Every report I send, free snapshot or full review, scores your website the same way, against the same checklist. This page publishes that checklist, so you can see exactly what earned each point, check my work, and compare one review with the next.
The score measures how ready your website is to be found, understood, and recommended by Google and by AI assistants such as ChatGPT, Gemini, Perplexity, and Claude. It covers what you control on your own site and your presence around the web. It doesn’t measure your traffic or sales; those are reported separately when you share access to your analytics.
In short:
- Six areas, each scored out of 100, weighted into one readiness score.
- A fatal problem, such as a page search engines can’t read, caps the score at 60 no matter how good everything else is.
- Most checks are machine-measured and repeatable; the rest are judged by me, and the report labels which is which.
- I don’t promise rankings or AI mentions. I promise specific findings, specific fixes, and a re-score that shows what changed.
The scores you’ll see
| Score | What it answers | How it’s built |
|---|---|---|
| Overall readiness | How ready is the whole site? | All six areas, weighted (below) |
| Google search (SEO) | Can Google reach, read, trust, and rank it? | Reach, technical health, content, business details |
| Answer engines (AEO) | Can Google’s AI Overviews and similar features pull clear answers from it? | Content, clear answers, business details |
| AI assistants (GEO) | Will ChatGPT, Gemini, Perplexity, and Claude name and recommend you? | Reach, clear answers, business details, reputation |
| Accessibility and writing | Can everyone use it, and is it free of errors? | Pages passing an automated accessibility screening, and confirmed spelling errors |
The three lens scores (SEO, AEO, GEO) use the same area weights as overall readiness, counting only the areas that matter for that lens.
Accessibility is reported as a count of pages that pass, not a 0 to 100 score. Automated checks catch only part of what the accessibility guidelines (WCAG 2.1 AA) require, and a single number would suggest a level of compliance no automated tool can confirm. It isn’t a legal review.
The six areas
| Area | Weight | The question it answers |
|---|---|---|
| 1. Can search engines reach the site | 20% | Can Google, Bing, and AI crawlers get to every page that matters? |
| 2. Technical health | 15% | Is the content readable without running JavaScript, and is the site fast, secure, and mobile-friendly? |
| 3. Content and page basics | 20% | Does each page clearly say what it is, answer what buyers ask, and show real proof? |
| 4. Clear, quotable answers | 15% | Can an AI system lift a correct, self-contained answer from your pages? |
| 5. Business details in code | 15% | Does the code tell machines who you are, where you are, and what you offer, accurately? |
| 6. Reputation beyond your site | 15% | Do reviews, directories, articles, and profiles vouch for you? |
Overall readiness = (area 1 × 20 + area 2 × 15 + area 3 × 20 + area 4 × 15 + area 5 × 15 + area 6 × 15) ÷ 100.
Each area is worth 100 points, split across specific checks. A check earns full, half, or no points. Checks that look at pages are run across a sample of the site, weighted toward the pages that bring in business: a check passes only when those pages pass and about 80% of the sampled pages do.
1. Can search engines reach the site (20%)
| Check | Points |
|---|---|
| robots.txt loads, doesn’t block valuable pages, and lists the sitemap | 15 |
| Google and Bing can reach every page that brings in business | 15 |
| AI search crawlers (ChatGPT, Claude, Perplexity) are allowed, or blocked on purpose | 15 |
| Firewalls and bot protection don’t challenge search or AI crawlers | 10 |
| Key pages load normally; old pages redirect in one step | 10 |
| A made-up address returns “not found,” not a normal page | 5 |
| Each page names one preferred address (canonical) that agrees with the sitemap | 15 |
| No accidental “don’t index” or “don’t show snippets” settings | 10 |
| The sitemap lists only real, indexable pages and is submitted to Google and Bing | 5 |
2. Technical health (15%)
| Check | Points |
|---|---|
| Main content, navigation, prices, and proof are in the HTML before any JavaScript runs | 30 |
| Real-visitor speed data from Google is “good” | 20 |
| Works well on phones: readable text, tappable buttons, no blocking pop-ups | 15 |
| Secure (HTTPS) everywhere | 10 |
| Links are real links, and key pages are within three clicks of the homepage | 5 |
| Key pages get links from at least three other pages’ body text | 10 |
| Language versions are tagged correctly (full points if there’s only one language) | 10 |
3. Content and page basics (20%)
| Check | Points |
|---|---|
| Every page has its own clear title and description | 10 |
| One main heading per page, matching the title, with a logical outline | 10 |
| The type of page matches what people searching for it want | 15 |
| Covers the questions buyers ask, as fully as the pages that outrank it | 15 |
| Shows first-hand proof: results, real screenshots, original numbers, lessons learned | 15 |
| Named authors or experts with relevant credentials | 15 |
| Dates are visible and facts are current | 10 |
| No contradicting facts or near-duplicate pages competing with each other | 5 |
| No confirmed spelling errors in titles, headings, or buttons | 5 |
4. Clear, quotable answers (15%)
| Check | Points |
|---|---|
| The main question is answered near the top of the page | 20 |
| Each section opens by naming its subject and giving the answer | 20 |
| Headings match how people actually ask | 10 |
| Comparisons are real tables; steps are numbered lists | 10 |
| Facts are specific and sourced: numbers, units, dates, links | 20 |
| The topic is covered from definition to cost to alternatives, inside useful pages (not dozens of thin ones) | 10 |
| Videos have transcripts; key facts aren’t locked in images or PDFs | 10 |
5. Business details in code (15%)
| Check | Points |
|---|---|
| Organization or local business markup with logo, description, and links to official profiles | 20 |
| Page-type markup on main pages (articles, products, services, breadcrumbs) | 20 |
| The markup is valid and complete | 15 |
| Every value in the markup is visible on the page; no self-awarded ratings | 15 |
| Business, website, pages, and people are connected in the markup | 10 |
| About and Contact pages with real people and at least two ways to reach you | 10 |
| Name, description, and category match across the site and major profiles | 10 |
6. Reputation beyond your site (15%)
| Check | Points |
|---|---|
| Other sites link to your key pages about as much as to the pages that outrank them | 20 |
| Reviews on the platforms your buyers use: enough, recent, good, and answered | 20 |
| You appear on the third-party pages AI assistants cite for your category | 20 |
| Press, expert quotes, unpaid mentions, podcasts or video, and LinkedIn activity | 15 |
| A Google Knowledge Panel or Wikidata entry | 10 |
| One consistent story across LinkedIn, review sites, directories, and your Google profile | 15 |
Caps, coverage, and what’s measured
Fatal problems cap the score
Some problems undo everything else. When one is found, its area is capped at 40 and overall readiness at 60, and it goes first in the report:
- A page that brings in business is hidden from search engines (set to “don’t index,” blocked, or pointed at another address).
