Roughly 75 percent of iOS users opt out of app tracking in 2026, and for many Meta advertisers that single Apple prompt is why reported conversions on iPhone traffic look like they collapsed overnight — even though the sales never stopped. App Tracking Transparency does not turn off your campaigns; it removes the deterministic signal Meta used to tie a purchase back to a click, so events that genuinely happen stop being reported the way they were.
This guide works forward from the opt-out through seven signal-loss points — ATT consent, the 8-event limit, domain verification, the Conversions API, modeled conversions, the attribution window and dropped breakdowns — so you fix measurement instead of chasing a phantom performance drop. To check your account against the most common post-ATT signal leaks automatically, run our free 5-axis ad account audit.
Updated 2026-05-17 with current Aggregated Event Measurement, Conversions API, and modeled-conversion behavior observed across US, UK and European accounts.
- Roughly 75 percent of iOS users opt out in 2026 — the drop is in reporting, not in real sales. 2. Aggregated Event Measurement caps you at 8 events — only the top-priority one a user completes is counted. 3. Domain verification comes first — without it you cannot configure or rank those eight slots. 4. The Conversions API recovers signal the pixel loses to blocking when deduplicated with a shared event ID. 5. Modeled conversions are estimates with a delay of up to 3 days — judge totals and trends, never a single conversion.
How do ATT opt-outs cut iOS conversion signal?
ATT is the first signal-loss point to understand because every constraint below it flows from one user choice. When an iOS user taps Ask App Not to Track, Meta loses the deterministic identifier it relied on, and the chain from click to conversion breaks at the source.
App Tracking Transparency — Apple's prompt asks each user whether an app may track them across other apps and websites. In 2026 the great majority decline, so for most iOS traffic Meta cannot observe the conversion the old way and must measure within a privacy-safe framework instead.
Deterministic loss — Without permission, there is no stable identifier to match a purchase to the specific click that drove it. The sale still happens, but the link your dashboard used to show simply is not available for that user, which is why reported numbers fall while revenue holds.
A reporting drop, not a performance drop — This distinction is the whole point. Before you pause campaigns or slash budgets, prove whether sales actually fell or whether they merely stopped being counted. Our post-ATT strategy guide covers the structural response in depth.
What is Aggregated Event Measurement and the 8-event limit?
Once a user opts out, Meta can no longer report that conversion freely; it must route it through Aggregated Event Measurement, and that framework imposes the constraint that surprises most advertisers — a hard cap of eight events per domain.
Aggregated Event Measurement — This is Meta's privacy-safe system for measuring web events from opted-out iOS users. It aggregates conversions so no single person can be identified, which is exactly why it strips the detail you used to see.
The 8-event limit — Each verified domain can configure only 8 conversion events. They are ranked by priority, and for any given user only the single highest-priority event they complete is counted. A purchase will outrank an add-to-cart, so the cart event is not measured if the same user buys.
Why the order matters — If you put a low-value event like a page view in a high slot, it can crowd out the purchase you actually care about. Configuring the eight slots is not optional housekeeping; it directly decides what Meta can measure for the majority of your iOS traffic. The next section walks the exact configuration in Events Manager.
How do I verify my domain and prioritize 8 events?
You cannot touch the eight-event priority list until Meta trusts that the domain is yours, so domain verification is the lock that opens the whole post-ATT configuration. Do it first, then rank your events.
Domain verification — In Business Manager, add the DNS TXT record, an HTML file, or a meta tag to prove ownership, then confirm it in Events Manager. Until this is green, you cannot configure or prioritize Aggregated Event Measurement events, and Meta may measure nothing useful for opted-out users.
Configuring the 8 events — Open Events Manager, select the verified domain, and choose the eight conversion events that matter. Map them to your real funnel: purchase, lead, initiate checkout, add to cart, and so on, only as far down as the eighth slot.
Prioritizing by value — Rank the highest-value action first. Because only the top event a user completes is counted, your money event belongs at the top and low-value micro-conversions belong at the bottom or off the list entirely. To compare this with Google's privacy setup, see our CAPI versus Enhanced Conversions comparison.
Can the Conversions API recover lost signal?
Verification and event priority decide what Meta is allowed to measure; the Conversions API decides how much of it actually arrives. This is the single highest-leverage recovery lever for post-ATT measurement.
The Conversions API — Instead of relying only on the browser pixel, which ATT, ad blockers, and ITP can degrade, the Conversions API sends events server-to-server from your backend straight to Meta. Because the server already knows a purchase happened, it can report the event even when the browser cannot.
Hashed matching — The server passes hashed customer data such as email and phone so Meta can match the event to a user privately. Better matching means more conversions are recovered and attributed, which improves the event match quality score you should watch in Events Manager.
Deduplication — Run the Conversions API alongside the pixel, not instead of it, and share a single event ID so the same purchase is not counted twice. Done right, the server feed fills the gaps the pixel leaves rather than inflating your numbers. For the build options, compare our CAPI Gateway versus server-side GTM guide.
What are modeled conversions and the reporting delay?
Even with a clean Conversions API feed, some opted-out conversions can never be observed directly. Rather than report zero, Meta estimates them — and those estimates arrive on a delay you have to plan around.
Modeled conversions — When Meta cannot directly observe a conversion because the user opted out, it uses aggregated and consented data to statistically estimate how many likely converted, then reports them as modeled. They keep your numbers honest in aggregate rather than understating performance.
The reporting delay — Aggregated events are processed in batches, so modeled and aggregated conversions appear with a delay of up to about three days. Today's spend will not show its full conversion picture until the data settles, which is normal under this framework.
