What Changed and Why Does It Matter?
With iOS 14.5, Apple required every app to obtain separate permission from users. When users chose 'Ask App Not to Track,' their behaviors — visiting a website, making a purchase, filling out a form — could no longer be followed by Meta's pixel system (the tracking code placed on websites). The result: ad managers lost visibility into a portion of conversions that actually happened, and budget decisions had to be made based on incomplete data.
declined tracking when shown the iOS permission prompt — industry estimates close to Meta's own data pointed in this direction, making the scale of data loss concrete.
Methods Used to Narrow the Measurement Gap
- Conversions API (CAPI): Sending data to Meta directly from your own server rather than relying on the pixel. Even if a user's browser blocks tracking, server-side information can still be passed. Using it alongside the pixel gives better measurement coverage.
- First-party data: Email lists, phone numbers, and CRM records you collected yourself. These can be uploaded to Meta to build lookalike audiences and reach existing customers — a foundation that doesn't rely on cookies.
- Modeled conversions: Meta uses machine learning to estimate missing conversion data. The numbers shown in Ads Manager may include these estimates, so every figure should be read as an approximation rather than an exact measurement.
- Aggregated Event Measurement: Meta's system that lets you prioritize up to eight conversion events per domain. You choose which actions to measure, and iOS restrictions are applied within that set of eight.
- Privacy-aware attribution windows: After the iOS change, the 28-day click attribution window was shortened. This can make ad impact look smaller than it is; choosing attribution windows deliberately has become more important.
