Another approach is to run a short test with varied settings and compare outcomes it gives a clearer picture without fully committing to one method.
I usually turn this switch off when testing creative assets to ensure my ad campaign has only one variable, allowing me to observe changes in ad performance data.
When testing ads, I’d avoid turning on advanced switches at first and keep things simple so results are easier to understand.
I do the same. When I'm testing ad audiences and keywords of interest, I turn this switch off to maintain variable control throughout the entire campaign.
No. Keep it ON, especially during testing.
It helps Facebook learn faster and find results.
Turn it OFF only later, when scaling with proven audiences.
Your suggestions have given me some new ideas. I will only activate this feature after testing the materials, and then I will test the interest keywords, setting only one interest keyword per ad group to observe its effectiveness.
it can actually help to leave the switch on briefly sometimes the controlled delivery highlights what works best before scaling up.
I usually turn it off during the testing phase when I'm not scaling up, because I feel that having it enabled introduces too many variables into my ad testing phase, making it difficult to properly analyze the ad performance.
Yes. During the ad testing phase, I proactively turned off (or limited) this switch to avoid sudden spending spikes and keep test data stable. Once the campaign was stable and scaled, I turned it back on or adjusted it accordingly.
Your testing method is exactly the same as mine. During the testing phase, I turn off two switches, and this is one of them. Once I've gathered enough data and identified relevant keywords, I'll turn them all back on and increase the budget.
Generally not—you shouldn't intentionally turn it off during testing; keep advanced optimizations/enhancements enabled first to let the system learn, then only disable specific options later if they hurt control or performance.
I used two testing methods: one involved turning off the advertising during the testing phase, and the other involved keeping it on. However, I noticed that the budget consumed during the testing phase fluctuated unpredictably. I'm still observing the results.
Yes, during testing it often makes sense. That switch can restrict delivery before the algorithm learns anything useful, so turning it off early gives cleaner data.
That's absolutely correct, I've tested it. Turning it off for ad testing resulted in very stable data, which will be a great help for my analysis. The data is very clean. I'm currently testing the data analysis without turning it off.