales
Compare the full basket and purchases over the following weeks between exposed and control customers to detect substitutions.
Three cross-sell offers, ranked by incremental margin
The purchase rate with the offer is compared with that of a control group that does not see it. The most purchased offer is not the one that earns the most.
| Cross-sell offer | Purchase rate with offer | Control-group purchase rate | Incremental purchase | Unit margin | Incremental margin per 1,000 baskets |
|---|---|---|---|---|---|
| Extended warranty | 8% | 2% | 6 points | €40 | €2,400 |
| Protective case | 12% | 9% | 3 points | €15 | €450 |
| Wireless mouse | 15% | 14% | 1 point | €10 | €100 |
Rank cross-sell offers by incremental margin, measured against a control group, not by purchase rate.
A discount that lifts sales by 60% and loses money
Illustrative example: across 10,000 orders, a 20% discount on a €50 accessory, which yields €25 of margin before discount, raises its purchase rate from 10% to 16%. The control group, without the offer, buys it in 10% of orders.
Margin added = 10,000 × (16% − 10%) × (€25 − €10) = €9,000. Discount given to buyers who would have paid full price = 10,000 × 10% × €10 = €10,000. Net effect = 9,000 − 10,000 = −€1,000.
The accessory's purchase rate rises by 60%, but the offer loses money: it mainly rewards customers who would have bought it at full price. A discount reserved for baskets without the accessory, or a smaller one, would change the result.
Acceptance conditions
What must be true to act.
- 01
Measure every cross-sell offer against a control group that does not see it.
- 02
Assess the offer on incremental margin, never on attributed revenue.
- 03
Include in the cost the discount given to buyers who would have paid full price.
- 04
Suspend an offer whose net incremental margin stays negative over two measurement periods.
Limits
What this analysis does not prove.
- The control group deprives some customers of a useful offer during the measurement.
- Effects on satisfaction and future loyalty escape a short-term measurement.
- Personalised recommendations change with the algorithm, which makes successive measurements hard to compare.
- A result obtained on one main product does not automatically transfer to others.
References
Works cited.
- Shah, Kumar, Qu & Chen (2012), Unprofitable Cross-Buying: Evidence from Consumer and Business Markets (opens in a new tab)
- Kumar, George & Pancras (2008), Cross-Buying in Retailing: Drivers and Consequences (opens in a new tab)
- Li, Sun & Wilcox (2005), Cross-Selling Sequentially Ordered Products: An Application to Consumer Banking Services (opens in a new tab)