RANK OUTLAW · RESEARCH WORKSHEET

Compare the measurement before the percentage

Use this worksheet to decide whether two AI-search statistics support a comparison. It is an editorial aid, not a ranking or revenue prediction model.

Save this HTML page for an offline copy. Enter your observations below, then print or save as PDF to preserve your answers. Entries are not automatically saved or sent anywhere.

Completed example: 14% versus 87.3%

Check	14% observation	87.3% observation
Surface	Google AI Mode	Google AI Mode
Unit	Exact URL	URL links
Ranking range	Google organic top 10	Google organic top 10
Citation type	Analyzed AI Mode sources	Links inside AIM SERP blocks only
Sample / design	Average across 9,721 keywords; US, non-logged-in collection	310 links in AIM SERP blocks; these appeared in 79 of 9,734 responses
Collection date	June 2025	June 2025
Calculation / denominator	Average URL overlap: 14%. The cited page does not specify the overlap formula’s denominator.	Coverage within AIM SERP blocks: 87.3%. The category contains 310 links, not the full set of 122,617 links.
Capture method	SE Ranking study; verify its original methods before reuse	Same study, restricted link category

Source: SE Ranking’s AI Mode study, June 2025 collection.

Decision: different measurements. The citation category and population differ. The missing overlap denominator also prevents a fully specified comparison. Do not treat 87.3% as a correction of 14%, calculate growth between them, or infer a performance change.

Your comparison

Observation ASource URL and reported percentage

Surface

Unit: URL, domain, or another unit

Ranking range

Citation type

Sample: population, design, and size

Collection date

Denominator and aggregation formula

Capture / parser method

Location, account state, and other limitations

Observation BSource URL and reported percentage

Surface

Unit: URL, domain, or another unit

Ranking range

Citation type

Sample: population, design, and size

Collection date

Denominator and aggregation formula

Capture / parser method

Location, account state, and other limitations

Choose a defensible conclusion

Missing definitions or calculation methods: leave the comparison unresolved.
Incompatible metrics, populations, sampling, or capture: label different measurements.
Comparable methods with different dates or sampled records: describe a trend cautiously, noting uncertainty.
Passing these checks supports only the measured difference. It does not establish causation, rankings, leads, or revenue.

Decision and supporting reason

Next action and evidence needed

Check verified eligibility problems first, fill specific answer gaps second, and gather comparable observations when no page defect is established.