Apple’s App Tracking Transparency requires third-party apps to ask permission before tracking users. The recurring criticism is that Apple’s own advertising operates under a narrower definition of tracking and a separate, less prominent consent flow.
Key takeaways
- App Tracking Transparency, usually shortened to ATT, is an iOS feature that makes third-party apps request explicit permission before tracking a user across apps and websites owned by other companies.
- The central criticism is not that ATT is ineffective, but that its definition of “tracking” excludes the kind of data combination Apple performs inside its own family of apps and services.
- Apple’s position has been that its advertising business does not share personal data with other companies and therefore does not meet its own definition of tracking, so the prompt does not apply to it.
- Competition authorities in several jurisdictions have examined whether ATT amounts to self-preferencing, though the status and outcome of individual proceedings is not something this article can confirm.
- The dispute matters because a permission prompt shown to some parties and not others can shift advertising revenue and measurement capability between competitors on the same platform.
What is actually being alleged
The claim that circulates in developer and advertising circles is one of asymmetry rather than deception. Under ATT, an app made by an independent developer that wants to use the device advertising identifier, or otherwise link a user’s activity to data held by another company, must display a system-generated prompt. If the user declines, the identifier becomes unavailable and cross-company measurement largely stops working.
Apple’s own apps and services sit outside that prompt. Personalised advertising in Apple’s own properties is governed by a separate setting located in the system preferences, which is a different interface, reached by a different path, and not surfaced at the moment a user opens an app. Critics argue that the practical result is a much higher rate of consent for one party than for everyone else, not because users prefer that party but because the two consent journeys are not comparable.
A second strand of the criticism concerns the definition itself. ATT treats “tracking” as the linking of user data with data from other companies for advertising or data-broker purposes. Combining data across services owned by a single company is not covered. Because Apple operates a large set of first-party services, this definitional boundary is said to fall in a place that happens to be favourable to the platform owner.
Why the question is resurfacing
Discussions of this kind tend to reappear when a technical analysis, a regulatory development or a developer’s own measurement work is shared and widely read. The topic has appeared on aggregator sites in that pattern more than once, and the arguments raised are broadly the same each time: comparisons of the two consent surfaces, readings of Apple’s published definitions, and observations from developers about how their measurement changed after ATT was introduced.
It is worth being precise about what a resurgence of interest does and does not indicate. It does not establish that a new finding has been made, that a regulator has ruled, or that Apple’s practices have changed. Without a verifiable primary document, the honest description is that a long-running argument has attracted renewed attention, and that the underlying design facts about ATT have been publicly documented since the feature launched.
The background a newcomer needs
Before ATT, mobile advertising on iOS relied heavily on a per-device advertising identifier that apps could read by default. Advertising networks used it to recognise the same device across different apps, to attribute an app install to the advertisement that produced it, and to build profiles for targeting. Users could reset or limit the identifier, but the control was buried and rarely used.
ATT changed the default. The identifier became unavailable unless the user granted permission through a standardised system prompt, with wording Apple controls. Apple presented this as a privacy measure that restores meaningful choice, and the low observed opt-in rates across the industry are generally consistent with the view that most users, when asked plainly, decline.
The commercial consequence was substantial. Businesses that depended on cross-app attribution lost precision, and some shifted spending towards channels where the advertiser owns the data. Platforms with large logged-in user bases and their own inventory were comparatively insulated, because they can measure inside their own walls without needing a cross-company identifier. Apple is one such platform, which is why the design of the boundary attracts scrutiny.
Who is affected and how
Independent app developers and small publishers are affected most directly. They rarely have a logged-in relationship across multiple properties, so they depended on third-party attribution to justify advertising spend. When that capability weakens, acquisition costs become harder to measure and, in many reported cases, effectively higher.
Advertising technology firms and measurement vendors face a structural problem, since their product was built on the identifier ATT restricted. Large advertising platforms are affected but have more room to adapt, through first-party data, modelled attribution and privacy-preserving measurement APIs.
Users are affected in a more ambiguous way. They receive a genuine control that did not exist before, and fewer of their identifiers travel between companies. Whether they experience a comparable level of control over the platform’s own advertising depends on whether they know the separate setting exists.
Where informed people disagree
There is no consensus on whether asymmetry here is a fault or a consequence of the concept. One view holds that “tracking” has always meant following someone between different parties, that a company using its own data about its own users is a different act in kind, and that requiring identical prompts would confuse users about a real distinction.
