Attribution Window Comparison Check

Separate a change in attribution settings from a change in campaign performance. The download provides a specific review process. It does not make decisions or act in an account.

What to provide

Two dated campaign reports, conversion definitions, attribution models, lookback windows and timezone.

How it works

1. Record model, window, event definition and export time for each report.
2. Align event and reporting periods. Do not compare conversion-date and click-date totals as equivalent.
3. Keep different windows separate. Calculate change only for a like-for-like pair.
4. Return comparable rows and setting gaps. Do not change analytics settings or claim causal lift.

The output

report | period | model/window | event | comparable pair | observed change | unknown

Worked fictional example

Fictional report A has 100 events in a 7-day window; B has 120 in a 28-day window. The difference of 20 is not proven growth. Matched 7-day B of 105 shows an observed increase of 5, not causal lift.

Checks and limits

Preserve source versions and original values. Unknowns stay unknown. Use a qualified reviewer for regulated, legal or hiring decisions. No send, share, account change, schedule, ledger update or payment is authorized. No live account test is claimed.

More marketing workflows · How-to guides

Method: reviewed source. Demand lead: Oct3liveGSCreportgeneration11impressions andmarketingcopilot37; exactintentvolumeunknown. These are site-query cluster impressions and method evidence, not exact new-topic monthly search volume or a traffic promise. Checked October 3, 2026.