A record of checks, not a ranking prediction
This scorecard helps prioritize observable problems. It deliberately has no combined “GEO score”: a missing canonical tag, a false service claim and a failed enquiry have different consequences and should not be hidden in one percentage.
Download the audit scorecard (CSV)Each row provides a check, evidence requirement and acceptance method. Complete the affected URL, status, owner and dates. Use “not applicable” only with a reason.
Prioritize by the consequence
| Priority | Use when | Example |
|---|---|---|
| P0 | The intended customer or crawler cannot complete a core task. | An important page is accidentally noindex, or a form reports success without saving. |
| P1 | A confirmed problem affects facts, buying decisions or a key page's usability. | Outdated prices conflict between a service page and the pricing table. |
| P2 | An improvement makes already accessible information easier to use. | A related guide needs a more useful internal link. |
These priorities organize work. They are not Google's ranking factors or a vendor's proprietary score.
Write a change that someone can verify
Replace “improve GEO” with an affected URL, exact issue and acceptance check. For example: “The service page promises a price that differs from Pricing. Confirm the current amount, update the conflicting sentence and compare both pages after deployment.”
Keep the original evidence, proposed change and retest evidence in separate fields. An implemented change can still be awaiting verification. A local test does not prove that the production deployment has the same result.
Close the issue, then monitor the outcome
After deployment, repeat the technical or content check that first identified the issue. Record the date and result. Monitor search impressions, AI observations and enquiries as separate outcomes over an appropriate period.
Use the checklist to find issues and the prompt tracker to record AI observations. A broader audit scope should specify who owns fixes and retesting.