Originally published November 2019. This video predates App Tracking Transparency (ATT), introduced in iOS 14.5. Its limited-ad-tracking adjustment is historical.
Short answer: this 2019 App Masters video explains a gap between Apple Ads (formerly Apple Search Ads) and mobile measurement partner (MMP) reports, then estimates unobserved revenue. The calculation assumes that unobserved users generate the same revenue per install as observed users. It is a model, not a verified correction. Despite the title’s “ROI”, the revenue-to-ad-spend calculation is ROAS, not profit ROI. Watch it on YouTube.
Page Contents
Why did the reports differ?
The video lists four explanations:
- Apple counted downloads, while MMPs measured installs at first open.
- MMPs deduplicated across networks, while Apple reported its own attribution.
- Apple included redownloads.
- Users with limited ad tracking (LAT) enabled could appear as organic in the MMP instead of being attributed to Apple’s ads.
The walkthrough focuses on the fourth gap. These describe the historical setup, not a complete diagnosis of a reporting discrepancy today.
How does the video adjust the estimate?
- Split Apple’s reported conversions into LAT-on and LAT-off groups.
- Calculate LAT-on divided by LAT-off, week by week, checking outliers. The example rounds a ratio near 98% to about 1:1.
- Multiply the MMP’s observed installs by one plus that ratio. A 1:1 ratio produces a multiplier of two, not one.
- Apply that multiplier to revenue only under the equal-revenue-per-install assumption. In the example, doubling estimated installs roughly halves CPI; doubling day 7 revenue takes day 7 ROAS from an implied 30% to 60%.
These are illustrative numbers from one historical account. The missing users could have different revenue behavior, so doubling their revenue is not established by the install ratio alone.
What has changed since?
Apple’s ATT rules require permission to access the advertising identifier; without it, the identifier is zeroed. Today’s Apple Ads attribution uses AdServices. Apple’s current documentation also excludes attribution for ad groups using age or gender targeting. Check the current integration and counting rules rather than carrying over the old LAT multiplier.
For current context, see MMPs after ATT, Apple Ads for subscription apps and Apple Ads with Meta.
Frequently asked questions
Can I reuse the 1:1 ratio? No. It was an example from one account in 2019, and both attribution coverage and user value must be checked for your own data.
Does an install gap prove revenue is understated by the same proportion? No. That estimate requires comparable revenue per install across observed and unobserved users.
