Creative testing methology

Curious how everyone is structuring their creative testing these days for subs apps.

-Which GEOs do you use for concept testing? Do you test in Tier 1 directly or use proxy countries?
-How do you structure campaigns/ad sets? (One concept per ad set? ABO vs CBO?)
-What metrics and thresholds determine whether a creative moves forward?
-Do winners get promoted into BAU/scaling campaigns, or do you scale them within the testing campaign?

I do testing in tier1 countries but with different structures (depending on budget, type of product and audience diversification). The main metric is always the cost per the event you optimize towards but I look at engagement metrics to detect ad fatigue and also learn from winners and losers.

I've been lately keeping all assets within the same ad group (divided by concept) rather than using the usual testing/BAU approach where you move winners to BAU

Hi @davamosk, how does an ad group behave when you keep all the creatives within the same ad group? Does a new winning creative automatically start taking most of the impressions, or do you need to turn off the older, fatigued ads first?

I tried this approach once, but the ad group basically stopped spending once I turned off the fatigued creative. I ended up having to create a new campaign with the winning creative.

That’s why I’d love to understand why you’ve started preferring this structure over the more traditional testing + BAU setup.
Also, if you don’t mind sharing, I’d love to hear how you approach scaling with this new testing structure.

If a new creative is good enough, it takes the first positions quickly so you don't even need to turn off older creatives but you need a lot of volume as this winning assets take 24-48h max in starting to absorb spend so if they don't, it means you will still depend on older performers.

This has been my most used structure for small and medium budgets lately when I have good creative capacity. I don't like the testing vs BAU approach that much because a creative can perform very different when you give it $100 vs $1000/day.

Okay, so once you find a new winning creative, do you usually scale it vertically by increasing the budget of the ad set that contains it, or do you duplicate that campaign/ad set into a new campaign with a relatively low daily budget?

Also, how would you approach this at a much higher spend level, say $10k+ per day?

From my own experience, I’ve been scaling by duplicating winning setups into multiple lower-budget campaigns, usually around $300-500/day each, $1000+ a day kills the winner really fast. But over time, I feel like this creates too many campaigns to manage, especially as creatives start fatiguing and need to be replaced more frequently.

For higher spend levels, do you think it’s better to keep managing a larger number of low-budget campaigns and continuously add new winning ad sets from the creative testing campaign, or consolidate spend into fewer campaigns with higher budgets per ad set?

when I find a winning, I just increase the budget where the winner is and if that winner absorbs the budget, I move the others to a new ad group, but never the winner.

For higher spend levels, I normally see Meta automatically replacing winners over time once I start uploading other potential winners although sometimes I have to force the redistribution of spend and pause the winners that got fatigue faster. I have found success on consolidating campaigns with fewer ad groups only when I have enough creative capacity to rotate 2-3 times a week with at least 5-10 new concepts on each rotation