💎 Growth Gems #36 - Gems from "How to Win with Facebook AAA (Automated App Ads) Campaigns"
Hi there,
Today's gems are mined ⛏️ from a Mobile User Acquisition Show podcast episode with Gina Kwong (Electronic Arts):
How to Win with Facebook AAA (Automated App Ads) Campaigns
💎 #1
Initially it was a challenge to optimize creatives because you could not get the creative data through your MMP or an API. Gina downloads the data from the UI and optimizes base on the relative performance (i.e which creatives get impressions).
💎 #2
While it's nice to have the possibility to have 50 creatives, it seems it's better to only have 20-25 creatives and refresh the ones that don't get spend after one or two weeks.
💎 #3
Facebook AAA is the fastest way to test creatives and very effective because it goes through a high amount of creatives and can tell you very quickly what will be favored.
💎 #4
If you add a new creative to a AAA campaign that already has winning creative(s) and you see it get some mid-range spend it means that it's performing well and a potential winner as well. Because Facebook will always keep serving the champion creative.
💎 #5
Gina recommends not to run AAA only so you can mitigate risks (included risks due to having only 1 campaign for geo-language-optimization combination). What you can do is use a combination of LAL and AAA campaigns.
💎 #6
If you're running campaigns in smaller countries/regions that have similar performance, "country-clustering" can be very effective.
💎 #7
For high cost products you can also test optimizing AAA towards an event higher in the funnel were you have more conversions (vs. purchase) so the algorithm can learn. Or run lookalike campaigns first to know what works.
💎 #8
Gina's recommendation is to put the creatives in AAA (vs. MAI), and to optimize with AEO to test creatives against what you're actually optimizing for. You can also test AAA with the new MAI + purchase optimization.
💎 #9
Group campaigns together by LTV tiers: tier 1/tier 2/tier 3 countries. Then review which countries bring you the best performance and regroup differently if needed.
💎 #10
Facebook says they exclude 90 days installers but it's not always the case. You may be still showing your ads to people that have it already but this can happen on any channel. But if these people don't convert they will stop showing them the ads.
💎 #11
Gina always starts campaigns with a few hundred dollars only. After $1k or $2k if it's not giving results then you know something is not working (creative? geo?).
💎 #12
Start with one path that you think will work, and if it does work then scale it by 20% increments. Then you can expand to different optimization types.
Stay savvy!
⛏️ Sylvain