Retention and lifetime value
Why what happens after the first order decides how much you are allowed to pay for it.
Acquisition cost is only meaningful next to what a customer is worth. Two brands with identical ad accounts can have completely different ceilings, and the difference is usually built after checkout rather than before it.
The ceiling is set by repeat purchase
If your customers buy once, you can only pay a fraction of first-order margin to acquire them. If they buy four times, you can outbid that competitor on every impression and still make more money. Retention is not a separate department from media buying; it is the input that decides the bid.
| Repeat behaviour | What it allows |
|---|---|
| One purchase, no repeat | Must be profitable on order one |
| Predictable repeat, long gap | Can break even on order one, profit later |
| Subscription or consumable | Can lose money on acquisition deliberately |
Measuring it without fooling yourself
Two traps are common:
- Averaging across all time. Old customers inflate lifetime value because they have had longer to buy again. Measure by cohort — customers acquired in the same month — and compare cohorts at the same age.
- Using revenue instead of margin. Lifetime value that ignores cost of goods, shipping, returns, and payment fees justifies a payback period you cannot fund.
A practical working number is contribution margin per customer at 60 or 90 days, by cohort. It is short enough to act on and long enough to include the second order.
Payback period
Payback is how long until a customer has returned what you paid to acquire them. It is a cash constraint, not an accounting one: a 12-month payback can be correct strategically and still bankrupt a business that cannot fund the gap. Set the acquisition target from the payback period you can actually finance, then scale against it. See unit economics.
What actually moves retention
| Lever | Effect |
|---|---|
| The first delivery experience | Sets whether there is a second order at all |
| Post-purchase sequence | Converts a buyer into a repeat buyer at near-zero media cost |
| Replenishment timing | Reaching people when the product runs out, not on a fixed schedule |
| Product range | A second thing to sell to people who already trust you |
| Win-back | Cheaper than acquisition, and the list is already yours |
The feedback into acquisition
Retention data also tells you who to acquire. Cohorts differ: some acquisition angles, offers, and discounts bring customers who never return. An offer that wins on first-order cost and loses on repeat rate is a loss disguised as a win — which is why deep discounts often make the account look better and the business worse.
Before raising budget, check whether repeat rate is stable. Scaling an account whose customers do not come back multiplies the acquisition cost problem instead of the revenue.