iOS ATT (App Tracking Transparency)

What is iOS ATT?

iOS ATT (App Tracking Transparency) is Apple’s opt-in prompt that asks a user’s permission before an app can track them across other apps and sites, and because most users decline, it collapsed the deterministic signal ad platforms relied on for optimization and attribution.

How it works and why signal collapsed

ATT shipped with iOS 14.5 in 2021 and turned cross-app tracking into a per-app permission most people refuse. With the identifier gone for the majority of iPhone users, platforms can no longer observe many conversions directly, so Meta now models a meaningful share of iOS conversions (statistically estimating them) rather than counting them. Modeled conversions arrive later, in aggregate, and with less certainty, which is why post-2021 reporting feels laggier and less granular than the deterministic era before it.

A worked example of the reporting gap

Walk it with round, illustrative numbers. Say 100 patients subscribe after clicking iPhone ads, but only 40 opted into tracking. The platform observes those 40 directly and models an estimate for the rest, so what it can attribute cleanly is a fraction of the true total. The other conversions still happened and still made money; the platform simply cannot see them well, so the CAC it reports reads worse than the CAC the bank account actually experienced.

Why health feels it more than most

Telehealth and supplement traffic skews heavily mobile and iPhone, so the ATT blind spot lands squarely on the categories that can least afford it. A weight-loss advertiser sees fewer matched conversions, laggier and coarser reporting, and an inflated platform CAC that can make a profitable campaign look like a loser and get it paused. The working response is to stop relying on the browser signal and rebuild it server-side:

• Send conversions from your server with CAPI, deduplicated against the pixel, so ATT opt-outs, ad blockers, and Safari cookie limits stop erasing events.
• Lean on first-party data (your own customer and intake records) to match conversions the browser can no longer.
• Feed clean conversion signal so the algorithm optimizes on nearly all of your conversions instead of a fraction.

Why it matters for health and DTC brands

Recovering signal is not optional in health, because the same server-side layer that fixes attribution is also where compliance lives. A server-side pipeline lets a brand restore optimization signal while stripping the condition, medication, and intake fields the FTC drew a hard line around in its GoodRx ($1.5M) and BetterHelp ($7.8M) actions in 2023. Done right, ATT stops being a reporting handicap and the platform optimizes a GLP-1 campaign on almost all of its true conversions, none of which carry protected health data. Pair CAPI with server-side tracking, and judge campaigns on blended marketing efficiency ratio (MER) and real CAC rather than the platform’s ATT-degraded numbers.

Related terms