Disagree on this one personally. While the algo uses those as references for what ad to serve, it's main priority is your optimization event (in most cases lowest cost conversion)
if you're not backing out on the campaign level over an extended period (7-14+ days) and it keeps spending on that ad then make some changes, but a big concern with this advice is that people see a high CPA or low ROAS on a historic winner and assume it's fatigued, pause it out then see worse account results over the long term despite 1-2 weeks of improvement.
Also in my opinion the BOF ad level data on meta is extremely faulty due to it being modeled, and making decisions on what to deliver budget to based on that isn't a great move. Plug the same account into triple whale and northbeam + look at it in ads manager and you'll walk away with three different stories of "what's working"
Had a case this year where I handed over an account, new manager turned off the historic top performers with high engagement and low ROAS, and saw better results for 1-2 weeks -- then nCAC spiked and revenue tanked. I'm certain meta does misallocate budget and there are plenty of cases where this action does help, but honestly I think it's much less necessary than media buyers want to believe.
How are you quantifying creative fatigue on these and determining when something needs to be paused out?
@max_blessen @bruno_BPW @max_blessen going to shoot you a DM, just had a case study with a client that went more native-style with this type of content for their ads and doubled revenue this month