Attribution Windows

What This Page Answers

Attribution windows define how long after an ad interaction a platform is allowed to claim conversion credit. They can change reported performance dramatically without changing actual business results.

Basic Definition

An attribution window has two parts:

  • Interaction type: click, view, engaged view, or another platform-specific interaction.

  • Time window: how many hours or days after that interaction the conversion can be counted.

Example:

7-day click, 1-day view

This means a platform can report a conversion if the person clicked within the last seven days or viewed within the last one day, depending on platform rules.

Why Attribution Windows Matter

Attribution windows affect:

  • Reported ROAS

  • Reported CPA

  • Campaign comparisons

  • Retargeting performance

  • Budget allocation

  • Executive confidence

  • Incrementality interpretation

A wider window usually reports more conversions. That does not mean the campaign caused more conversions.

Click vs View Windows

Click-through attribution gives credit after a user clicks an ad and later converts. View-through attribution gives credit after a user sees an ad but does not click, then converts later. View-through credit can be useful for upper-funnel and video campaigns, but it can also inflate platform performance, especially in retargeting or high-awareness audiences.

Engaged-View Windows

Engaged-view attribution is common in video environments where watching a meaningful portion of a video indicates stronger engagement than a simple impression. It sits between view-through and click-through attribution:

  • More intentional than a passive impression.

  • Less explicit than a click.

  • Useful for video-heavy platforms like TikTok and YouTube.

Platform Differences

Meta, Google, TikTok, and ChatGPT Ads do not use identical attribution logic. Key differences can include:

  • Default attribution windows.

  • Whether view-through credit is included.

  • Whether engaged-view conversions exist.

  • Whether modeled conversions are included.

  • Whether reporting is click-date or conversion-date based.

  • How conversions are deduplicated across campaigns.

Never compare platform ROAS without checking attribution settings.

Historical Window Changes

Attribution settings also change over time. A historical report can look stronger simply because the platform used a wider attribution window at the time.

Meta used to support a 28-day click attribution window. After Apple privacy changes around iOS 14.5 in 2021, Meta reporting moved away from that long click window for standard campaign reporting. If you compare current Meta performance to older pre-2021 reporting that used 28-day click attribution, current numbers can look worse even when business results are similar.

The reason is mechanical: a shorter window gives the platform fewer days to claim conversion credit. Fewer attributed conversions lowers reported ROAS and raises reported CPA, even if backend revenue did not change.

When reviewing old Meta benchmarks, always ask:

  1. Which attribution window was used?

  2. Was view-through credit included?

  3. Was the report built before or after the iOS 14.5-era reporting change?

  4. Are you comparing platform-reported results or backend business results?

Practical Example

A campaign reports:

  • 7-day click ROAS: 3.5

  • 1-day click ROAS: 2.1

  • 1-day view ROAS: 4.8

The campaign did not become more profitable across three realities. The reporting lens changed. The difference tells you something about the role of the campaign:

  • Strong 1-day click performance suggests direct response.

  • Strong 7-day click but weak 1-day click suggests longer consideration.

  • Large view-through lift suggests influence or possible over-credit.

How To Standardize Reporting

For cross-channel reporting:

  1. Record each platform's default attribution window.

  2. Choose one internal reporting standard for weekly decisions.

  3. Keep platform-native reporting for optimization diagnostics.

  4. Use blended MER to sanity-check total performance.

  5. Use incrementality tests when platform attribution is strategically important.

  6. Avoid switching windows mid-test unless you clearly mark the change.

Common Mistakes

  • Comparing Meta 7-day click, 1-day view to Google last-click reporting as if they are equivalent.

  • Scaling retargeting based on wide view-through windows.

  • Treating modeled conversions as directly observed conversions.

  • Changing attribution windows and interpreting the result as a performance change.

  • Reporting only the window that makes performance look best.

Practical Rule

Attribution windows are reporting assumptions. They are not proof of causality.

Source Notes