App Tracking Transparency (ATT) is Apple’s framework requiring every iPhone and iPad app to ask a person’s permission before tracking them across apps and websites owned by other companies. It arrived with iOS 14.5 in April 2021, and it is the reason the “Ask App Not to Track” prompt appears when you open many apps for the first time.
How App Tracking Transparency works
Before ATT, apps could read a device’s advertising identifier and use it to connect what someone did in one app with what they did elsewhere. That connection is what let Meta see that a person who viewed your ad in Instagram later bought something on your website.
Under ATT, an app that wants to track must show Apple’s standard prompt. If the person taps “Ask App Not to Track”, the app cannot access the advertising identifier and is not allowed to track them by other means either, such as fingerprinting the device. People can also switch tracking off for all apps at once in their iPhone settings.
The prompt is separate from your own website’s cookie banner. A UK visitor on an iPhone may decline tracking in the Instagram app and then accept cookies on your site, or the other way round. Each decision limits a different part of the data trail.
Why it matters
For advertisers, ATT removed a large part of the signal Meta used to measure and optimise campaigns. Meta responded with Aggregated Event Measurement, delayed and modelled reporting, and shorter default attribution windows. It also pushed advertisers towards sending conversion data from their own servers through the Conversions API.
The UK effect is noticeable because iPhones make up a large share of smartphones here. A clinic in Bath or a candle shop in Brighton advertising to mobile users will find that a meaningful portion of their conversions are reported late, in aggregate or not at all. Campaigns may be working better than Ads Manager suggests, or worse; you cannot tell from Meta’s figures alone.
App advertisers face a stricter version of the same problem. Install and in-app event measurement for opted-out users runs through Apple’s own privacy-preserving attribution systems, which report fewer details and with delays.
Common mistakes
- Assuming ATT is a Meta setting. It is Apple’s rule. You cannot switch it off or work around it, and trying to identify opted-out users breaks Apple’s policy.
- Treating it as a passing problem. It has been in place since 2021 and the tracking environment has only tightened since.
- Confusing it with cookie consent. Your UK GDPR and PECR consent duties apply on your website regardless of what Apple does.
- Measuring only inside Ads Manager. With partial data, platform reports need checking against your own sales records.
How to act on it
Start with the data you own. Make sure purchases and leads are recorded in your own systems with a source where possible, and that the Meta Pixel is backed by server-side events with good customer matching. Ask for consent properly on your site, because clean consented data is worth more than patchy data collected without it.
Then change how you judge results: look at weekly trends, compare Meta’s totals with your actual orders, and test with occasional holdout periods if your budget allows. I set up this kind of measurement as standard in my Facebook ads management service.