As we move closer to the release of iOS14.5 and the likely implementation of Apple’s ATT framework we wanted to share our approach to solving some of the largest problems that you as app marketers will face over the coming weeks.
This change by Apple will significantly inhibit the ability to deterministically attribute installs to campaigns and ad networks. Where users don't give access to their IDFA through the ATT framework the "campaign" and "source" fields will be missing from the attribution files you receive from your MMP’s.
We wanted to share how AlgoLift is helping to solve this problem and can continue to predict the LTV performance of your campaigns and ad networks and adjust bids and budgets on campaigns to maximize for long-term returns.
AlgoLift user-level LTV predictions
There will be no interruption in the AlgoLift LTV prediction service and we can continue to provide LTV predictions for your iOS users if you’re sharing the attribution, revenue and engagement data against an advertiser generated user ID (eg a hashed IDFV or some other internal ID).
AlgoLift has built a campaign measurement solution for iOS that outputs a campaign pROAS schema. This will enable you to measure the long-term performance of their SKAdnetwork campaigns and ad networks and optimize accordingly. While this service will be available to you without any additional charge, enrollment in the service is not an automatic ‘opt-in’ as it will require you to provide supplemental data to the existing data pipelines for the LTV predictions. We intend on delivering the campaign pROAS estimates via s3, alongside the ongoing LTV deployments.
Our iOS14.5 campaign measurement solution requires two data sets:
- User-level app data: this is the data set we use to predict LTV. To use our campaign measurement solution we need you to start sending us conversion_value as a field, details on where to send that are in our documentation. The AlgoLift solution works regardless of the conversion_value definition but AlgoLift has built an algorithm to determine the optimal conversion_value for your app. Let us know if you’d like us to help you with this.
- SKAdnetwork postbacks: we request these postbacks through S3, details on how to send us this data is here. We’ve outlined how each of the major MMP’s is supporting exporting these postbacks below:
- Appsflyer - currently restricted to subscribers of their “data locker” product
- Kochava - scheduled report for all customers, custom query for enterprise customers
- Tenjin - not yet announced
AlgoLift UA Automation and cross-channel budgeting
AlgoLift will make bid/budget adjustments based on our iOS14.5 campaign measurement solution and will continue to operate our cross-channel monthly budgeting product to allocate the optimal budget at the channel level. However, we can continue to optimize campaign bids and budgets while SKAdnetwork postbacks are unavailable.
To do this we'll use modeled ROI from previous historical data, ongoing Android and iOS(<14.5) campaigns, iOS14.5 opt-in users as priors. We will continue to make changes to campaigns across Facebook, Google, Apple Search, Unity, Ironsource, Vungle and Applovin based on long-term predicted returns.
Let us know if you have any questions on the above and would like to jump on a call to discuss further. We look forward to getting you set up on our iOS14.5 campaign measurement solution and ensuring your success in this new environment.
The AlgoLift Team