eComStrategicsTalk with Dan

Methodology 6.1.0

See exactly how the audit reaches its score.

The audit checks the initial server response for the submitted URL, then adds a bounded sample of representative pages from verified sitemap documents. Recurring metadata, page-health, heading, image, and discovery problems can therefore affect the result instead of hiding behind a clean homepage.

Run the audit

Scoring model

The number cannot outrun the evidence.

Verified on-site readiness is earned points divided by all applicable possible points. Tested quality is earned points divided by tested points. Evidence coverage is tested points divided by applicable possible points. A check that was not collected remains unverified in the readiness denominator; it is never silently converted into a pass or mislabeled as a confirmed defect. Genuine advisories and non-applicable checks do not affect the score.

Confidence combines evidence coverage and source quality. A numeric readiness score is shown only when coverage reaches 60% and confidence reaches 60%. The published weights are this audit's prioritization model—not Google, Bing, or an AI platform's private ranking formula.

Discovery & eligibility, 43%

Methodology weight for verified access, indexing, snippet, canonical, crawler, and sampled site-health evidence.

Page clarity & differentiation, 40%

Methodology weight for page-level topic clarity, useful links, metadata, and sampled sitewide differentiation.

Experience & machine clarity, 17%

Methodology weight for image alternatives, landmarks, applicable answer and entity evidence, and sampled accessibility patterns.

Site pattern scan

A bounded cross-page sample completes more of the score.

The audit inspects up to 32 representative same-origin pages discovered in verified sitemap documents, allowing modest sites to be sampled in full when time permits. It reports and scores verified cross-page patterns involving response health, canonicals, title and description duplication, topic differentiation, internal discovery, and missing image alt attributes. Prevalence determines partial credit, so a recurring template problem matters more than one isolated page.

Every stage shares a fixed time budget. If a source is slow or unavailable, completed evidence is returned as a partial result and the gap is named. A sample is an efficient pattern detector, not a promise that every URL on the site was crawled.

Performance and platform verification

Volatile and account-owned evidence is intentionally checked at its source.

PageSpeed results can change between runs, and Search Console or Bing indexing data is only available to verified site owners. Those sources are valuable, but combining them with a stable technical-readiness score would create false precision. Each result therefore links to the appropriate external tool instead.

External results do not raise or lower the readiness score.

Deliberate exclusions

The audit does not invent AEO or GEO penalties.

Google says its generative search features use the same foundational Search systems, require no special AI markup, and do not guarantee crawling, indexing, rankings, or citation inclusion. For that reason, the following items may appear as contextual guidance when relevant, but their absence alone does not lower readiness:

  • FAQ schema or a generic FAQ section on every page.
  • llms.txt, AI text files, artificial answer chunks, or special AI markup.
  • Exact title, description, heading, or word-count targets.
  • Exactly one H1 when the page still communicates one clear primary topic.
  • Schema quantity without relevance, accuracy, or agreement with visible content.
  • PageSpeed availability, social metadata, or optional non-search crawler policy.

Core score checks

The same weighted checks power every score.

8 weighted pointsSuccessful page response

The final audited URL returns a successful HTTP response.

10 weighted pointsIndexing is permitted

HTML and HTTP robots directives do not contain noindex.

5 weighted pointsSearch snippets are permitted

HTML and HTTP robots directives do not restrict Google or Bing text snippets; Google also applies these controls to AI-feature previews.

4 weighted pointsCanonical is valid

A declared canonical is singular, parseable, and same-origin. Its absence is advisory, not a generic failure.

8 weighted pointsGoogle and Bing crawlers are allowed

robots.txt permits Googlebot and Bingbot. Third-party AI search crawler choices are reported separately.

AdvisorySitemap is discoverable

Advisory: a sitemap can support discovery and site inventory but does not guarantee indexing or ranking.

AdvisoryCrawler policy is explicit

Advisory: document an intentional robots.txt policy. A missing file does not block crawlers.

AdvisoryAI and non-search crawler controls

Advisory: AI search, training, model-control, and user-initiated agents are reported separately and never change Google/Bing eligibility.

5 weighted pointsDescriptive page title

The initial HTML contains a non-empty page title.

