- 01crawlup to 500same site, HTML only · 1,000 with deep crawl
- 02judge12 judgments7 scores + 5 probabilities, one Jev call per page
- 03check18 checkstitles, canonicals, structured data, speed, links
- 04report0–100 + gradeissues ranked by pages affected, share link
A score is a starting point, not a ranking.
dotmap checks each fetched HTML page for technical issues, then uses Jev to estimate how well its title, description, content, headings and direct answers serve a reader. The site score averages the analyzed pages, leaving out any marked noindex. A high score means the inspected pages fit this rubric; it does not mean Google has indexed them or that ChatGPT will cite them.
Start with the page list. Open a low-scoring URL and look at the exact issue and recommendation. Fix broken responses, accidental noindex directives and canonical mistakes before polishing prose. Then improve weak pages with a direct answer, useful details and clear headings. Recrawl the same site to see whether the specific pages improved.
Coverage matters. The crawler stays on the same site, follows reachable links and stops at your chosen page limit and depth. A signed-in deep crawl also reads sitemaps and can inspect up to 1,000 pages. Review the uncrawled list before describing a report as complete. Login walls, inaccessible URLs and pages outside the limit cannot be judged from their actual content.
Connect page changes to real outcomes.
For search performance, compare the affected URLs and queries in Google Search Console. For business impact, use analytics or CRM conversions. An audit can show a technical or editorial regression; it cannot prove that the regression caused a drop in traffic or revenue. Keep the dates and measurements together so your team can investigate rather than guess.
If you sign in, dotmap saves site history and lets you explore a scored graph of your crawled pages. That map helps reveal thin sections and missing pathways from the root. Repeat audits make the comparison useful: a new broken page or a large score drop is a prompt to verify the source page, not a substitute for human review.