A blank workbook using the same severity and verification labels as my client reports. One row per finding, and every row has to prove itself.
On this page
This website audit report template uses the same severity and verification labels as the reports in my case studies, with no client detail in it. It is a spreadsheet, not a PDF, because a report you cannot sort and filter is a report nobody works through.
The format matters less than one column in it: how you know each finding is true.
Confirmed means reproduced by hand against the live site. Likely means strong evidence that was not independently reproduced. Unverified means a tool reported it: it stays visible, but it is not asserted.
Automated audit tools, including my own systems, report things that are not true. A report without this column asks the reader to trust every row equally. With it, they know exactly which rows to act on first.
Write the How to check column before the fix. If you cannot write down how to reproduce a finding, it is not ready to go in the report yet.
Sort by severity, then by confidence. A Critical finding marked Unverified is a reason to look harder, not a reason to act.
The template is the format. If you would rather have the findings, send a URL: three to five real ones come back within two working days, free, filled in exactly this way.
Real engagements, described by sector only. Each includes the report and the workbook.
Full-site technical audit, then an organic growth plan and a topical-authority cluster architecture, with a forecast model built on leading indicators rather than invented targets.
Three automated warnings were rejected before publishing — including a 'duplicate title on every page' that turned out to be the titles inside SVG icons.
Audit delivered; no live-site changes recorded.
Full-site technical, on-page, content, schema and performance audit, with every content page crawled and a random sample of product pages checked.
One apparent crawler block turned out to be rate-limiting caused by the audit itself, and was rejected before it reached the report.
Audit delivered; implementation not recorded.
Mine runs in two passes. The first is an agentic audit of the whole site through systems I designed: indexation, canonicals, sitemaps, structured data, performance, images, AI crawler access, and the copy itself checked for factual errors.
The second pass is the one that matters. Every finding is re-tested by hand against the live site and labelled Confirmed, Likely or Unverified before it goes in the report, with a check you can run yourself.
The tool matters less than checking what it tells you. I work from Search Console, Analytics, crawl data and audit systems I designed, and I treat every report, including my own systems' output, as a claim to verify rather than a finding.
Send a URL. Three to five verified findings come back within two working days, free, with no call required.
Ask my AI
Answers from notes I wrote. It shows its sources and says when it does not know.
Ask about the case studies, how I would approach your site, or what actually goes into an audit. Every answer shows the note it came from.
AI-generated from my own notes. Verify anything that matters.