Win-back of lapsed customers: campaign payback and break-even response
Calculates whether a campaign to win back lapsed customers pays off: how many respond, how much margin is left after the discount, what part would have bought anyway and at what response the campaign breaks even. The discount and the response are your numbers.
At a glance
- What question does it answer?
- Will a win-back campaign pay off and at what response does it break even.
- What data is needed?
- Customers in the mailing, the cost per contact, the expected response and the share of responders who would have bought anyway.
- How to read the result?
- On the example of a mailing to 20,000 lapsed customers (6 per contact, 120,000) and a response of 3%, 600 customers respond and make 840 orders. The margin of an order on the offer is 720 (ticket 1,800, margin 55% minus a discount of 15%), 604,800 in all. 20% of responders would have bought without the mailing: that is 166,320 of margin that cannot be credited to the campaign. The additional margin is 438,480, the result 318,480, the payback 265.4%. The campaign breaks even at a response of 0.82%. Example on the default values: Responders: 600; Orders of the responders: 840; Margin on orders after the discount: RUB 604,800; Margin that would have come anyway: RUB 166,320. The figures are illustrative: enter your own to see your result.
Why and how to reason
A mailing to lapsed customers seems cheap, but it has a hidden cost: the discount that also goes to those who would have bought anyway. So an honest calculation counts only the additional margin: margin of promotional orders minus margin of those who would have come on their own. Of 20,000 customers 3% respond, 600; each makes 1.4 orders, 840 in all. Promotional order margin = order value × (margin before discount − discount) = 1,800 × (55% − 15%) = 720, total 604,800. Among responders 20% would have bought anyway: 840 × 20% × 990 (margin without discount) = 166,320. Additional margin is 438,480, contact costs 20,000 × 6 = 120,000, result 318,480. Break-even response is 0.82%: below it the campaign loses. Discount and response are your numbers, so the calculation gives the threshold at which the decision changes.
When to use it
Use it before a mailing or a promotion for dormant customers to decide which discount and list are justified. It fits a brand or retailer with a contact base and mailing consent. It does not fit when the response is unknown: first run a test on a small group. Assumptions: margin and order value are constant, the share “would have bought anyway” is your estimate, one response period.
How to read the result
The key figure is “break-even response”. If your test gives a higher response the campaign pays; if close the risk is high. Then compare the additional margin with costs: payback shows how much profit a rouble of cost brings. If response is sensitive to the discount, calculate several variants: a smaller discount can give more profit at a slightly lower response. Check the “would have bought anyway” share too: at 50% the promotion mostly gives the discount to your own. The decision: scale, lower the discount or narrow the list to those who left recently. Do not forget to check the result against a control group with no mailing: only it shows how many responses you created rather than would have received anyway.
Common mistakes
Counting all revenue of responders rather than margin after discount. Forgetting the cost of contact and segmentation. Mailing without the customer’s consent: this breaks advertising and personal data requirements. One more mistake: calling customers “lapsed” by too short a window: some buy rarely and would return on their own.
What data is needed?
- Customers in the mailing, the cost per contact, the expected response and the share of responders who would have bought anyway.
- Orders per responder, the average ticket excl. VAT, the margin before the discount and the discount in the offer.
Inputs
| Field | Unit | Value in the example | Meaning |
|---|---|---|---|
| Customers in the mailing | customers | 20,000 | Lapsed customers who consented to communications; check the consent to mailings with a lawyer (laws 152-FZ and 38-FZ). |
| Cost per contact | ₽ | 6 | A mailing, a message, a bonus per contact. Amount excl. VAT: the general rate is 22% from 01.01.2026. The rate follows the shipment date; confirm your regime and rate with an accountant. |
| Response: the share who bought after contact | % | 3 | Your own estimate from past campaigns; there is no single standard in Russia. |
| Of the responders would have bought anyway | % | 20 | Determined by a control group without the mailing. An example, not a standard. |
| Orders per responder in the period | × | 1.4 | How many orders a customer who bought after the campaign makes. |
| Average ticket excl. VAT before the discount | ₽ | 1,800 | Amount excl. VAT: the general rate is 22% from 01.01.2026. The rate follows the shipment date; confirm your regime and rate with an accountant. |
| Gross margin before the discount | % | 55 | The share of margin in a ticket with no discount. |
| Discount in the win-back offer | % | 15 | The discount that every responder gets. |
How it is calculated
| Step | Formula |
|---|---|
| Order margin on the offer | ticket × (margin before the discount − discount) |
| Additional margin | orders × margin on the offer − orders × “anyway” share × margin with no discount |
| Break-even response | costs ÷ (customers × orders per responder × (margin on the offer − “anyway” share × margin with no discount)) |
Worked numbers
Default values from the app; in the interactive version you enter your own.
- Responders
- 600
- Orders of the responders
- 840
- Margin on orders after the discount
- 604,800 ₽
- Margin that would have come anyway
- 166,320 ₽
- Additional margin of the campaign
- 438,480 ₽
- Contact costs
- 120,000 ₽
- Campaign result
- 318,480 ₽
- Payback: result ÷ costs
- 265.4 %
- Break-even response
- 0.8 % · now 3 %
Margin per order
| Item | Value |
|---|---|
| Order margin with no discount | 990 ₽ |
| Order margin on the offer | 720 ₽ |
| Additional margin per responder | 731 ₽ |
Determine the “would have bought anyway” share with a control group that got no mailing; without it the result is overstated. Only one campaign period is counted.
Nuances and limits
- Having no control group: without it the result is overstated.
- Forgetting that the discount is also received by those who would have bought at full price.
- Mailing without checking consent; check the requirements with a lawyer (152-FZ, 38-FZ).