Investigate the gap before using the number.
A blank record for comparing a GA4 report with an automated report. Establish what differs, retain the evidence and decide the next check.
Click or tab into a bordered field to write. Entries stay in this page; treat reload or closing as losing your edits. Your browser may restore page state.
When you have finished, download a PDF snapshot of this page. Your entries are included directly; no second worksheet or new tab is opened. Keep confidential account data out of copies you share.
Are these the same question?
Use an internal reference without personal or confidential customer details.
Property and report
Property ID, report name and surface: Reports, Explore, Data API or another tool.
Period and time zone
Exact start and end dates; property time zone. Note any relative date setting.
Metric and scope
Exact metric label or API name, and what it counts. Note the scope where relevant.
Dimensions and totals
Breakdowns and dimension scope; record whether comparing a row, sum of rows or overall total.
Filters and settings
Filters, comparisons, segments, reporting identity and attribution settings where relevant.
Freshness and extraction
When generated or retrieved; last refresh; whether the period includes today or recently processed data.
Record what you can observe.
A warning or metadata flag is evidence to investigate. Its presence alone does not establish why these two numbers differ.
Record both values, units, exact row or total, and retrieval time. Leave unavailable values blank; do not replace them with zero.
Record visible sampling or threshold warnings, an (other) row, API metadata, row count, pagination and connector refresh status where available. Note “not available” when you cannot inspect something.
Retain report links, screenshots, exported settings or a sanitised request/response in your approved records. Put references here, not credentials or access tokens.
What could explain the difference? What one setting or step will you check while keeping the rest comparable?
What changed, what stayed the same and what evidence supports the result? State what remains unexplained.
Resolve, qualify or investigate further.
- Fix
- Correct an evidenced query, filter, extraction or presentation issue, then rerun the comparison.
- Recheck
- Set a specific later check when processing, refresh or another temporary condition may matter.
- Qualify
- Keep an understood limitation beside the number and explain its effect on the decision.
- Escalate
- Hand over the reproducible comparison and evidence when the cause is still unresolved.
Write fix, recheck, qualify or escalate. Explain what supports the choice and whether the number is suitable for the intended decision.
Write what the reader needs to know, including unresolved limitations. Do not turn a plausible explanation into a confirmed cause.