3 weighted pointsMeta description context

The initial HTML contains page-specific snippet context. This is a low-weight presentation signal, not an indexing blocker.

5 weighted pointsClear primary topic

The visible main content contains at least one non-empty H1. Multiple H1 elements are not an automatic failure.

2 weighted pointsExtractable answer blocks

Advisory: question-led sections may help when they fit the page's intent; they are not a universal requirement.

5 weighted pointsCrawlable internal paths

The visible page content exposes at least one normal internal link to another URL.

4 weighted pointsInternal link purpose is clear

Crawlable internal links have accessible, destination-specific names rather than unnamed or generic labels.

AdvisoryContent structure review

Advisory: descriptive subheadings can aid readers, but no universal word or heading count is scored.

2 weighted pointsPresent JSON-LD parses

JSON-LD that is present parses successfully. A page without JSON-LD is not automatically defective.

3 weighted pointsPrimary entity is identified

Advisory: use page-specific entity schema only where it accurately matches visible content and an eligible feature.

1 weighted pointsEntity identifiers are stable

Advisory: stable @id and carefully chosen sameAs values can clarify entities but are not universal requirements.

3 weighted pointsVisible images declare text alternatives

Visible, non-presentational images include an alt attribute; an empty alt is accepted for decorative images.

2 weighted pointsOne visible main landmark

The document exposes one visible native main element or ARIA main landmark for readers and assistive technology.

4 weighted pointsSampled site page health

A representative site sample is checked for recurring response, crawler-access, indexability, and snippet-eligibility failures.

4 weighted pointsSampled canonical consistency

Canonical declarations are compared across a representative site sample.

8 weighted pointsSampled title uniqueness

Titles are checked across sampled pages for missing values and recurring duplication.

4 weighted pointsSampled description uniqueness

Meta descriptions are checked across sampled pages for missing values and recurring duplication.

3 weighted pointsSampled topic differentiation

Primary page topics are compared across sampled pages for distinct intent.

3 weighted pointsSampled internal discovery

A representative site sample is checked for pages that are difficult to reach internally.

4 weighted pointsSampled image alternative coverage

Image alternative patterns are compared across a representative site sample.

Crawler taxonomy

Search crawlers have different jobs.

Googlebot and Bingbot affect the scored discovery-access check. AI search crawlers, training agents, model-control agents, and user-requested agents appear separately for context but do not affect the score.

Scored search crawlerGooglebot

Google Search crawling, including content eligible for Google's AI search features.

Scored search crawlerBingbot

Bing search discovery and indexing.

Advisory: searchOAI-SearchBot

Surfaces websites in ChatGPT search results.

Advisory: searchClaude-SearchBot

Search crawling used to improve relevance for Claude users.

Advisory: searchPerplexityBot

Search indexing used to answer Perplexity queries.

Advisory: trainingGPTBot

OpenAI model-training control; it is not the ChatGPT search crawler.

Advisory: trainingClaudeBot

Anthropic model-training control; it is separate from Claude-SearchBot.

Advisory: model controlGoogle-Extended

A standalone control token for Gemini training and grounding; it does not control Google Search inclusion.

Advisory: user initiatedChatGPT-User

A fetch initiated by a ChatGPT user or agent action.

Advisory: user initiatedClaude-User

A fetch initiated by a Claude user request.

When robots.txt is missing, the audit reports that no restriction was found. HTTP 429, authentication responses, server errors, and fetch failures are marked as not tested instead of being guessed.

Scope

What this audit can and cannot establish

  • The scan reads public server-delivered evidence for the submitted URL and a bounded set of same-origin sitemap pages.
  • Missing or inaccessible evidence lowers coverage; it is never counted as a pass.
  • Rankings, traffic, conversions, and live AI citations require separate measurement.
  • Recommendations need human review before implementation on a production site.

Data handling

See your score without sharing contact details.

The public scan only needs a URL. If you request the detailed roadmap, the audit and contact details you submit are stored in the eComStrategics database so Dan can follow up, and Resend handles the requested roadmap email. Standard hosting and service logs may still apply. See the privacy notice.

Primary documentation

Sources used for the methodology