Apple changed the terms of mobile customer acquisition
App Tracking Transparency required permission for tracking across other companies' apps and websites. The change tied advertising measurement and product analytics to users' privacy choices and the platform's rules.
When Apple released iOS 14.5 on 26 April 2021, a new permission prompt changed the information available to mobile businesses. App Tracking Transparency required apps to obtain permission before tracking data across apps or websites owned by other companies for advertising, or sharing data with data brokers. Users could review their choices in Settings and change them later.
For a mobile startup, the implications extended beyond adding a dialog. Advertising could help bring a user into the app, while the evidence used to assign that user to a campaign depended on data flows controlled by several companies. A change in the operating system could therefore alter how the startup measured growth even if the product itself continued to work.
The requirement had been signalled before launch
Apple's developer notice in January said it had temporarily deferred the requirement to give developers more preparation time. It announced the requirement would return in an upcoming beta, with a wider rollout in early spring. Without permission, tracking would not be allowed and the device's advertising identifier would be all zeros.
- 27 January 2021
Apple describes the privacy changes
Its Data Privacy Day announcement explains the permission requirement and distinguishes it from App Store privacy labels.
- 28 January 2021
Developers receive a requirement update
Apple describes the upcoming beta requirement and the advertising identifier becoming all zeros without permission.
- 26 April 2021
iOS 14.5 is released
App Tracking Transparency arrives in the publicly available software update, with choices accessible in Settings.
Apple's announcements and software release. Availability did not mean every user had updated or made the same permission choice.
The public release of iOS 14.5 brought the permission requirement into effect. The proportion of users granting permission, the pace of device updates and the advertising costs faced by individual businesses still had to be measured. The prompt alone could not establish those commercial effects.
Tracking and ordinary product measurement were different questions
Apple's rule concerned links across other companies' apps and websites and sharing with data brokers. Businesses could still observe activity inside their own products. A purchase, subscription renewal or completed onboarding remained a business event; the restricted data path was how activity became connected across services and used for advertising.
That distinction was useful for growth teams. A campaign dashboard might attribute fewer purchases after a tracking change without proving an equal reduction in total purchases. Conversely, stable total revenue would not prove every campaign remained efficient. Teams needed to reconcile advertising reports with aggregate sales and product activity, and acknowledge uncertainty when the link between a campaign and a later event became less observable.
The measurement design had to change with the permissions
The developer notice also described planned improvements to SKAdNetwork and support for Private Click Measurement. Apple presented them as ways to measure advertising while preserving privacy, with further details to follow. These alternatives gave developers other measurement mechanisms to consider, although Apple did not promise they would reproduce every signal available through individual tracking.
Scroll to view all columns
| Comparison | Observation | Additional evidence needed |
|---|---|---|
| A customer subscribes | The product records a completed subscription | Whether a particular advertisement caused the subscription |
| An advertising report changes | The platform attributes a different number of conversions | Whether demand changed, measurement changed, or both |
| A test group records higher sales | A defined test group performs differently | Whether groups were comparable and the result persists |
Questions for evaluating acquisition evidence under Apple's permission and measurement changes. The examples illustrate what to test; they do not report campaign results.
An experiment needed a defensible comparison
A young company could respond by testing budgets against observable business outcomes, comparing groups or markets where a defensible experiment was possible and checking whether reported acquisition costs used consistent definitions. Seasonality, product changes and other campaigns could affect the same outcomes. Spending decisions had to account for these limitations, including the uncertainty behind apparently precise attribution figures.
For example, a team could compare aggregate paid subscriptions during a defined campaign test while holding other major changes constant. The test would still need enough time and volume to distinguish a persistent effect from ordinary variation. This approach illustrates one way to evaluate acquisition; its feasibility would depend on the business and the data available.
A third-party library could be part of the product's promise
Apple separately described App Store privacy labels as summaries based on practices reported by developers. A label explained those practices to prospective users; a tracking permission asked for a decision about tracking. Implementing either accurately required a clear understanding of what the app and its embedded suppliers actually did with data.
For software startups, that placed supplier selection beside growth strategy. An advertising or analytics library might send information beyond the company's own systems. Product, engineering and commercial teams needed a shared account of that behaviour, including the permissions involved and how a change affected the user experience. A vendor promising improved measurement still had to explain which data supported the claim and whether the approach fit the platform's requirements.
Distribution dependence belonged in the forecast
The investment issue was concentration in a route to customers. A startup could own its application and customer proposition while depending on an operating system, advertising network or measurement provider for crucial parts of acquisition. Forecasts based on historical campaign efficiency needed to identify those assumptions and show how the business would respond if visibility or permissions changed.
The announcements did not establish how much advertising costs would change or whether every participant would benefit. They made privacy choices an explicit condition of important advertising data flows. Investors needed to examine how a company explained its data practices, measured customer value and made spending decisions with the information it could legitimately obtain.
Investment implications
- A loss of attribution detail is not automatically a loss of all demand; measurement and customer behaviour must be examined separately.
- The data practices of embedded advertising and analytics suppliers belong in the product review.
- Acquisition models should identify which platform permissions and external data links their forecasts assume.
Sources
- Data Privacy Day at Apple: Improving transparency and empowering users (opens in a new tab)
Apple
- AppTrackingTransparency requirement update (opens in a new tab)
Apple Developer
- iOS 14.5 delivers Unlock iPhone with Apple Watch, more diverse Siri voice options, and new privacy controls (opens in a new tab)
Apple
Published
