What counts as “recovered revenue”? Honest attribution, explained

Here is the uncomfortable fact under every recovery dashboard: abandoned carts complete on their own all the time. The shopper was always coming back after payday; the card declined and they retried at lunch. A tool that claims every completion that follows its email is taking credit for gravity. Understanding how honest counting works protects you from paying — in subscription or in trust — for recoveries that were never at risk.

Attribution: the day-to-day speedometer

Practical recovery attribution is a matching discipline. A completion counts as recovered only when it matches the same shopper and the same cart, within a bounded window of a recovery touch. Tighter still: a completion that arrives before any touch went out is an honest zero — the program did nothing and should say so. These rules cost the dashboard some flattering numbers, which is precisely what makes the remaining numbers usable.

The baseline problem attribution can’t solve

Even tight matching can’t answer the deepest question: would this cart have completed anyway? Some would. The clean answer is the holdout: exclude a random slice of qualifying carts from all recovery touches, then compare completion rates. If touched carts complete at 18% and held-out carts at 11%, your ladder caused seven points — the rest was coming back regardless. Randomness is what makes the comparison legitimate: the two groups differ only in treatment, so the gap has one explanation.

Reading recovery numbers without self-deception

• Expect a baseline. If a dashboard implies zero organic completions, it is measuring generosity, not recovery.
• Windows should fit the behavior. Recovery is a days-scale event (the intent curve); a 45-day recovery window would be claiming next month’s ordinary shopping.
• Small numbers lie. Ten carts per group can produce any ratio by luck. Let cohorts reach three digits before believing a lift figure.
• Per-rung truth matters most. The email rungs are free — they only need to beat zero. The postcard costs money, so its incremental rate is what should feed the threshold math, not its attributed rate.

Why honesty is the profitable setting

Inflated recovery numbers aren’t just vanity — they misprice decisions. They justify discounts that weren’t needed, thresholds set too low, and budgets defended with revenue that would have arrived anyway. Honest measurement runs the program on real margins. It reports smaller numbers and makes you more money — the trade every serious operator takes.