See More → View Navigation Close Facebook Twitter LinkedIn Quote
Main Content

October 31, 2017

In 2017, 84% of app marketers measure performance based on revenue instead of install volume.

In line with that shift to direct-response advertising, Facebook launched App Event Optimization (AEO) last summer. Since then, we’ve implemented AEO campaigns for several of our clients across industries and are ready to share some of our findings.

But first, a bit about AEO…

 

Facebook App Event Optimization 101
App event optimization allows advertisers to target people who are most likely to take a specific action within an app based on behaviors exhibited across other apps.
AEO goes further than mobile app install campaigns, driving real business value, whether through in-app purchases, viewing content such as in-app ads, or revenue-producing activities within games. Specifically, advertisers have 14 standard app events for which they can optimize…
We’ve worked closely with our clients to implement AEO strategies, deciding which events to target, testing them against MAI, and more. In addition to weighing in on some of Facebook’s best practices, here are our learnings so far.

 

1. Volume Matters

At the highest level, the ideal app event to optimize for should be one that is important to your business. Sometimes, however, your revenue-generating event might not work for AEO. To give the AEO algorithm enough data to show meaningful results and cost improvements, you should choose events that meet the following criteria…
Note: It’s important that the event occurs less than 50% of the time or you might not see significant differences from your other campaigns. Also, keep in mind that it’s important to give your ad sets ample time to gather data; usually two weeks should do it.

If your high business value event (i.e., purchase) doesn’t meet those criteria, what should you do?

 

2. Take a Step Back

Often times (especially for small or new apps), high business value events are too far downstream in a buyer’s journey, meaning they don’t occur often enough.

In that event, how do you choose what alternative event to optimize for? For linear app funnels, simply choose the conversion step before your revenue-generating conversion event.

For example, if your business cares about purchases, but there aren’t enough weekly purchases, you will likely see more conversion volume if you take a step back and optimize for ‘initiated checkout,’ or even ‘added to cart.’ For some apps, however, this might not be so straightforward…

 

3. Get Creative, and Always be Testing

For many businesses like gaming apps or publishers, there might not be an event that is simply a step back in the funnel. In that instance, you’ll want to get creative and look inward at your historical user behavior data. For nonlinear user journeys, such as those within gaming apps, we suggest doing behavioral analytics research to identify strongly correlated behaviors.

 


 For example, if your business profits from users clicking in-app ads, look at historical data to see what other behaviors those high LTV users do, whether that’s searching for content or completing tutorials.

When in doubt, start testing. Facebook’s Ad Study feature, enabled through their developer API, is a great way to A/B test identical ad campaigns with different bid types or optimized events. Here are a few Ad Study best practices…

  • Only test one variable at a time
  • Use only new campaigns in order to have an accurate comparison
  • Allow enough time (1-2 weeks minimum) to collect enough data for statistical relevance
  • Try to avoid making changes while the Ad Study is running. If you do make changes, make the same changes to both campaigns.

With those results, analyze downstream data to inform the allocation of budget across each event. You can also apply Ad Studies to compare MAI campaigns to AEO campaigns.

 

4. Try AEO with ‘Super-combined’ Audiences

It goes without saying that reaching users with higher lifetime value (LTV) is harder to do. Remember, only 6% of users come back to the app after the first week.

So, for AEO campaigns to scale while maintaining (or ideally driving down) cost per acquisition (CPA), casting a wide net is crucial. Facebook recommends that in order for app event optimization to really work, audience sizes should be at least 1 million people. We’ve found that bigger is better; shoot for audiences of at least 5 million, especially when running Ad Studies.

We also find that AEO is a great way to target audiences that have already shown promise in app install campaigns. To increase learnings as fast as possible, try combining high-performing audiences, whether they’re lookalikes or built off of keywords, into single, “super-combined” audiences. You should consider running ‘super-combined’ audiences especially when there’s 20% or more audience overlap.

Once you have a large, high-quality audience and an idea of what events to optimize for, you’re ready to go!

 

5. Scale up incrementally

If you’ve never run app ads, however, you probably shouldn’t jump straight into AEO, as Facebook won’t have enough data to start spending.

One way we try to avoid ‘cold start’ issues, is by launching our super-combined audiences with install ads first. Once we’ve collected significant data through goal completions (installs), we then transition the ad sets to AEO. We aim for achieving at least 150 goals before graduating from MAI to AEO. We’ve also found success applying the same strategy to testing events further downstream in the conversion funnel.

Through this tactic, we helped one of our clients, a media streaming app, drive CPA down 33% while maintaining acquisition volume goals.


AEO is fundamentally changing how growth professionals think about user acquisition, prioritizing lifetime value and recalibrating the KPIs that they track. Since AEO launched, however, many advertisers held misconceptions that it only works in specific verticals such as gaming.

While we’re still in early days of testing App Event Optimization, we’ve seen promising results across many other verticals; for a leading fitness app, a media streaming app, a news app, and an e-commerce app to name a few.

The bottom line is… if you’re marketing an app, there’s no reason you shouldn’t try AEO. By approaching it strategically and constantly analyzing downstream costs and performance, you should have a solid idea of how it’s working for your business.

Interested in learning how we can help optimize your Facebook ad campaigns? Check out our Facebook advertising services and expertise >

About Bamboo

We design and execute highly scalable advertising programs for trailblazing brands.

More About Us