Did iOS 14.5 actually reduce my conversions, or just my reported conversions?

Almost always just the reporting. The purchases still occur, but ATT removed the identifier that let the ad platform attribute them, so they disappear from Ads Manager while still appearing in your store analytics. Comparing platform-reported revenue to actual store revenue tells you which one you are dealing with.

What is a good Event Match Quality score?

6.0 out of 10 is acceptable and 7.0 or higher is strong. Scores below 5.0 usually mean you are sending only email, or missing click identifiers and IP/user-agent data. Each additional hashed identifier typically adds a few tenths of a point.

Will a Conversion API double-count my conversions?

Only if you deploy it without deduplication. Send the same event_id from both the browser pixel and the server event, and the platform collapses them into one conversion. Verify the match rate in Events Manager before trusting the numbers.

Should I remove the browser pixel once server-side tracking is live?

No. The browser pixel supplies the cookie-based identifiers and the deduplication partner for your server events. Running both paths produces higher match quality than either alone.

How long before recovery shows up in performance?

Match quality improves within days of deployment, but the optimization algorithms need roughly two to four weeks of cleaner feedback before delivery changes. Judge the outcome on blended ROAS over six weeks, not on the first week of platform-reported numbers.

iOS Tracking Recovery Guide: Rebuilding Conversion Signal After ATT

By CAPIbara Team · Published July 31, 2026 · Updated July 31, 2026

iOS Tracking Recovery Guide: Rebuilding Conversion Signal After ATT Short answer: iOS App Tracking Transparency did not delete your conversions — it deleted the browser-side evidence of them. Recovery means rebuilding that evidence server-side with first-party data, then validating the result against blended ROAS instead of platform-reported ROAS. This guide walks the full recovery path: diagnosing how much signal you actually lost, restoring it with a Conversion API, raising match quality, and confirming the fix in numbers you can trust. # What ATT actually broke Short answer: ATT stopped ad platforms from reading the device identifier that linked a click to a purchase. The purchase still happens; the platform just can no longer prove it was theirs. Three things degraded at once when iOS 14.5 shipped: Identifier loss. IDFA opt-in rates settled in the 20–30% range, so most iOS sessions arrive anonymous. Cookie lifespan collapse. Intelligent Tracking Prevention caps client-side cookies at 7 days — and 24 hours for click-decorated links — so a browser pixel forgets the visitor long before a considered purchase closes. Delayed, modelled reporting. Aggregated Event Measurement batches conversions and fills gaps with statistical modelling, which means the numbers in Ads Manager are an estimate, not a ledger. The practical consequence: campaigns that are profitable look flat, the algorithm optimizes on partial feedback, and budget quietly drifts toward whatever channel still report…