Switching to 1-day-click erased 22% of 'reported' ROAS

Switching to 1-day-click erased 22% of 'reported' ROAS
Subscription box, 60d, $96k spend, same campaigns measured two ways.

Attribution window (how long after a click a sale still counts):
— 7-day-click / 1-day-view: 3.2x ROAS
— 1-day-click only: 2.5x ROAS ▼22%

Matched against the backend revenue from a holdout, true incremental ROAS was 2.4x — within rounding of the 1-day number. The 7-day view-through was claiming sales that would've happened anyway.

Read: pick your attribution window before you set ROAS targets, or you're optimizing toward a mirage.