I often get asked a deceptively simple question: “How big should our welcome bonus be if our average order value (AOV) is £250 and we want it to pay back within three purchases?” The short answer is: there’s no single magic number — it depends on margins, redemption behaviour and how much of the activity is truly incremental. But we can turn that mess into a clear formula and a practical calculator you can use right now. Below I’ll walk you through the logic I use when advising SMEs, show worked examples, and give a ready-to-use table so you can plug in your own inputs.
What “pays back within three purchases” actually means
When I say “pays back within three purchases” I mean the incremental gross profit generated by the customer across their first three purchases should cover the full cost of the welcome bonus (face value + fulfilment/operational cost). This is important: we’re talking incremental profit, not revenue. If a customer would have bought anyway, their purchases aren’t incremental and don’t count toward payback.
So the core components we need are:
The simple payback formula
Here’s a compact way to express payback across N purchases (we’ll use N=3):
Payback condition: Cost_of_Bonus ≤ Sum_{i=1..N} (AOV_i × Margin × Incrementality_i)
If AOV is stable and we assume the same incrementality and margin across the first three purchases, it simplifies to:
Cost_of_Bonus ≤ N × AOV × Margin × Incrementality
Rearrange to solve for the maximum Cost_of_Bonus you can afford:
Max_Cost_of_Bonus = N × AOV × Margin × Incrementality
That cost can be the face value of the welcome bonus plus any fulfilment fees. If your program gives points rather than a straight discount, convert expected monetary value accordingly.
Plugging in realistic numbers — worked example
We’ll use numbers I commonly see in SME retail clients:
First we calculate the incremental gross profit across three purchases:
Incremental_Gross_Profit = 3 × £250 × 0.40 × 0.60 = 3 × £250 × 0.24 = 3 × £60 = £180
This £180 is the maximum total cost we can assign to the welcome package (face value + fulfilment costs) if we want payback inside three purchases. If fulfilment is ≈5% of face value, we solve for face_value (FV):
Cost_of_Bonus = FV + 0.05 × FV = 1.05 × FV ≤ £180 → FV ≤ £171.43
So, in this example you could offer a welcome discount roughly worth £170 and still expect payback within three purchases — assuming our margin and incrementality assumptions hold.
Why this can look surprisingly high
Many founders are surprised that the allowed welcome discount can be this large. That’s because customers making £250 orders with high margins generate a lot of gross profit per purchase. Two important caveats:
Sensitivity table — see how inputs change the result
The table below shows Max_Face_Value for a few common scenarios (3 purchases), with fulfilment cost assumed at 5% of face value.
| Scenario | Margin | Incrementality | Max Cost of Bonus | Max Face Value (≈) |
|---|---|---|---|---|
| Optimistic | 50% | 70% | 3×250×0.5×0.7 = £262.50 | £262.50 / 1.05 = £250.00 |
| Base | 40% | 60% | £180.00 | £171.43 |
| Conservative | 35% | 40% | 3×250×0.35×0.4 = £105.00 | £100.00 |
| Low incrementality | 40% | 30% | £90.00 | £85.71 |
What to watch for — practical adjustments I always make
When I run these numbers with a client, we don’t stop at the headline result. Here are the adjustments I always consider:
Example welcome offer structures that work for a £250 AOV
Depending on the number we settle on, here are pragmatic offer types I recommend:
How to validate quickly (A/B test)
The final step is always an experiment. My usual test plan:
One client I worked with used a £60 welcome voucher for £240 AOV customers and thought it was expensive. After a 90-day cohort analysis they found a 35% uplift in second purchase rate and a three-purchase payback — because their margins and incrementality were strong. Data replaced gut instinct.
If you want, send me your actual margin and an estimate for how incremental you think those first purchases will be and I’ll run the numbers and suggest realistic offer structures you can A/B test. I include a small model in my onboarding pack for clients, but I’m happy to sketch something based on the inputs you give me here.
