Limit-Ad-Tracking in iOS 10Apple’s Limit-Ad-Tracking feature was rolled out with iOS 6 and allows users to opt out of cross-app-marketing by randomizing their device’s IDFA, by app. This was designed to preserve and protect a user’s desire for anonymity, while still allowing attribution to happen on an app-by-app basis. In this spirit Kochava has been a leader in security and privacy throughout the platform, including syndicating the Limit-Ad-Tracking flag as set by each user.
With the release of iOS 10, Apple has further matured the Limit-Ad-Tracking feature to zero out the IDFA instead of randomizing it. This added level of abstraction ensures that bad actors will have a harder time cross-marketing to users who have opted out. Because of this, an estimated 15%-20% of devices will include identical IDFAs (all zeros), which would, by default, cause any installs to be deduplicated. Kochava continues to respect both the Limit-Ad-Tracking flag and the spirit in which it was implemented by syndicating the flag to all partners, while creating a way for attribution to happen on an app-by-app basis, as originally intended and supported by Apple.
So What?A unique device ID is required for device-based reconciliation to work. Because devices with Limit-Ad-Tracking enabled will still have a device ID present, the waterfall will never move on to fingerprint-based matching. Moreover, since all devices with Limit-Ad-Tracking enabled have the same IDFA (all zeroes) they will be deduplicated based on the standard attribution waterfall.
The GistKochava has the most flexible and powerful attribution engine. The logic has already been updated to recognize IDFAs comprised exclusively of zeroes and fall back to fingerprint-based matching. The solution accounts for the ~15%-20% of devices with Limit-Ad-Tracking enabled (which causes the IDFA to be all zeroes) while maintaining the highest level of integrity and configurability. This is another example of how the investment Kochava has made into a modular and powerful infrastructure gives its customers an unfair advantage over companies using other attribution solutions.
- Scenario 1: If a user is not running iOS 10, device reconciliation can be performed and subsequent waterfall attribution methods happen as necessary.
- Scenario 2: If a user is running iOS 10 and “Limit-Ad-Tracking” is not turned on, device reconciliation can be performed and subsequent waterfall attribution methods happen as necessary.
- Scenario 3: If a user is running iOS 10 and “Limit-Ad-Tracking” is turned on (expected to be ~15%-20%), device reconciliation can’t be performed and subsequent waterfall attribution methods like fingerprinting will be used for attribution.
The SolutionAs Kochava customers begin to loosen reconciliation settings to allow for fingerprint matches, it is important to take advantage of other features offered within Kochava.
- Fingerprint Equalizer: The Fingerprint Equalizer defines a length of time during which Kochava will consider a fingerprint match at the same level of attribution as a device-based match. Since we know that your iOS app is expected to have ~15%-20% of IDFA’s returning zeroes, you can work with your trusted networks. This will help you maintain your expected ratio of attributed installs to organic installs by allowing fingerprint matching for a specified timeframe.
- Traffic Verification: Traffic Verification determines if the real-time traffic for the specified tracker is valid based on your custom settings in Kochava. This also allows you to perform actions on invalid traffic. With more installs being reconciled further down the attribution waterfall, the integrity of the match will be impacted. This will put even more burden on advertisers to ensure the attributed installs they are paying for are verified in real-time.
- Limit-Ad-Tracking Fallback: Within the Kochava Platform, Default Reconciliation Settings may be configured to automatically fall back to Fingerprint attribution in cases where Limit-Ad-Tracking is enabled and the device ID is all zeros.
Learn more about Configurable Attribution