"Should we raise our cashback from 5% to 7%?" — this is the question we're asked most often. The answer depends on 3 factors: your gross margin, your AOV, and your sector's psychological threshold. We analyzed 47 A/B tests run in our portfolio between 2024 and 2026. Here's the real data.
The marginal returns curve
First lesson, perhaps the most important: the relationship between cashback rate and share rate is not linear. It looks more like an S-curve:
- 0-2%: near-zero effect. Sharing stays at "organic" level (customers who share out of brand love, without incentive).
- 3-5%: take-off zone. Each additional point of cashback adds 6-9 points of share rate.
- 5-7%: optimal zone for most sectors. Marginal gains of 3-5 points per %.
- 7-10%: marginal gains drop to 1-2 points per %. Margin erodes faster than share rises.
- >10%: ceiling effect. Some customers think "scam" and share less than at 8%.
Global portfolio sweet spot: 5-7% by sector, with an absolute cap in euros.
The 3 psychological thresholds
Beyond the quantitative curve, we identified 3 psychological thresholds causing non-linear jumps:
Threshold 1: the "meaningful amount" (min €3)
Below €3 of cashback in absolute value, sharing drops heavily. For a €30 basket, that's a 10% floor. This is why food (low AOV) performs better in voucher than in cash.
Threshold 2: the "psychological ratio" (10% perceived)
Customers prefer a €5 cashback on €50 (10%) over €7 on €100 (7%) — even if absolute value is higher in the second case. Ratio speaks louder than amount.
Threshold 3: the "credibility ceiling" (max 15%)
Above 15%, customers grow suspicious. Sharing drops, friends receiving the link wonder "is this a scam?". The rare brands that can afford >15% are those with a strong brand and very high margins (luxury, editorial).
3 notable sector A/B tests
Case 1: beauty, move from 4% to 6%
Beauty brand, €280K/month, AOV €62. 50/50 test over 6 weeks.
- 4% variant: share rate 54%, commissions paid €6,048
- 6% variant: share rate 71%, commissions paid €11,928
- Delta referred revenue: +31%
- Delta net margin after cashback: +19% (incrementality offsets the extra cost)
Decision: stay at 6%. Marginal ROI is positive.
Case 2: food, test cash 4% vs. voucher €12
Food brand, €180K/month, AOV €38. Equivalent budget.
- Cash 4% (= €1.50): share rate 19%
- Voucher €12 usable from €40: share rate 47%
- Voucher redemption rate: 38%
- Additional orders: +212%
Decision: voucher for life. Cash doesn't work under €40 AOV.
Case 3: premium accessories, test 7% vs. 10%
Accessories brand, €420K/month, AOV €185. 50/50 test over 4 weeks.
- 7% (= €13): share rate 48%
- 10% (= €18.50): share rate 49%
- Delta referred revenue: +2% (negligible)
- Delta margin after cashback: -11%
Decision: back to 7%. Beyond this threshold, marginal is negative.
Recommendations by sector
Based on 47 A/B tests and portfolio expertise:
- Beauty: 5-6% cash, €20/share cap
- Fashion: 5-7% voucher (not cash) usable >€60
- Sport & outdoor: 4-5% cash, €25 cap
- Accessories & jewelry: 6-7% cash, €35 cap
- Food & beverage: 3% cash + €8-12 voucher referee (not cash-only referrer)
- Home & décor: 7-8% cash, validation >€120, 90-day window
- Luxury & premium: 3-4% cash, €80 cap (argument matters more than amount)
The 3 A/B tests worth running (and those that aren't)
Not all tests are worth it. Here are the 3 to launch as priority:
- Cash vs. voucher (if AOV <€50): near-systematic on food, basic cosmetics, small décor. Duration: 4 weeks. Min traffic: 800 orders.
- Rate tier test (-1 point below sector threshold): test 1 point below your current rate to verify you're not above the diminishing returns threshold. Duration: 6 weeks.
- Absolute cap (change the cap, not the %): often underestimated. Moving the cap from €15 to €25 on an €85 AOV can boost sharing 8 points without changing the %.
Not worth it: testing >10% vs. 10%, testing non-round rates (5.3%, 6.7% — the brain rounds), testing cash vs. wallet if AOV >€100 (cash always wins).
Conclusion: data beats intuition
Cashback rate is the most debated variable internally among merchants — and the most often decided by founder intuition. A/B tests cost 6 weeks and return 3-20% net margin. That's probably the best test ROI you can run in 2026.
Our advice: get out of your head, run the test. The data is there. You just have to ask the question.