Scaling a Non-Gaming Subscription App on iOS: What is the budget and patience threshold for Meta AEM?

roasd365

Active Member
We are currently in the process of scaling a non-gaming subscription-based app on iOS using Meta Ads. I have AEM, but I’m hitting some "classic" crossroads regarding budget allocation and patience. I’d love to get some feedback from those who have successfully scaled in the current 2026 environment.

I’ve just started a campaign optimizing for Purchase(One Time Purchases+Subs), but the Day 1 and Day 2 numbers are coming in at 3x–4x my target CPA. Which means I can't feed the algo with the necessary daily events.

For those scaling successfully right now, do you tolerate high CPAs in the first few days of a Purchase campaign, or do you prefer starting with a upper funnel event campaign and switching to Purchase later?
 
If you don't have enough budget to generate 10-15 events per day at least, I'd wait and work in improving the conversion before going upper in the funnel (unless you have an event which is very correlated to the purchase)

(I'm assuming the purchase is happening in day zero for you)
 
If you don't have enough budget to generate 10-15 events per day at least, I'd wait and work in improving the conversion before going upper in the funnel (unless you have an event which is very correlated to the purchase)

(I'm assuming the purchase is happening in day zero for you)
Thanks for the reply. Roughly how many events are needed for performance to stabilize? Do you have a particular benchmark like 10 events per day for at least 3 or 7 days?
 
Hey guy, now you mentioned it. It is pretty hard to collect 10 purchases a day, you will need a huge daily budget, which is quite costly to run a campaign, especially when you are in the first phase of exploring Meta.

From what I’ve seen, running a Purchase campaign with a low budget in the first few days often leads to high CPAs, unstable performance, and not enough data to properly evaluate, let alone optimize.

Is there any other way to approach testing on Meta or increasing budget to a point that it would enable at least 10 purchases a days is the only way to go?
 
Hey guy, now you mentioned it. It is pretty hard to collect 10 purchases a day, you will need a huge daily budget, which is quite costly to run a campaign, especially when you are in the first phase of exploring Meta.

From what I’ve seen, running a Purchase campaign with a low budget in the first few days often leads to high CPAs, unstable performance, and not enough data to properly evaluate, let alone optimize.

Is there any other way to approach testing on Meta or increasing budget to a point that it would enable at least 10 purchases a days is the only way to go?
try with a proxy event which is very correlated to the purchase - i got success some times with that approach in small accounts but it will mainly depend on how good that proxy event is built
 
Hey guy, now you mentioned it. It is pretty hard to collect 10 purchases a day, you will need a huge daily budget, which is quite costly to run a campaign, especially when you are in the first phase of exploring Meta.

From what I’ve seen, running a Purchase campaign with a low budget in the first few days often leads to high CPAs, unstable performance, and not enough data to properly evaluate, let alone optimize.

Is there any other way to approach testing on Meta or increasing budget to a point that it would enable at least 10 purchases a days is the only way to go?
Would go for a upper funnel event like "registered" or "reach lvl 2" stuff like that. That makes it easier to get more events per day with less budget.
 
10 per day is the standard yes!
Do you have a POV on using a 7 day click attribution window in the ad set vs. a 1 day click attribution window to achieve the 10 results per day? Assuming a 1 day click will drive the algorithm to find more users with higher intent, but a 7 day click could help with the learning phase, especially if budget is low (just assumptions about the algorithm, though). Annual subscription app product with a 7 day free trial for context.
 
Last edited:
Do you have a POV on using a 7 day click attribution window in the ad set vs. a 1 day click attribution window to achieve the 10 results per day? Assuming a 1 day click will drive the algorithm to find more users with higher intent, but a 7 day click could help with the learning phase, especially if budget is low (just assumptions about the algorithm, though). Annual subscription app product with a 7 day free trial for context.
I normally use 7d windows with AEO campaigns because otherwise the performance fluctuates a lot. In your case, I would focus a lot of refining the trial_qualified event
 
try with a proxy event which is very correlated to the purchase - i got success some times with that approach in small accounts but it will mainly depend on how good that proxy event is built
I tried with the proxy event but there are some custom events after I run some ad campaigns to gain some traffic that it shows up at the adset set up, some dont. And i have no idea how this happens. Is it necessary to set up a custom conversion to the custom event that didn't show up David? I'm exploring w2a btw.
For example, i created a custom event called QuizEngaged (users completed Q4) and it didnt show up here (pic 1). But another event called BetterMeIntent showed up... and the QuizEngaged is more top funnel, added later than the BetterMeIntent ~3days.
 

Attachments

  • 1782101665955.png
    1782101665955.png
    45.5 KB · Views: 5
Did you review the events in the events manager? You have to accept them before they're available to use
This is what I view in the events manager. The QuizEngaged event showed up here but still not in the adgroup set up and I still cant figure it out why..
 

Attachments

  • 1782182422199.png
    1782182422199.png
    74.6 KB · Views: 4
This is what I view in the events manager. The QuizEngaged event showed up here but still not in the adgroup set up and I still cant figure it out why..
you need to go to the settings on the events manager, then event review, then confirm the events. They will appear as available to target after that
 
Back
Top