How attribution works in Meta Ads in 2026: updated windows, iOS 14.5 impact, incremental attribution, and comparison with external measurement tools.
Attribution in Meta Ads changed radically between 2021 and 2026. The Facebook Attribution tool was retired in August 2021. The 28-day windows disappeared after iOS 14.5. Meta introduced incremental attribution in 2025 and completely redefined what counts as a “click” in March 2026.
If your knowledge of Facebook or Meta Ads attribution comes from before 2022, this guide is for you.
To assign conversions, Meta Ads relies on three elements:
The most important element in practice is the attribution window, because it determines which conversions appear in your reports and (more critically) how Meta’s algorithm optimises ad delivery.
The attribution windows currently available in Meta Ads Manager are:
| Window | Description |
|---|---|
| 1-day click | Conversion within 24 hours of clicking an ad link |
| 7-day click | Conversion within 7 days of clicking an ad link (default) |
| 1-day view | Conversion within 24 hours of viewing an ad without clicking |
| 1-day video engagement | For videos: conversion after watching at least 5 seconds, within 24 hours |
Default setting in 2026: 7-day click + 1-day view.
The 28-day windows (click and view) were eliminated in 2021 as a direct consequence of iOS 14.5. The 7-day view window was removed from the Meta Insights API on 12 January 2026, causing 15–30% drops in attributed conversions for advertisers who were using it.
This is a critical point many advertisers miss: the window you configure is not just a reporting preference. It determines how Meta’s algorithm optimises who it shows your ads to.
For most businesses, the 7-day click + 1-day view setting is the right balance.
In April 2021, Apple launched iOS 14.5 with the App Tracking Transparency (ATT) framework, which requires apps to ask users for explicit permission before tracking their activity. 70–75% of iOS users opted out of tracking.
The effect on Meta attribution was structural and permanent:
The correct technical response is the Meta Conversions API (CAPI). Unlike the pixel, which operates from the user’s browser, CAPI sends data directly from your server to Meta, without depending on the device or browser tracking consent.
With CAPI correctly implemented:
If you don’t have CAPI, it’s the first technical optimisation to address before attempting to interpret your campaign attribution.
The Facebook Attribution tool was permanently retired in August 2021. It was the standalone panel inside Business Manager that allowed comparison of multi-touch models (last click, linear, positional, data-driven) and cross-channel conversion path analysis.
Meta had stopped onboarding new advertisers from 1 May 2021. The reason was explicit: iOS 14.5 restrictions had made reliable cross-publisher tracking impossible.
What replaced it:
In 2026, Meta conversion measurement rests on three pillars:
There is currently no native Meta tool that allows comparison of multi-touch attribution models across channels. Third-party attribution tools exist for that purpose.
In March 2026, Meta published a significant restructuring of its attribution model under the title “Simplifying Ad Measurement for a Social-First World”. The main changes:
Before 2026, any interaction with an ad (likes, shares, saves, comments) could count as a “click” for attribution purposes. This artificially inflated attributed conversions.
From March 2026, only clicks on the ad’s link count for click-through attribution. Social interactions (likes, shares, etc.) no longer generate attributed conversions.
A new category called engage-through captures:
If someone takes one of these actions and converts within 1 day, the conversion is recorded as engage-through, not click-through.
This change may cause a significant drop in reported click-through conversions — not because your campaign is performing worse, but because Meta is no longer attributing conversions to interactions that weren’t real clicks on your ad. If you see a drop in attributed conversions from March 2026 onwards, check whether this adjustment is the cause before changing strategy.
Meta also announced partnerships with Northbeam and Triple Whale to incorporate Meta view data into their attribution models, reducing the historical gap between Meta’s reported data and third-party platforms.
In April 2025, Meta launched incremental attribution as an optional model within Ads Manager.
Unlike last-click attribution (which assigns all credit to the final touchpoint), incremental attribution uses randomised lift tests to identify conversions that were caused by the ad — not just correlated with it.
Basic mechanism: Meta creates a control group that doesn’t see the ad and compares their conversions with the exposed group. The difference is the incremental conversion attributable to the campaign.
Results from initial tests (January–June 2024): A 20%+ improvement in incremental conversion measurement efficiency across 45 advertisers in 11 industries.
Why it matters: Last-click attribution tends to overestimate the impact of retargeting ads and underestimate prospecting ads. Incremental attribution gives a more honest picture of each campaign type’s true contribution.
Keep the default setting: 7-day click + 1-day view. It’s the balance Meta recommends for most campaign objectives.
Consider using 1-day click for more conservative data. It helps evaluate real ROAS without the noise of conversions that would have happened anyway.
The 7-day window may be insufficient to capture the complete journey. Complement with a third-party attribution tool that can track the journey beyond what Meta can see.
Without the Conversions API, any attribution analysis in Meta is incomplete by definition. CAPI is the prerequisite for any attribution configuration to be reliable.
If you use incremental attribution, keep in mind that retargeting ads tend to show low incrementality — most of those conversions would have happened anyway. This doesn’t mean retargeting isn’t valuable, it means you should evaluate it with different metrics than prospecting.
Meta, like any advertising platform, measures its own performance and scores itself. This creates a permanent discrepancy with Google Analytics, GA4, and third-party tools.
The March 2026 changes (click-through = link clicks only) were explicitly designed to reduce this gap and align Meta’s data with how GA4 counts sessions and conversions.
Even so, Meta’s native attribution is a walled garden: it doesn’t see what happens in other channels (Google, email, organic). For a complete view, the most-used tools in 2026 are:
| Tool | Methodology | Ideal profile |
|---|---|---|
| Triple Whale | Click-based + proprietary pixel | Shopify eCommerce $5M–$50M |
| Northbeam | Probabilistic ML, fractional multi-touch credit | Complex multi-channel, from ~$1,000/month |
| Measured | Lift testing (true causal attribution) | Enterprise, rigorous spend optimisation |
The key structural difference: Meta measures within its own ecosystem. External tools aggregate data across all paid, organic, and direct channels to give a complete ROI picture.
If you want to review how attribution is configured in your Meta Ads account and whether your current windows accurately reflect the impact of your campaigns, book a free audit.
30 minutes to review your situation and tell you exactly what I would change. No pitch, no sales proposal.