The opposing view holds that the user’s interest is in not being profiled, and that whether the profiling happens inside or across corporate boundaries is an accounting detail invisible to the person affected. On this reading, a rule written by a firm that also competes in the regulated market will tend to land where that firm’s activities are on the permitted side, whether or not anyone intended it.
A third position separates the privacy question from the competition question entirely: ATT may improve privacy outcomes and still constitute self-preferencing, and regulators are not obliged to choose one framing.
The practical implications
For developers, the working assumption should be that cross-company identifiers are no longer a dependable foundation. That points towards first-party relationships, server-side measurement of what a business can legitimately observe about its own users, aggregated and privacy-preserving attribution APIs, and greater tolerance for uncertainty in marketing measurement.
For advertisers, the shift rewards channels where the buyer can measure outcomes without third-party linkage, which mechanically favours large platforms with owned inventory. For users who want parity of control, the practical step is to review the platform’s own personalised-advertising setting rather than assuming the ATT prompt covers everything.
For anyone assessing the argument, the useful discipline is to distinguish what is documented in the platform’s own developer materials — the definition of tracking, the scope of the prompt, the location of the separate setting — from contested inferences about intent and effect.
What to watch next
The most consequential developments are likely to be regulatory rather than technical. Competition and data-protection authorities in several jurisdictions have taken an interest in how platform-level privacy controls interact with the platform owner’s own commercial activities, and any binding decision would set expectations well beyond a single feature.
Also worth watching: whether the definition of tracking is revised to cover intra-company data combination, whether the platform’s own advertising consent moves into a comparable prompt, and how privacy-preserving measurement standards mature. A further signal is whether independent, reproducible technical analysis of the two consent flows becomes available, since much of the current debate rests on reasoning from published policy rather than measurement.
Frequently asked questions
What does App Tracking Transparency actually do?
It requires apps on Apple’s mobile platforms to obtain explicit permission before tracking a user across apps and websites owned by other companies. The permission is requested through a standardised system prompt whose wording the platform controls. If the user declines, the app cannot access the device advertising identifier or link the user’s data to data held by other firms for advertising purposes.
Does Apple show the same prompt for its own advertising?
No. Apple’s own personalised advertising is governed by a separate control in the device settings rather than by the ATT prompt. Apple’s stated reasoning is that its advertising does not share personal data with other companies and therefore does not meet its published definition of tracking. Critics respond that the two consent journeys are not equivalent in visibility or timing, whatever the definitional argument.
Is this illegal?
That is unresolved and depends on jurisdiction. Privacy law and competition law ask different questions: one about lawful processing of personal data, the other about whether a dominant platform advantages itself against rivals. Competition authorities in several places have examined the issue. This article cannot confirm the current status or outcome of any particular proceeding, and readers should check primary regulatory sources.
Why did ATT reduce advertising revenue for some companies?
Cross-app advertising depended on a shared identifier to attribute installs and purchases to specific advertisements. When most users decline permission, that attribution breaks, and advertisers become less willing to pay for inventory whose results they cannot measure. Companies that sell advertising against other firms’ audiences were hit hardest; those measuring conversions inside their own logged-in services were comparatively insulated.
Can users turn off Apple’s own personalised advertising?
Yes. There is a setting within the device’s privacy and advertising preferences that controls whether the platform’s own services deliver personalised advertisements. It is separate from any answer given to an ATT prompt, so declining tracking in third-party apps does not change it. Users who want consistent treatment need to adjust it themselves, and many are unaware it exists.
Does this mean ATT was not really about privacy?
Not necessarily. The two explanations are not mutually exclusive. ATT can genuinely reduce cross-company data flows while also strengthening the position of the company that wrote the rule. Assessing motive from outside is speculative; assessing effect is more tractable, and the observable effect is that the restriction binds third parties more tightly than it binds the platform’s own advertising business.
Sources and further reading
- Apple’s published developer documentation on App Tracking Transparency, which sets out the definition of tracking and the scope of the permission prompt.
- Apple’s user-facing privacy and advertising support pages, which describe the separate personalised-advertising control.
- Published decisions and press releases from European competition authorities that have examined platform privacy controls and self-preferencing.
- Trade coverage in advertising and mobile-marketing publications reporting developer-side measurement of opt-in rates and attribution changes.
Surfaced from the hackernews signal “platform privacy rule asymmetry”. AI-assisted draft, editorially reviewed.
