CASE STUDY | USER ACQUISITION
Playrix runs non-personalized UA campaign with Liftoff, CPAs decrease
CPIs, CPA and D7 ROAS on par or better than personalized traffic
Preparing for iOS14With Apple announcing upcoming changes to IDFA, Playrix turned to Liftoff to run campaigns promoting all their casual games, targeting iOS users with Limited Ad Tracking (LAT) enabled on their device. The purpose of the on-going campaigns is two-fold:
- Understand the performance impact of bidding on “non-personalized” zero IDFA traffic including CPI, CPA (in-app purchase) and Day 7 ROAS
- Understand the spend potential on iOS of non-personalized LAT traffic compared to “personalized” traffic
Playrix Limited Ad Tracking Campaign SetupTo evaluate the performance of non-personalized traffic ahead of the release of iOS14, Liftoff added a segment of LAT users to existing Playrix campaigns to enable bidding against this audience. All other aspects of the campaign remained the same including the setup, KPIs, ad creatives, regions, and ML models.
Because Liftoff systems were built to not solely rely on device IDs, adapting our bidding strategy to accommodate traffic without device IDs was straight-forward. For attribution purposes, we relied on the MMP to accurately track installs and post-install events.
Performance was Strong Right From the StartNon-personalized traffic performed well from the get-go. CPIs were lower across every major region. CPA (cost per in-app purchase), though initially higher compared to IDFA traffic, eventually decreased to less than personalized traffic over the first few weeks, suggesting that non-personalized traffic has the potential to outperform personalized traffic.
D7 ROAS is on par or better than personalized traffic in most cases.
Scale wasn’t an issue as non-personalized traffic quickly ramped up to become a very significant percentage of Playrix’s overall iOS spend with Liftoff.
Because of how the Liftoff ML platform was designed, setting up a campaign targeting non-personalized users did not require any major overhaul to a standard IDFA-targeted campaign. By utilizing the complete Liftoff creative and technology stack, including Liftoff ML bidding, Smart Pacer and Dynamic Ads, buying and optimizing non-personalized traffic was seamless, and gave Playrix a performance edge against other buyers who are not able to buy non-personalized traffic with as much optimization. Liftoff ML was able to more accurately value and optimize non-personalized inventory, delivering campaign performance on par or better than personalized traffic.
In addition, we achieved significant scale targeting non-personalized traffic as evidenced by the percentage of spend we were able to allocate towards the campaign, and the plan is to continue scaling-up non-personalized traffic. Furthermore, post iOS 14, we expect the majority of iOS traffic to be non-personalized hence Liftoff is preparing to ensure that Playrix and our other customers are ready to embrace Apple’s new data privacy rules related to IDFA.