Read them as directional — Modeled numbers are most reliable at the account and campaign level and least reliable in granular breakdowns. Judge trends and totals, never a single conversion, and never react to one day of unsettled data. To turn recovered conversion value into a return figure, use our pixel and CAPI mismatch fix to confirm the feed is clean first.
The post-ATT Meta measurement diagnostic table
Work this table top to bottom — it is ordered by how foundational each signal-loss point is and how often it is the real reason your iOS conversions look broken.
When ATT cuts your reported conversions, pausing iOS campaigns feels like the obvious fix, but it usually kills real sales that were still happening and never showing up in the pixel. A campaign whose purchases land in your backend but not your dashboard is a measurement problem, not a performance one. Recover the signal with domain verification, the eight-event priority list, and the Conversions API first, then judge the campaign on settled, modeled totals — not on the raw pixel number the day after Apple's prompt.
What can and cannot you measure after ATT?
You will not get the pre-ATT dashboard back, and chasing it wastes time. The goal is to recover as much trustworthy signal as the framework allows, then build reporting around what is genuinely measurable.
What you can measure — Aggregate conversion totals, campaign-level trends, modeled conversions, and server-confirmed events through the Conversions API. With verification and the eight events configured, these are reliable enough to optimize on at the account and campaign level.
What you cannot measure — Granular breakdowns by age, gender, region, and placement are limited or gone for opted-out iOS users, fine-grained per-user attribution is unavailable, and the attribution window has narrowed to 7-day click and 1-day view with a roughly 72-hour processing delay on aggregated events.
Build reporting around totals. Set stakeholder expectations to trends and aggregates, not single rows or fine slices, and accept slightly older numbers as the cost of privacy-safe measurement. Re-measure matched conversions and event match quality after each fix, not all at once, so you know which lever moved the result. To recover every recoverable signal automatically, run the SteerAds free 5-axis audit, and size the return on your recovered conversion value with our ROAS calculator.
Sources
Official sources consulted for this guide:
-
facebook.com — Aggregated Event Measurement
-
facebook.com — domain verification
-
developers.facebook.com — Conversions API docs
-
facebook.com — Meta Ads
FAQ
Why did my Meta conversions drop on iOS after ATT?
App Tracking Transparency forces a prompt that lets iOS users deny app-level tracking, and in 2026 roughly 75 percent of them opt out. When a user opts out, Meta loses the deterministic identifier it used to attribute a conversion back to the click, so events that genuinely happened simply stop being reported the way they were. The drop you see is usually a reporting drop, not a real performance drop. Sales still occur, but they arrive as modeled or delayed numbers, or fall outside the limited measurement Aggregated Event Measurement allows. Fix measurement before you judge any campaign.
What is Aggregated Event Measurement and why only 8 events?
Aggregated Event Measurement is Meta's privacy-safe framework for measuring web events from iOS users who did not grant tracking permission. To protect individual privacy, Apple's rules cap each verified domain at eight configured conversion events, ranked by priority. Only the single highest-priority event a user completes is counted, and only one of those eight slots fires per session. That is why you must choose and rank your eight events deliberately in Events Manager. Anything past the eighth slot, or any event you forgot to configure, is invisible for opted-out iOS traffic and never reaches your reports.
Does domain verification matter for Meta measurement?
Yes, and it is non-negotiable for the post-ATT setup. Until you verify the domain in Business Manager, you cannot configure or prioritize the eight Aggregated Event Measurement slots that decide what gets measured for opted-out iOS users. An unverified domain means Meta picks events for you or measures nothing useful, and you lose control over which conversion takes the single available slot. Verification also protects your events from being claimed by another business. Add the DNS TXT record or meta tag, confirm it in Events Manager, then configure your priority list before you spend another dollar.
How does the Conversions API recover lost signal?
The Conversions API sends conversion events server-to-server, from your backend straight to Meta, instead of relying only on the browser pixel that ATT and trackers can block. Because the server already knows a purchase happened, it can pass hashed customer data and the event even when the pixel is blocked, which recovers signal the browser drops. Pair it with the pixel and deduplicate with a shared event ID so the same purchase is not counted twice. Properly deployed, the Conversions API typically lifts matched conversions and improves the data feeding Meta's optimization on iOS traffic.
What are modeled conversions in Meta Ads reporting?
Modeled conversions are statistical estimates Meta reports for conversions it can no longer observe directly because the user opted out of tracking. Rather than show zero, Meta uses aggregated and consented data to estimate how many opted-out users likely converted, then attributes them in reporting. They are labeled as modeled and appear with a reporting delay of up to about three days. Treat them as directional, not exact: they are best at the account and campaign level and least reliable for granular breakdowns. Judge trends and totals on modeled data, never a single conversion.
Why are my Meta breakdowns missing after ATT?
Aggregated Event Measurement deliberately strips detailed breakdowns to protect privacy, so age, gender, region, and placement splits are limited or unavailable for opted-out iOS conversions. Attribution is also confined to a shorter window, and the default has moved to a 7-day click and 1-day view model with a roughly 72-hour processing delay on aggregated events. You will see fewer dimensions and slightly older numbers than you did before ATT. This is a permanent constraint of the privacy framework, not a bug, so build your reporting around totals and trends rather than fine-grained slices.
How long does it take to recover Meta signal after ATT?
The foundational fixes ship within a day or two but need data to prove out. Domain verification and configuring your eight events take an afternoon, and the priority list takes effect quickly. A Conversions API deployment usually takes a few days of engineering plus a verification window to confirm the event match quality score is healthy. After that, allow one to two weeks for modeled conversions and optimization to stabilize on the richer signal. Sequence the work so verification and event priority land first, then layer the Conversions API and let learning settle.