Knowledge base
When should I stop letting Meta's AI control my budget?
Once Advantage+ has generated a few weeks of real placement data, that is your cue to review, not necessarily to pull back.
Not automatically at any fixed date. The useful trigger is data, not time: once Advantage+ has run long enough to generate a real read on which placements (Feed, Reels, Stories, Audience Network) are actually converting, at least two to three weeks at a meaningful daily budget, you have enough to make an informed decision instead of a guess.
At that point there are three reasonable outcomes, not just one. You might leave it fully automated because the results are strong across the board. You might exclude one or two placements that are clearly underperforming while leaving the rest automated. Or you might pull the best-performing segment into a manual campaign you control more tightly. "Take the reins" does not mean turning automation off, it means deciding on purpose instead of by default.
The tell that it is time to look: cost per result rising for two straight weeks with no change on your end. That is the platform's own signal that something in the automated mix has shifted, and it is worth a manual look regardless of what triggered it.