How does Meta attribution work?

Some actions happen on the ad itself, such as a video view, a reaction or a link click, and Ads Manager counts them directly. Other actions happen elsewhere, such as a purchase on your website. Meta credits these off-ad actions to an ad when they happen within a set number of days after someone viewed, clicked or engaged with it. The rules for that crediting are the attribution setting, and you choose them for each ad set.

Each ad set also has an attribution model. Standard attribution credits conversions from impressions, clicks or engagements within fixed windows. Incremental attribution uses machine learning models that predict whether an ad caused a conversion, and optimises delivery towards those conversions. Meta's help page also describes custom attribution, which lets an advertiser share attribution data from an external analytics tool so delivery follows that logic.

The choice affects delivery as well as reporting. Meta says the model informs ad delivery, and it warns against comparing ad sets that use different models side by side in the campaign table, because each model counts conversions differently.

Under standard attribution, three kinds of interaction can earn credit for website and in-store conversions. Meta's help page lists the windows it currently supports for each.

Meta's stated default is 7-day click, 1-day view and 1-day engagement. Under that default, a purchase six days after a link click counts, and so does a purchase within a day of someone seeing the ad without clicking. A report set to 1-day click alone would count neither of those purchases. Meta notes that some accounts may still use the earlier click-through and engage-through definitions while the update rolls out.

On 16 October 2025, Meta told developers that from 12 January 2026 the Ads Insights API would stop returning the 7-day view and 28-day view attribution windows. The 1-day click, 7-day click, 28-day click, 1-day engaged view and 1-day view windows stayed available. The same notice limited unique-count and hourly breakdowns to 13 months of history and frequency breakdowns to 6 months, while total values remain available for up to 37 months. Dashboards that pulled 7-day or 28-day view figures through the API stopped receiving them.

On 3 March 2026, Meta announced a change to what counts as a click. For campaigns optimising for website or in-store conversions, click-through attribution now includes only link clicks. Conversions that followed a share, a save or another non-link click moved to engaged-view attribution, which Meta renamed engage-through. Meta also cut the engaged view for video ads from 10 seconds to 5. The rollout began later that month and reached accounts at different times, and Meta said billing would not change.

Meta's stated reason was alignment. Many third-party tools, Google Analytics among them, attribute only to link clicks, so counting likes and saves as clicks made Ads Manager disagree with them. Reports that span early 2026 can mix the old and new definitions, so compare those periods with care.

Meta can only attribute a conversion it hears about, and on websites it hears through two channels. The Meta Pixel is a snippet of JavaScript on your site. It loads a small library of functions that the site calls whenever a visitor takes an action you want to track, such as adding to cart or buying. The Conversions API sends events from your server, website platform, app or CRM (customer relationship management system) to Meta, and Meta processes those server events like pixel events for measurement, reporting and optimisation.

Meta recommends running both, which it calls a redundant setup. The same purchase can then arrive twice, so Meta deduplicates a browser event and a server event when their event IDs and event names match and they arrive within 48 hours of the first. The pixel runs in the visitor's browser, so an ad blocker can stop it firing, which Meta lists as one reason Ads Manager can report fewer conversions than your own records. Server events do not depend on the browser.

Since iOS 14.5, Apple requires apps to ask permission through its App Tracking Transparency framework before they track people or read the device's advertising identifier. Apple defines tracking as linking data from your app with data from other companies' apps, websites or offline properties for targeted advertising or advertising measurement. When someone declines, the advertising identifier reads as all zeros.

Meta's response is Aggregated Event Measurement, a protocol for measuring web and app events from people on iOS 14 and later. Meta says it removes identifiers, adds differential privacy and aggregates data across users. For websites, Meta has since removed the setup it once required. Advertisers no longer prioritise eight conversion events per domain, the Aggregated Event Measurement tab in Events Manager is gone, and domain verification is no longer needed for event configuration. Events sent through the Conversions API may also be processed under its limits.

Where conversion data is missing or partial, Meta may use statistical modelling to account for some conversions, and Ads Manager marks those results with an annotation. Modelled conversions are one more reason Meta's totals differ from a count of orders.

Meta's help pages list the usual causes of mismatched conversion counts. Most come down to each tool using its own rules for which ad or visit gets the credit.

• View-through and engage-through credit. Meta can credit a purchase to an ad the buyer saw and never clicked. Meta says some third-party tools struggle to attribute conversions to impressions, and that Google Analytics attributes only to link clicks.

• Different models and windows. Shopify's marketing reports default to last non-direct click, which gives all the credit to the last channel before the purchase, excluding direct visits. Google Analytics 4 uses its own reporting attribution model with a default lookback window of 90 days for most key events. Meta uses the ad set's attribution setting.

• Cross-device journeys. Meta reports conversions across devices for Meta accounts, while tools that rely on cookies may not connect a click on a phone with a purchase on a laptop.

• Lost referrers. Many tools credit a sale using the referring URL. If someone clicks the ad, leaves, and returns later in a new tab, the referrer is gone and the tool treats the buyer as a different visitor.

• Ad blockers and setup faults. A blocked or misconfigured pixel lowers Meta's count, and duplicate pixel fires raise it.

• Dates and time zones. Ads Reporting can report a conversion on the date it happened or on the date of the impression, and two tools may use different time zones.

Meta's troubleshooting advice is to match attribution windows, time zones and conversion dates across tools, then check raw pixel fires in Events Manager. Tagging ad links with UTM parameters, for example with a UTM builder, lets Shopify and Google Analytics see which visits came from which ad. For what each Ads Manager column means, see what Meta ad metrics measure.