This is the subhead for the blog post
Have you struggled to improve post-install performance on your mobile app campaigns? Facebook now has the answer! In July they released a tool, called App Event Optimization, which allows you to optimize further down the funnel and therefore improve your long-term goals for app users.
App Event Optimization has a lot of similarities to oCPM bidding from website conversions campaigns, which may make the feature a little easier to understand and manage. If you have the Facebook SDK set up, you can transfer this type of bidding to your mobile app install campaigns by simply choosing to optimize towards “App Event” and choosing the conversion event you would like to optimize for.
But that leads us to a question: What event should you optimize towards? I had a chance to conduct a test comparing the performance of three different optimizations, from the beginning of the funnel with the basic install to the end of the funnel with app purchases, and I was able to find which works best in certain situations and why.
The data below shows results from six days of optimizing for install; add to cart; and purchase:
When optimizing to install, you may have sufficient volume in spend, clicks, and installs, but the lack of quality of that volume becomes more apparent further down the funnel. As you can see, when I optimized towards install, the purchase CPA was high, so the main goal was to improve those numbers – and App Event Optimization was the key. It does appear that the easy fix would be to just optimize directly for purchases, but there are a few items you need to take into account first.
Optimizing for the end of the funnel can give you decent results with a strong end of funnel CVR; however, it can also may have a low ability to scale. When optimizing to a certain event, the Facebook algorithm requires at least 20 of those conversions a day in order to identify and target the people likely to convert. Since my test had under 20 purchases a day, Facebook could not efficiently identify the people who were likely to convert, and since they could not determine who was best to target they began to limit spend. Iif your ad set can push more than 20 conversions a day at a consistent rate, optimizing for the end of the funnel is a great place to start – but if not, you have to optimize earlier in the funnel, which was my next step.
The event right before purchasing was add to cart, and for a mid-volume account it represented the sweet spot where we could both receive an adequate amount of spend and realize strong returns. We easily hit 20 add-to-cart conversions a day, which allowed the Facebook algorithm to function properly and enabled strong performance throughout our flow. Overall, when optimizing towards add to cart, we were able to decrease our CPA by over 30% from optimizing to install while also increasing the volume (a capability that optimizing towards purchases lacked).
App Event Optimizations is a great tool to help you drive the right people to your app and achieve strong long-term effects, but before you can fully benefit from it you just have to determine which event should be your target for optimization.