Meta Just Removed Your Attribution Windows. Here's What Broke.

What Happened on January 12

Meta permanently removed two attribution windows from the Ads Insights API:

What remains:

The API parameters action_attribution_windows=7d_view and action_attribution_windows=28d_view simply stopped returning data. No error message. No fallback. Just empty results.

Three days later—on January 15—Meta also stopped delivering campaigns that used targeting interests deprecated in June 2025. If your campaigns weren't updated, they went dark without warning.

Why This Matters More Than You Think

This isn't a minor API tweak. For many advertisers, it broke reporting, erased historical baselines, and set off false alarms about campaign performance.

30-40% of Conversions Just Disappeared

For advertisers with longer consideration cycles, view-through conversions between days 2 and 28 represented a significant share of reported performance:

Historical Data Is Gone Too

This isn't just a forward-looking change. Historical reports that queried 7-day or 28-day view windows can no longer be regenerated. If you were tracking trends using these windows, your historical baselines are broken.

Budget Decisions Based on Incomplete Data

If your media team optimizes based on Meta-reported ROAS, the numbers just changed underneath them:

The actual performance of your campaigns is identical. Only Meta's ability to claim credit changed. But if someone is watching these numbers without context, they'll conclude Meta performance collapsed—and potentially cut budget from campaigns that are working fine.

The Pattern: Platforms Keep Changing the Rules

This is not an isolated event. It's part of a pattern:

Every time a platform changes its attribution rules, advertisers lose data and have to recalibrate. This is the fundamental problem with platform-controlled measurement: you don't own the ruler.

Google decides how to credit. Meta decides what to count. And they can change the definitions at any time—without your input, often without your awareness.

What to Do Now

1. Don't Panic-Cut Meta Budget

The most important thing: separate measurement changes from performance changes.

If your Meta campaigns were profitable before January 12, they're almost certainly still profitable. The conversions are still happening—Meta just can't see them anymore.

Before making any budget decisions:
- Compare conversion volume in your own analytics (not Meta's)
- Check your actual revenue numbers for the same period
- Look at CRM data for lead quality and pipeline
- If nothing changed on your end, the campaigns are fine

2. Update Your Reporting Baselines

If you were tracking Meta performance using view-through windows, your year-over-year comparisons are broken. You need new baselines:

3. Switch to Independent Attribution

Platform attribution has a structural problem: the platform measuring your ads is the same platform selling you ads. When they change the rules—and they will keep changing them—your data breaks.

Independent attribution solves this by:

1. Tracking touchpoints yourself — Server-side, across all channels
2. Applying consistent models — You choose the attribution logic, not the platform
3. Owning your data — No one can remove your attribution windows
4. Seeing the full journey — Not just what one platform can see

4. Measure View-Through Independently

View-through conversions are real. Someone sees an ad, doesn't click, and converts later. That's a legitimate touchpoint. Meta just can't track it beyond 1 day anymore.

Server-side tracking can capture these journeys independently:

1. Track ad impressions via the Meta Conversions API
2. Match to downstream conversions in your own system
3. Apply your own attribution window (7 days, 28 days, whatever fits your business)
4. No dependency on Meta's API limitations

This is more work than relying on Meta's numbers. But it's also more accurate, more durable, and entirely under your control.

5. Run Incrementality Tests

If you're unsure whether Meta is actually driving value (now that the numbers look different), test it directly:

Incrementality tests are immune to platform attribution changes because they measure real-world outcomes, not platform-reported credit.

The Bigger Picture

Meta's attribution window removal is a symptom of a larger shift: platforms are retreating from measurement.

• Apple made tracking opt-in (iOS 14.5)
• Google removed attribution models from GA4
• Meta is shrinking attribution windows
• Chrome's Privacy Sandbox limits cross-site tracking

Each change reduces what platforms can tell you about your marketing. The companies that thrive in this environment are the ones building their own measurement infrastructure—first-party data, server-side tracking, independent attribution.

The companies that keep relying on platform-reported numbers will keep getting surprised when those numbers change.

Summary

On January 12, 2026, Meta removed 7-day and 28-day view-through attribution windows from the Ads Insights API.

What broke:
- 30-40% of view-through conversions are no longer reported
- Historical data using these windows is gone
- Year-over-year baselines are broken
- ROAS appears to drop even though actual performance hasn't changed

What to do:

The lesson: Every time you trust a platform to measure itself, you accept the risk that the rules change without warning. The only durable measurement is the measurement you control.

Further Reading

On Platform Attribution Problems:
- Why Your Platform Reports Don't Match — Cross-platform reconciliation
- Why Did GA4 Remove Attribution Models? — Google's parallel move

On Building Independent Measurement:
- What is Multi-Touch Attribution? — The complete MTA guide
- Server-Side vs Client-Side Tracking — Why server-side captures more

Sources:
- Meta Ads Attribution Window Changes (Dataslayer)
- Facebook Ads: Attribution Window Removals (Supermetrics)
- Meta Restricts Attribution Windows (PPC Land)

Holly Henderson

Co-Founder, mbuzz

Holly Henderson is Co-Founder of mbuzz. With 10+ years in marketing including roles at Westpac, Avon, and Forebrite, she's obsessed with making measurement actually useful.

Harvard Extension School Forebrite Westpac Avon