"It paid off — I'm up 15% since switching"

"It paid off — I'm up 15% since switching"

Granted: a 15% steady-state lift is meaningful and might fully justify the move.

But nobody books the switching cost. The migration week of broken ads and zero fill, the 4 to 8 weeks of the new network's algorithm cold-starting on your inventory, the temporary CLS hit, the support hours. People remember month-six RPM and forget the trough that came first. The honest figure is annualized net, trough included.

🚩 Hidden variable: the migration J-curve. Earnings dip before they rise, and the dip is real money the 15% has to repay before it's profit.

Amortize the first-quarter loss against the lift. A 15% gain that took a 25% trough to reach hasn't paid off until month nine.

Not saying the switch was wrong — saying the comparison is rigged.