Meta ads testing on low budget
HI there! Wanted to get thoughts on my current testing framework and if it should be changed.
I test multiple digital products ($9 price, trying to get ~$15 AOV by improving upsells/downsells).
For each product, I allocate ~$150 budget for testing.
Cold pixel. Purchase/sales objective. 1 Campaign each product, 1 adset, 8-12 ads. Full broad, no targeting restrictions (let andromeda do the targeting)
Each $150 test: 8-12 creatives. Completely different formats/pain points/target personas - all targeted via the creative so andromeda can go after various specific targetgroups).
All 8-12 creatives into one adset. Auto stop-pause ad via meta MCP if no sales after $15 spent and $1+ CPC (so the budget gets allocated to other ads in the adset) & stop-pause if CPC $2+ after $10 spent (means CTR or CPM or both too high, ad not sustainable for my AOV).
Question is: does this framework make sense? How much should I trust meta regarding the specific ad it pushes? Should I remove the stop-pause rules and just fully trust meta with its creative choice, or try to implement the rules so that all ads get a chance at spending?
I've also noticed some creatives getting a couple purchases on first $5 and then spending $20 more and getting none. What does this signal?
I understand that for this kind of budget, testing is hard, since there's not that many sales to statistically judge the performance of the ads. But what kind of early signals (besides purchases) could show that the ad has potential and should be let to run for longer, even if it doesn't it gets 0.2-0.5 early ROAS? Is high CTR + low CPM (resulting in cheap CPC website clicks) a good signal, if the campaign has sales/purchase objective, or can the cheap CPC be misleading?