ATT and the Networks That Absorbed the Loss

How a permission prompt rewired mobile advertising

When Apple pushed iOS 14.5 to users in April 2021, it required every app to ask permission before accessing the device's Identifier for Advertisers — the IDFA, the per-device tracking token that had underpinned mobile ad targeting since 2012. The prompt was two lines of text and two buttons. Its financial consequences took the better part of a year to show up fully in earnings reports, but when they did, the figures were among the largest single-policy losses recorded in the digital advertising industry.

App Tracking Transparency did not ban tracking. It required an explicit opt-in rather than the existing opt-out, and on that single change the entire mobile attribution stack for third-party advertisers depended. Opt-in rates varied across markets and app categories, but Apple's ATT framework documentation ↗ made clear that apps receiving a "denied" response were prohibited from reading the IDFA, which effectively collapsed the cross-app identifier pool available to ad networks outside Apple's own ecosystem.

Snap warned first; Meta counted the cost

Snap was the first major platform to quantify the damage in public. On its second-quarter 2021 earnings call, the company told investors that the ATT changes had disrupted its ability to measure and target advertising, and in subsequent quarters it revised revenue guidance downward, attributing the shortfall explicitly to measurement loss. Snap's stock fell more than 20 percent in after-hours trading following its third-quarter 2021 results — a quarter in which it reported that iOS changes had created "a revenue headwind" it was still working to address. The language was careful, but the direction was not.

Meta offered the largest single figure. On its fourth-quarter 2021 earnings call, the company said it expected the impact of Apple's iOS changes to represent a headwind of approximately $10 billion to its 2022 advertising revenue. That estimate — attributed to the company's own forecast and widely cited in subsequent regulatory filings — became the most-quoted number in the ATT debate. Meta had the scale and the internal measurement apparatus to produce a plausible estimate where smaller networks could not; it also had the motive to publicise the figure, given its public criticism of the policy.

Both companies had previously built their mobile advertising products around the assumption that IDFA availability was the default. The opt-in requirement broke that assumption in a matter of weeks.

SKAdNetwork and what replaced the identifier

Apple had introduced SKAdNetwork ↗ in 2018 as a privacy-preserving attribution framework, but adoption had been negligible while the IDFA remained freely available. After ATT, it became the primary measurement mechanism available to iOS advertisers. SKAdNetwork works by routing aggregate conversion signals through Apple rather than through ad networks directly — meaning ad networks lose access to the individual-level data that made precise targeting possible.

The framework has real constraints: it returns a single, delayed, coarse conversion value per install, and it does not identify the device or user. For large platforms running statistical models at scale the loss was recoverable, at cost. For smaller networks and mobile measurement partners — companies whose entire business rested on per-device attribution — the transition was structural rather than cyclical.

By 2022, Meta, Snap and others had invested heavily in building probabilistic modelling systems to substitute for deterministic IDFA-based attribution. Meta's Aggregated Event Measurement and its Conversions API were designed to supplement, and partly replace, the signal ATT had removed. These systems rely partly on first-party data that users have consented to share directly with the platforms — which placed larger platforms with large logged-in user bases at a measurable advantage over ad networks lacking that first-party relationship.

The competitive asymmetry was noted by regulators as well as investors. The UK Competition and Markets Authority examined ATT as part of its broader 2021 investigation into Apple and Google's mobile ecosystems, observing that the policy, whatever its privacy merits, had consequences that differed in practice depending on whether a network had first-party logged-in data or relied entirely on cross-app tracking. Apple's own advertising business — operating under different rules by virtue of the App Store being a first-party context — was not subject to the same consent requirement.