Win-back ROI: which lapsed customers are worth a postcard
Win-back mail earns its keep on a simple asymmetry: the card costs the same $1.50 for every customer, but the customers are worth wildly different amounts. The whole ROI game is aiming the fixed cost at the variable prize — which is why the central tool here is a lifetime-value threshold, and the math fits on an index card.
The two-step math
Expected value of a card = response rate × expected next-order profit.
Mail when that exceeds the card’s cost — and since higher-LTV customers respond more and order bigger, the rule collapses to: mail lapsed customers above a lifetime-spend bar.
Worked example: a 4% response and a $60 average next order at 60% margin gives an expected $1.44 of profit per card — breakeven against $1.50. Move to your $100+ lifetime-spend customers, whose next orders run larger and whose response runs warmer, and the same card is comfortably profitable. That’s the threshold logic in one paragraph.
ROI calculator
Breakeven response rate: 3.33% — below that, this mailing loses money. Retention mail to your own customers typically responds far above cold-list rates, which is the whole argument for it.
Reading the calculator honestly
- Cards mailed — your lapsed pool above the LTV bar, not the whole quiet list. The bar is doing the heavy lifting.
- Response rate — assume less than you hope; these customers ignored email first. Warmer segments (recent lapses, high spend) sit at the top of the range (catching lapses early helps here more than anything).
- Average order — use the segment’s typical order, margin-adjusted for the truthful version. High-LTV customers return with high-LTV orders.
Three levers, in order of power
- The threshold. Raising the LTV bar lifts response and order size simultaneously — the rare lever that improves both sides of the fraction.
- Freshness. A customer mailed one week past their lapse point beats one mailed three months later, at identical cost. Continuous detection is an ROI feature.
- The offer. Last, deliberately — it works, but it spends margin the first two levers don’t (offer strategy).
Then audit the answer
Response-rate ROI still overcounts, because some mailed customers were returning anyway. The number that should set next quarter’s threshold is incremental returns — holdout-measured lift — per dollar of card spend. It runs smaller than the attributed number and it’s the one that’s true.
Common questions
›What response rate should I plan for on win-back cards?
Conservative low single digits — these are, by definition, customers who already ignored two emails. The math still clears comfortably because the card only mails to high-LTV customers, where one response funds many quiet ones. Replace the assumption with your measured rate after the first cohorts.
›Should I value a win-back at one order or at future orders too?
The next order is the number to run the ROI on — it is real and near. But remember what you are actually buying: a resumed rhythm. If your returned customers go on to order several more times, the true value per response is a multiple of the first order, which justifies a lower LTV bar than the single-order math suggests.
›Why gate on lifetime value instead of mailing every lapsed customer?
Because the card costs real money against very different prizes. The one-time $19 buyer and the $400 regular are both "lapsed," but only one of them is worth $1.50 of postage on the odds. The LTV threshold encodes that judgment once, and the program applies it automatically.