METHODOLOGY

We show how we measure.

What we inspect, how we calculate scores, and where the limits are.

Technical readiness score

Seven rules are applied to the initial HTML of the requested page. Each passed check earns its full weight; a failed check earns zero. The sum is 0–100. These weights are DoctorGeoLab’s own checklist, not an AI platform’s official ranking formula.

CheckPassing conditionPoints
Secure connectionHTTPS10
Page title10–90 characters15
Meta description50–200 characters15
Main headingExactly one H110
ContentAt least 250 words20
JSON-LDParseable data with @context and @type20
CanonicalA canonical link tag10

30 technical checks with a clear scope

Sixteen additional diagnostics cover indexing directives, viewport, language, encoding, favicon, sharing tags, H2 headings, internal links, image attributes and dimensions, HTTPS resources, canonical matching and sitemap presence. Seven further checks cover HTTP response, JSON-LD syntax, Organization / LocalBusiness, ContactPoint, WebSite, @id and sameAs. These diagnostics do not change the score.

Sitemap inspection is limited to two child sitemaps and 1,000 URLs, with up to 40 URLs displayed. The sitemap list is not a content audit; a separate crawl inspects initial HTML from up to 20 same-origin pages. The report is available as a PDF summarizing technical findings and priority fixes, with up to six pages including the crawl appendix.

Crawler access is a separate check.

The checker inspects robot groups and Allow / Disallow rules. It uses the most specific matching agent group and longest matching path, with Allow winning ties. A 404/410 robots.txt is treated as no robots restriction; other access failures are unknown.

Training crawlers and search crawlers are not the same. GPTBot and OAI-SearchBot, for example, serve different purposes. User-initiated tools such as ChatGPT-User may not apply robots.txt rules. This check does not prove an actual visit, indexing or recommendation.

What isn’t measured?

When the selected OpenAI or Perplexity provider is configured, Answer test retrieves a web-backed API answer to one customer question. Brand names are matched after the answer, never sent to the model. This sample is not an exact consumer app result or recommendation ranking. Workspace mention and citation rates cover only saved answers. Gemini is pending usage review. Weekly email briefing infrastructure is configured; reports require account tracking and email consent. Configuration does not prove that a report was delivered.

The scan does not execute JavaScript or crawl the full website. The JSON-LD check is not a full Schema.org validator. Word count alone does not establish content quality or accuracy.