Attribution window choice is one of the most underrated campaign settings. How getting it wrong kills working campaigns early or masks a fading one.
Attribution window choice is one of the most underrated settings at campaign launch, and getting it wrong either overstates or understates real results badly enough to drive genuinely bad scaling decisions.
A too-short window kills working campaigns
If an offer naturally has a longer conversion cycle — forex deposits, high-ticket e-commerce, subscription trials — a 1-day click window will systematically understate real conversions, making a profitable campaign look unprofitable before it's had a chance to show its real numbers.
A too-long window masks a fade
A 30-day window on an offer that genuinely converts fast (most nutra, most sweepstakes) can keep crediting conversions to a stale creative long after it's actually stopped working, delaying the point where you'd otherwise notice and fix a real decline.
Match the window to the offer's actual conversion behavior
Set window length to the offer's actual conversion behavior, not the platform's default. Pull the real time-to-conversion distribution from the tracker for that specific offer type before setting the attribution window — platform defaults are a generic starting point, not a deliberate choice for your specific offer.
Cross-device attribution is still underrated
A noticeable share of conversions from mobile-initiated campaigns complete on a different device — single-device attribution always understates real results to some degree, and that's worth factoring in whenever you're skeptical of a "disappointing" early CPA.
In short
- Offers with a long conversion cycle (forex, high-ticket, subscription trials) get understated by a short 1-day click window.
- A 30-day window on a fast-converting offer can keep crediting conversions to a stale, no-longer-working creative.
- Pull the real time-to-conversion distribution from the tracker, not the platform's default setting.
- Single-device attribution always understates results to some degree — factor that in when judging an early CPA.
If you're planning to enter this direction — run the economics before launch, not after the first thousand dollars spent.