First-order discount: CAC and LTV
Calculates the margin of a discounted first order, repeat purchases over 90 and 180 days, the acquisition cost and the customer payback period; compares with spending the same money without a discount if you have a control group.
At a glance
- What question does it answer?
- Does a customer who came on a first-order discount pay back and is that better than acquisition without a discount.
- What data is needed?
- The discount, number of new customers, first basket excl. VAT, margin at full price.
- How to read the result?
- At a 20% discount the margin of the first order is 1,750 (loss on the discount 1,000) and CAC is 2,000: the customer pays back within 90 days (2,245), over 180 days 2,575 and the cohort earns 575,000. 75% did not come back. A control group without a discount for the same spend earns 695,000, which is 120,000 more. Example on the default values: Margin of the first order — RUB 1,750; Loss on the discount per customer — RUB 1,000; Customer margin over 90 days — RUB 2,245; Customer margin over 180 days — RUB 2,575.
Why and how to reason
A first-order discount buys a new customer, but the brand pays out of margin. So we compute two numbers: what the customer brought and what he cost. First order: ticket RUB 5,000, margin 55%, discount 20%, so order margin 5,000 × (0.55 − 0.20) = RUB 1,750, loss on the discount RUB 1,000. Repeat orders: in 90 days 0.18 orders × 5,500 × 50% = RUB 495, in 180 days 0.3 × 2,750 = RUB 825. Customer margin over 90 days RUB 2,245, over 180 days RUB 2,575. Acquisition cost (CAC) 2 million ÷ 1,000 = RUB 2,000: the customer pays back in 90 days. Cohort profit over 180 days 1,000 × (2,575 − 2,000) = 575 thousand ₽. But a control group without discount (700 customers, 0.4 orders in 180 days) would have brought 700 × 3,850 − 2 million = 695 thousand: the discount does not win here.
When to use it
Reach for it when launching a new-customer discount and evaluating its effect by cohort. You need real repeat orders at 90 and 180 days, and a control group without discount makes the conclusion reliable. It does not fit one-off purchase goods or products where the second purchase comes after a year. If there is no control group, compare at least with a past cohort and remember that without control the conclusion is weak.
How to read the result
Look at customer payback time and cohort profit. If a customer pays back longer than the horizon you are willing to fund, the discount is too deep. Comparison with control answers the main question: did the discount bring more customers than it lost in margin. If control gives more profit, as here, cut the discount or replace it with a gift. The non-returning share (75%) suggests repeat-purchase work matters more than discount depth. If the control shows the discount does not win, test a replacement: a gift with the first order, free delivery, or a smaller discount with a higher basket threshold.
Common mistakes
Typical trap: counting customer margin before the discount; using yearly repeat orders instead of 180 days; ignoring that discount hunters come and do not return; leaving out acquisition costs. Acquisition costs must be the same as in the control. Do not wait for a yearly customer payback if you pay for acquisition every month.
What data is needed?
- The discount, number of new customers, first basket excl. VAT, margin at full price.
- Repeat orders per customer over 90 and 180 days and the share who came back.
- Optional: acquisition spend and a control group without a discount.
Inputs
| Field | Unit | Value in the example | Meaning |
|---|---|---|---|
| First-order discount | % | 20 | The first-order discount in percent. |
| New customers | people | 1,000 | How many new customers came through the offer. |
| Average first basket before discount, excl. VAT | ₽ | 5,000 | Price 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. |
| Margin at full price | % | 55 | Gross margin of a first order at full price, as a percent of revenue excl. VAT. |
| Acquisition spend (optional) | ₽ | 2,000,000 | Advertising and everything that brought these customers. Empty: zero. |
| Repeat orders per customer in 90 days | orders | 0.18 | Cumulative: how many repeat orders came per new customer. |
| Repeat orders per customer in 180 days | orders | 0.3 | The average number of repeat orders per customer within 180 days after the first. |
| Came back within 180 days | % | 25 | Share of customers with at least one repeat order. The rest did not come back (“bargain hunters” and one-off buyers). Empty: not shown. |
| Repeat basket excl. VAT (optional) | ₽ | 5,500 | The average repeat-order ticket excl. VAT among customers who bought again; empty means the repeat ticket equals the first. |
| Repeat-order margin (optional) | % | 50 | Empty: same as the full-price margin. |
| New customers without a discount | people | 700 | How many customers the same spend would bring without a discount (from a control group or a test). |
| Repeat orders per customer in 180 days (control) | orders | 0.4 | The same for a control group without the discount. Without a control it cannot be said that the discount led to repeat orders. |
| Threshold: share not back within 180 days | % | — | If more did not come back, the measure turns amber. Empty: no threshold, the color does not change. There is no single standard in Russia: enter your own. |
How it is calculated
| Step | Formula |
|---|---|
| First order | basket × (full-price margin − discount) |
| Customer over a period | first order + repeat orders × basket × repeat margin |
| CAC | acquisition spend ÷ number of new customers |
| Cohort profit | customers × (margin over 180 days − CAC) |
Worked numbers
Default values from the app; in the interactive version you enter your own.
- Margin of the first order
- 1,750 ₽Basket × (margin − discount). The loss on the discount: basket × discount.
- Loss on the discount per customer
- 1,000 ₽
- Customer margin over 90 days
- 2,245 ₽
- Customer margin over 180 days
- 2,575 ₽
- Customer acquisition cost (CAC)
- 2,000 ₽
- Customer payback period
- 90 days · within 90 days
- Cohort profit over 180 days
- 575,000 ₽
- Did not come back within 180 days
- 75 %
- Without a discount (control): profit over 180 days
- 695,000 ₽ · no discount better by 120,000 ₽
Data as a table
| Measure | Cohort profit |
|---|---|
| With the discount | 575k |
| No discount (control) | 695k |
Customer: with and without a discount
| Case | First order | Over 180 days | CAC | Cohort profit |
|---|---|---|---|---|
| With a discount | 1,750 ₽ | 2,575 ₽ | 2,000 ₽ | 575,000 ₽ |
| Without (control) | 2,750 ₽ | 3,850 ₽ | 2,857 ₽ | 695,000 ₽ |
Without a discount the same spend gives more profit: the discount brings more customers but less valuable ones. For established customers a deep discount does not raise future purchases (Anderson and Simester, 2004).
Abuse through several accounts and “bargain hunters” pull repeat orders down: count the cohort on real data, not expectation.
Nuances and limits
- Without a control you cannot say whether the customer would have come without the discount.
- Several accounts and “bargain hunters” pull repeat orders down.
- A deep discount may pay for new customers but not for established ones (Anderson and Simester, 2004).