Facebook Ads (Meta) API updates – June 10, 2025: reach and attribution changes

Facebook (Meta) has rolled out important updates to their Marketing API starting June 10, 2025, which affect how Reach and Attribution data are fetched and interpreted. These changes impact how data is loaded and displayed within your Windsor.ai dashboard, especially for users running long-range queries and custom attribution setups.

This document outlines:

• What has changed in the Meta API
• How can these changes affect your existing queries and dashboards
• What steps are required to ensure data integrity and remove discrepancies

As of June 10, 2025, Meta limits Reach data to the past 13 months when used with breakdown dimensions like Age, Gender, or Country. Queries outside this window will fail, and data warehouse lookback settings must be adjusted accordingly. 

This change improves API performance but restricts the depth of historical reporting.

What changed:

• When Reach is queried with breakdowns (like gender, age, placement, country, etc.), Meta now limits the date range to 13 months (394 days).
• Previously, Reach could be fetched for up to 37 months, even with breakdowns.
  

Important notes for Windsor.ai users:

• Dashboards or data loads that include Reach + breakdowns and request data older than 13 months will now return null in the Reach column.
• Scheduled full-refresh jobs with Reach + breakdowns will give incorrect reach data if they span more than 13 months.
  

What you can do:

• Update your custom lookback window to 13 months or less for data warehouse refreshes.
• If you need to get Reach data beyond 13 months, you’ll have to fetch it without any breakdowns. Then, retrieve breakdown data separately and merge the two datasets manually to achieve the desired insights.
• Avoid using Reach alongside the following breakdown fields beyond 394 days:

Gender, Age, Country, Placement, Impression Device, Platform Position,
Publisher Platform, Action Type, Device Platform, Canvas Component Name,
Region, Audience Type, Carousel Card, Dynamic Creative Assets, etc.

One of the most critical changes introduced by Meta on June 10, 2025, is how actions (such as conversions, purchases, leads) are time-attributed in the Marketing API. 

This directly affects how daily, weekly, and monthly performance metrics are visualized and compared within your Windsor.ai dashboard, especially for custom reporting.

What changed:

​​Before the update, users could explicitly define how actions were reported:

• Impression time: Actions were counted on the day the ad was viewed.
• Conversion time: Actions were counted on the day the conversion actually occurred.
• Or a “Mixed” setting for blended reporting.
  

Now, Meta enforces attribution logic automatically:

• On-Meta events (e.g., link clicks, page views, Facebook lead forms) are attributed to impression time.
• Off-Meta events (e.g., website purchases, external signups) are attributed to conversion time.

Important notes for Windsor.ai users:

– Shifts in daily/weekly trends:

• Website purchases and other offsite conversions now show up on the actual conversion day, not when the ad was shown.
• This will alter daily and weekly performance curves, possibly creating the impression of fewer conversions on ad-heavy days. This makes it difficult to compare new data to older periods where attribution may have been impression-based.

– You may now see closer alignment between Windsor.ai and Ads Manager — but only because both are using Meta’s enforced default attribution logic.

However, if your previous Windsor.ai queries relied on custom attribution windows or logic, you’ll see a break in historical comparison consistency.

– Reporting issues:

Tools like Looker Studio/Power BI often rely on simplified aggregations (e.g., monthly or YTD totals). With these attribution changes:

• Fields like Website Purchases may return blank or zero if they aren’t matched properly within the new attribution timing.
• Conversion data may be reported at a different time than when the ads were shown, or vice versa, depending on the type of conversion and the reporting logic used.

What you can do:

• Review your dashboards:
  • Pay attention to conversion date shifts in time-based visualizations.
  • Expect to see metric “spikes” or “gaps” due to changed attribution logic.
• Review time-series visualizations: Be aware that the same conversion may now appear on a different day compared to pre-June 10 data.
• Avoid comparing pre- and post-June 10 data in aggregate unless attribution logic is taken into account.
• Update date filters and explanations on dashboards, especially in executive summaries or client-facing views.
• In Looker Studio, Power BI, or other BI tools, replace broken scorecards or widgets that relied on outdated attribution settings.

By following these tips, users can resolve common data mismatches when fetching data from Facebook Ads via Windsor.ai after June 10, 2025.

If the issue persists, contact our customer support for further assistance.

Please, follow this quick guide to effectively report on your data discrepancy: https://windsor.ai/documentation/how-to-report-data-integrity-mismatch-issues. In this format, our dev team will be able to identify the problem much quicker.

Usually, investigating data discrepancies takes time (1-3 days), so it’s better to organize the report and send it to [email protected]. We’ll take a detailed look and reply to you as soon as we find a resolution.