Log in

SKAN LTV Prediction

According to Apple's policy, you can get the first ad data on iOS 14.5+ only with explicit user permission or using SKAdNetwork (SKAN)

To protect users' privacy, Apple sends SKAN postbacks with some delay, does not give any information about the install time and the device advertising ID. These restrictions prevent the ability to link traffic sources to user activity, assess incoming traffic, and determine the LTV of engaged users.

MyTracker allows you to build a forecast on SKAdNetwork data, to predict the LTV of the engaged iOS audience, and to evaluate the campaign success while maintaining user privacy.

How it works

The LTV prediction based on SKAdNetwork data is available only for iOS applications. The forecast is based on a probabilistic machine learning model. The model analyzes aggregate app data:

  • the number of SKAN conversions and the conversion value, transferred directly from Apple or ad networks;
  • data on the devices whose users permitted tracking: installs, sessions, launches, etc.
  • data on user behavior within the app: payments, launches, sessions.

The model analyzes the pattern of LTV distribution among ad campaigns, identifies probable organics, and builds an LTV forecast for SKAN conversions for 1 and 6 months ahead. You can view forecast by individual partners and ad campaigns, but deeper data detail (for example, by date and country) is not possible due to Apple's data tracking restrictions.

The prediction mechanism follows Apple's privacy policy. The model does not identify the device or the user and operates at the campaign level.

How to use

  1. Check that the MyTracker SDK has been integrated into your app.
  2. Make sure that you support SKAdNetwork attribution: ad networks are registered with Apple and added to the MyTracker's SKAN partner list Learn more
  3. In the Builder, select a report period. The forecast will be made for campaigns that have installs or SKAN conversion for the selected period.
  4. Note that Apple transmits SKAN conversions with some delay (at least 24 hours) and a conversion threshold of 128. And installs, which can also be necessary for prediction, can only be tracked with the user's explicit permission. For details, refer to the iOS & SKAdNetwork section

  5. Add SKAN LTV Prediction metrics to your report. LTV can be forecast for periods for 1 and 6 months after the first conversion is received.
  6. Add additional dimensions and metrics to evaluate the effectiveness of your ad campaigns, with both predictive and actual data. Add Partner and Campaign dimensions and Installs and SKAN conversions metrics.
  7. Due to Apple's iOS 14+ data tracking restrictions, detailed data by dates, countries, and other dimensions are impossible. For details, refer to the iOS & SKAdNetwork section.

  8. Click Calculate. Metrics SKAN LTV Prediction will show expected profits from iOS 14+ ad campaigns one month and six months ahead.

You can get reports with LTV forecast using Export API

How to improve prediction

The LTV prediction will be more accurate if MyTracker receives reliable information about the SKAN conversion values.

In SKAN postbacks, the conversion value is not encrypted and open to manipulation. To protect your data and verify postbacks from ad networks, we recommend sending copies of SKAN postbacks from Apple to MyTracker:

  1. Make sure you support the Conversion Value update method — updatePostbackConversionValue. For details, refer to the Conversion value section.
  2. Configure sending of direct SKAN signals to MyTracker. For details, refer to the SKAN postbacks from Apple section.
Was this article helpful?