BotView Methodology: 40+ Checks Explained

The visibility score asks two questions: can crawlers reach your content, and is it easy for AI to understand and quote? Here is how both are measured and the 40+ checks we run, in plain English.

What We Measure

BotView measures crawler visibility — can Googlebot and AI crawlers actually see and understand your content? This is different from traditional SEO audits that check meta tag length or keyword density.

We render each page using Playwright with a Googlebot user agent, then compare the initial HTML (what crawlers see first) against the fully rendered DOM (what JavaScript adds later). The gap between these two tells us how much content depends on JavaScript to appear. You can inspect this comparison for any page with the Googlebot view tool.

Reaching a page is only half of it. A page crawlers can fetch but that has no title, a few dozen words or nothing an answer engine can lift is still invisible in practice, so the score also rates content readiness. Keyword density, social tags and other traditional SEO factors are reported but do not affect the score.

What is a GEO score?

A GEO score rates how ready a page is to be quoted by an AI answer engine. BotView's GEO Optimisation Score grades seven on-page dimensions — extractability, Q&A structure, headings, structured data, entity density, citations and freshness — weights them, and returns one 0–100 number per page.

Crawler access asks whether crawlers can reach and render your pages; the GEO score asks whether what they find is shaped so an answer engine can quote it. GEO findings are reported under their own GEO_OPT_ codes, and the GEO score is the main input to the content-readiness half of the visibility score.

Dimension	What it measures	Why answer engines care	Weight
Answer extractability	Short, self-contained paragraphs (40–500 characters), bullet lists and tables — versus walls of text over 1,200 characters.	A passage that makes one point on its own is easier for a reader or an engine to quote accurately. Google says chunking is not required for its AI features, so this rewards clarity, not a format.	25
Q&A / FAQ structure	Question-form H2/H3 headings, FAQPage or QAPage JSON-LD, definition lists. Three or more question headings, or FAQ markup for a Q&A section, score full marks.	People ask engines questions. A page that states the question and answers it right below gives them a passage that matches.	15
Heading structure	One descriptive H1 plus H2 sections. Points come off for a missing H1, multiple H1s, a thin H1, or no H2s at all.	A clear outline tells readers and crawlers what each section covers.	15
Structured data	Presence and type of JSON-LD. Answer-friendly types (Article, FAQPage, HowTo, Recipe) score highest; supporting types (Organization, WebSite, BreadcrumbList) add a little.	Makes visible facts machine-readable and can qualify pages for Google rich results. Google says its AI features need no special schema.	15
Entity & factual density	Heuristic ratio of named entities and figures to word count, with a bonus for five or more figures. Pages under 150 words score low.	AI answers prefer sources with specific, checkable substance over vague copy.	10
Citations & sources	Distinct external domains you link out to, ignoring your own domain and social/share hosts. Three or more scores 90.	Links to primary sources let readers, and engines, check your claims.	10
Freshness	Whether we can find a machine-readable date (fresh / stale / unknown). We report whether the domain serves an /llms.txt but do not score it, because Google Search ignores it.	A date tells readers and engines how current the information is.	10

Each dimension is scored 0–100, multiplied by its weight, and summed to the page score. Anything below 70 becomes a fix in the report's Issues and Fixes tabs. Multi-page scans average each dimension across the pages crawled.

What this score is not. The seven dimensions and their weights are BotView's heuristics, not ranking factors published by OpenAI, Google, Anthropic or Perplexity, and we have not calibrated the score against observed citations. A GEO score of 80 is not an 80% chance of being cited. Google says its AI features need no llms.txt, special schema or chunking (Google's guide). Use a low dimension as a prompt to check the page, and fix blocked crawlers and wrong facts first.

What is a GEO score?

A GEO score measures how well a web page is structured to be quoted by AI answer engines. BotView's GEO Optimisation Score rates seven on-page dimensions — answer extractability, Q&A structure, heading structure, structured data, entity density, citations and freshness — and combines them into one 0–100 score per page.

How is a GEO score different from an AI visibility score?

Crawler access asks whether OAI-SearchBot, Claude-SearchBot, PerplexityBot, Googlebot and the rest can reach and render your pages. A GEO score measures answer-readiness: once they are in, is the content shaped so it can be quoted? BotView reports both, and since October 14, 2026 the visibility score combines them: 45% access, 55% content readiness (the GEO score minus missing basics such as a title or H1).

How the Score is Calculated

Each page gets two numbers. Crawler access starts at 100 and loses points for issues that stop crawlers seeing the content (below). Content readiness is the page's GEO score minus up to 40 points for missing basics: no title (15), thin content (12), no H1 (8), no meta description (6) and smaller ones such as a missing canonical or viewport.

The page's visibility score is 45% access plus 55% content readiness, then capped where good content can't make up for the problem:

Under 100 words of content: at most 15 for an empty page, rising to 40 at 100 words. Under 300 words: at most 70.
A JavaScript challenge or an empty single-page-app shell: at most 60.
A noindex directive: at most 20.
AI search crawlers (OAI-SearchBot, Claude-SearchBot, PerplexityBot) blocked in robots.txt: the site score is capped at 70, 55 or 40 for one, two or three blocked; all AI crawlers blocked caps it at 30.

Crawler access deductions by severity:

-15

Critical

Content is invisible or inaccessible to crawlers. Bot walls, heavy JS dependency, error status codes.

-5

Warning

Content is partially visible but degraded. Moderate JS dependency, slow rendering, blocked resources.

-1

Info

Minor visibility concerns. Small rendering differences, missing alt text on decorative images.

Multi-Page Scoring

When scanning multiple pages, the overall site score uses a weighted average: 70% straight average across all pages, 30% weighted toward the lowest-scoring pages. This means your worst pages pull the score down more than your best pages pull it up — because a single broken page can hurt your site's crawl visibility disproportionately.

The Five Visibility Groups

Every visibility issue falls into one of five groups. Each group gets its own sub-score in the report.

1. Rendering

Can crawlers see your content without executing JavaScript?

JS content dependencySPA shell detectionClient-side renderingLoading states in HTMLJS-rendered headings & links

2. Crawler Access

Are crawlers allowed to reach your pages and their resources?

AI crawler blocking (robots.txt)Blocked JavaScript/CSS/fontsBot walls & WAFsHTTPS availability

3. Indexability

Will search engines actually add your page to their index?

noindex directivesSoft 404 detectionError status codesRedirect chainsHreflang errors

4. Performance

Can crawlers finish rendering your page within their time budget?

Page load timeDOM ready timingRender budget exceededFirst & Largest Contentful Paint

5. Content Visibility

Is your important content actually visible in the DOM?

Hidden content (display:none)Modal/dialog contentInfinite scrollMissing internal linksEmpty anchor text

Score Grades

90–100

Excellent

Crawlers reach it and the content is easy to quote

70–89

Good

Reachable, with some content gaps

50–69

Needs Work

Significant access or content gaps

0–49

Poor

Blocked, thin or hard to read

What Doesn't Affect the Score

BotView reports on many traditional SEO factors, but these are informational only. They appear in your report but don't change your visibility score:

Open Graph / social tags
Image file formats
Keyword density or usage
Blocking training crawlers such as GPTBot or ClaudeBot (a policy choice)
Bot challenges our scanner hits from a datacenter IP (shown as unverified)

Smaller basics such as title length or a missing canonical cost a point or two of content readiness; they never outweigh access or content problems.

See Your Score

Run a free scan to see exactly what Google and AI crawlers can see on your website.

Free scan, no signup required