Meta Ads After iOS 14.5: What Actually Changed for Tracking

In April 2021, Apple shipped iOS 14.5 with App Tracking Transparency (ATT): any app that wanted to access the IDFA — the device-level identifier advertisers had relied on for years — now had to show an explicit opt-in prompt first. Overwhelmingly, users who saw that prompt said no. That single interface change is why "our Meta ads stopped performing the way they used to" is still a live conversation years later, and why the fix was never a setting to flip once — it's an ongoing tracking-architecture decision.

What actually broke is specific: Meta's Pixel-only, browser-and-device tracking lost reliable visibility into a meaningful share of iOS conversion events at the individual level. Meta could no longer consistently see one specific person click an ad and, in a separate app or session later, complete a purchase — the deterministic, one-to-one link between an ad click and a conversion event got structurally thinner, not just less precise.

Meta's response had two real parts. Aggregated Event Measurement (AEM) caps and prioritizes web conversion events per domain instead of tracking every event without limit, which forces a real decision: rank which conversion actions actually matter most for that domain, because only a limited set will be measured this way at all. Conversions API (CAPI) is the more consequential half — server-to-server event sharing sent directly from a business's own server, CRM, or e-commerce platform straight to Meta, deduplicated against browser Pixel events using a shared event ID, built on hashed first-party identifiers like email and phone instead of depending solely on a cookie or device identifier a user can decline.

This is also why a number sitting in Ads Manager today should be read differently than it was in 2020. Meta now statistically models a portion of the conversions it can no longer directly observe, filling gaps with estimates rather than one-to-one recorded events. That's not the platform being dishonest — it's the mechanical consequence of reduced signal — but it means a reported number is a directional estimate, not a literal ledger entry, and treating it as gospel at the account level is a mistake the platform's own documentation would tell you not to make if you read it closely.

Practically, for an account being set up today rather than patched from 2021: Conversions API is close to mandatory, not a nice-to-have — a Pixel-only setup on meaningfully iOS-heavy traffic is structurally under-counting regardless of how well the campaigns are actually performing. The 8-event cap under AEM means event-priority order has to be a deliberate choice — usually purchase or qualified lead ranked above add-to-cart, ranked above view-content — rather than left on whatever the platform defaults to. And attribution windows genuinely shrank across this period, with many advertisers seeing default windows move toward shorter click-attribution periods, because the platform can no longer reliably attribute across a long, cross-device gap without consent-based signal to bridge it.

This is exactly the same discipline the Signal-to-Revenue Framework applies everywhere else, not something unique to Google Ads: the last stage reconciles attribution against CRM-stage data instead of trusting platform-reported conversions outright. On a Meta account specifically, that means CAPI configured and properly deduplicated against the Pixel, event priority set deliberately instead of left default, and the real read on performance is what the CRM shows as closed revenue, not what Ads Manager reports as a modeled conversion.

The bigger structural lesson is that first-party data — a CRM's own contact and deal records, hashed and matched back to Meta as a Custom Audience — is now a more durable foundation for both tracking and targeting than anything purely browser-based ever was. That shift quietly rewards businesses with real CRM discipline over ones relying on ad-platform tracking as their only source of truth, which is as much a CRM problem as it is a media-buying one.