Apple Search Ads & SearchAdsHQ Amid IDFA Deprecation
Anastasia Sidoryk 
Apple’s IDFA deprecation is altering the ecosystem. Although the changes haven’t been implemented yet, mobile measurement partners, ad platforms and third-party mobile app marketing tools are taking steps to adapt themselves to the new reality.
This is a list of frequently asked questions that might arise on Apple Search Ads, tracking and SearchAdsHQ (by SplitMetrics) amid the upcoming Apple’s changes to IDFA. The FAQ will be updated as new information on the changes comes in.
Installs from iOS 14 users are tracked in the total Installs count in Apple Search Ads. LAT On Installs and LAT Off Installs don’t include installs coming from iOS 14 devices. Therefore, the sum of LAT On and LAT Off Installs may not match total Installs, because part of installs may be generated by iOS 14 users.
When you apply age and gender refinements, you automatically exclude customers with LAT turned On. On iOS 14 devices, the device’s Advertising Identifier is controlled by the tracking option “Allow Apps to Request to Track”. Device users who decline permission will not be served ads if the demographic refinements are set to a specific age or gender.
iOS 14 users may also choose to turn off Personalized Ads on their devices. If they do, they won’t be served your ads with specific age or gender refinements, or user types like “Returning users” or “Users of my other apps”.
We are not expecting a drop in ad spend after the iOS 14 release. Apple is the owner of true data knowing exactly which acquired users performed best. The channel itself is driving high-quality users for our customers and showing constant month-to-month growth. Compared to other ad channels, we are even expecting Apple Search Ads growth as a result of the release.
IDFA is not critical for SearchAdsHQ. In fact, our system doesn’t need it because we provide analytics at the app level, while IDFA is crucial for tracking users across apps of different publishers. Our system might use any unique identificator that is specific to a user for a particular app.
IDFA deprecation will not affect any SearchAdsHQ functionality except for one feature – Count Policy for Custom Conversions. This feature enables SearchAdsHQ clients to count either every conversion or only the first one by setting the count policy to “One Conversion.” Without IDFAs, the conversion type will be set to “Every” and the system will count every conversion (not the first one) that happens after a user taps on your ad.
Once the App Tracking Transparency (ATT) framework is fully implemented by Apple, we are going to switch to the internal identifiers specific to each MMP. Trackers are going to provide us with their specific identifiers so that we gather analytics and understand whether a user is unique.
The affected feature – Count Policy for Custom Conversions – will also be based on specific internal identifiers from trackers, which will be used to make a rough approximation of user-level data and thus enable the “One Conversion” count policy (counting only the first conversion resulting from the ad).
Similarly to Apple Search Ads, downloads from iOS 14 users are tracked under the total Downloads. LAT On Downloads and LAT Off Downloads exclude downloads coming from iOS 14.
Both in Apple Search Ads and SearchAdsHQ, LAT On and LAT Off Downloads don’t include downloads generated by devices running on iOS 14. These devices are counted in total Downloads. Therefore, you might see LAT On Downloads Share and LAT Off Downloads Share decreasing and not totalling to 100%.
Devices running on iOS 13 have the Limit Ad Tracking toggle in the Privacy Settings. By default, IDFAs of such devices are available unless users opt out.
Until Apple implements their privacy changes in early 2021, devices running on iOS 14 have accessible IDFAs. Still, users can deny access to their IDFAs via device settings where they can turn off “Allow Apps to Request to Track”. Similarly to LAT-enabled devices running on earlier iOS versions, this option, if turned off, will send all zeroes for IDFA.
Now that the new ATT framework is being rolled out, the True CPG formula has to take into account users with iOS 14 devices. AppsFlyer suggested that the number of users with zeroed IDFAs is growing, currently being at around 45% of all iOS 14 devices globally. Here’s are two modifications of a formula to calculate True CPG accounting for users who opt out of tracking:
True CPG (iOS 14) = Spend / (Downloads * Goals Rate * Q)
OR
True CPG (iOS 14) = Spend / (Q * Goals / Install Rate)
Where:
1) 5-10% of users download the app but never open it;
2) the conversion potential of trackable and non-trackable users is different: it is assumed that LAT On users and iOS 14 users with personalized ads off may get converted worse than trackable users.
By using Q = 1.0 or Q = 0.9, you don’t account for the conversion potential and the percentage of users who don’t open your app.
Use a lower Q-coefficient if you want to account for the percentage of users who don’t open your app and assume that LAT On and iOS 14 users with disabled personalized ads get converted worse than your trackable user bucket.
We suggest inserting 0.7 or 0.8 in the formula.
In SearchAdsHQ, you can configure the above formula (or any other formula) as a custom column. The system will automatically calculate your True CPG corrected for iOS 14 devices and show the result on the dashboard so that you can optimize your performance properly. Read more on custom columns here.
With the introduction of the ATT framework, the True ROAS formula has to account for users with iOS 14 devices. Here’s the formula we suggest using to account for users who opt out of tracking:
True ROAS (iOS 14) = (Q * Revenue / Install Rate) / Spend
Where:
1) 5-10% of users download the app but never open it;
2) the conversion potential of trackable and non-trackable users is different: it is assumed that LAT On users and iOS 14 users with personalized ads off may get converted worse than trackable users.
By using Q = 1.0 or Q = 0.9, you don’t account for the conversion potential and the percentage of users who don’t open your app.
Use a lower Q-coefficient if you want to account for the percentage of users who don’t open your app and assume that LAT On and iOS 14 users with disabled personalized ads get converted worse than your trackable user bucket.
We suggest inserting 0.7 or 0.8 in the formula.
In SearchAdsHQ, you can configure the above formula (or any other formula) as a custom column. The system will automatically calculate your True ROAS corrected for iOS 14 devices and show the result on the dashboard so that you can optimize your performance properly. Read more on custom columns here.