- Google or Bing is blocked from key sections, or every AI search crawler is blocked without a deliberate decision.
- The main content of a key page isn’t in the HTML until JavaScript runs.
- The site contradicts itself on facts like prices, hours, counts, or location.
- Reviews, ratings, credentials, or statistics that are made up or can’t be verified; undisclosed affiliate links.
- Business markup claims things the page doesn’t show, such as self-awarded star ratings.
- A firewall serves search crawlers a block or a stripped page (confirmed in logs or settings).
- Placeholder text (“Lorem ipsum,” “TODO”) on a live page that brings in business.
The cap is why a site with excellent content can still score in the 50s: if crawlers can’t read the content, the quality doesn’t count yet.
Coverage: what couldn’t be checked
Some checks need data only you have, such as Google Search Console, analytics, or server logs. A check I can’t run is marked “not assessed,” the area is scored on what was assessed, and the report shows the coverage percentage. An area with less than half its checks assessed is reported as “insufficient data,” not as a score.
Measured or judged
| Type | Examples | How it’s done |
|---|---|---|
| Measured | Crawler access, status codes, made-up-address test, canonicals, sitemap, words in the HTML, titles, headings, markup validity, secure connections, links, accessibility screening | Automated tools; anyone repeating the test gets the same result |
| Measured, with your access | Real-visitor speed, search traffic and queries, index status, AI crawler visits | Google Search Console, analytics, server logs |
| Judged | Whether a page fits what searchers want, depth, proof, author credentials, answer quality, whether markup matches the page, reputation | Reviewed by me against the checks above, with the evidence cited in the report |
What the numbers mean
| Readiness | Meaning |
|---|---|
| 85 to 100 | Strong; focus on reputation and measured visibility |
| 70 to 84 | Solid foundation with clear gaps |
| 50 to 69 | Significant gaps limiting visibility |
| Below 50 | Foundational problems; fix access and content first |
Comparing over time
Each site keeps a fixed set of tracked pages, the same checklist version, and the same set of AI questions, so a later review measures change, not a different test. When any of those has to change, the report says so and compares only what stayed the same.
What I promise, and what I don’t
I don’t promise rankings or AI mentions. Nobody controls Google’s results or what an AI assistant says. Google’s own guidance puts it plainly: “No one can guarantee a #1 ranking on Google.” Be wary of anyone who does.
What I do promise:
- Specific findings. Every issue comes with the evidence: the page, the exact text or code, and what each crawler received.
- Specific fixes. For each issue, what to change and where, with code ready to paste and steps for your website platform.
- A test for every fix. A pass or fail check you or your developer can run, so you know the fix worked without taking my word for it.
- A projected score. Because the checklist is fixed, the report shows what your readiness score becomes when the listed fixes are made.
- An honest re-score. The same checklist, the same pages, the same questions, so the next review shows exactly what changed.
The guarantee is tied to the work. Make the fixes the report lists, then tell me they’re done. I’ll re-score your site at no charge, using the same checklist and the same pages. If every listed fix passes its test and your readiness score hasn’t gone up, I’ll refund the report.
The fixes have to be made within 90 days of the report. Fixes that weren’t made, or that don’t pass their test, aren’t covered, because the score can only measure changes that actually happened.
With monthly tracking, I promise measurement, not outcomes: each month you’ll know whether AI assistants name you, whether what they say about you is correct, how your scores moved, and what to fix next.
Tools, sources, and updates
How a review runs. Each page is fetched twice: as plain HTML, the way most AI crawlers see it, and in a real browser after JavaScript runs. The homepage is also fetched as each major search and AI crawler (Googlebot, Bingbot, GPTBot, OAI-SearchBot, ClaudeBot, Claude-SearchBot, PerplexityBot) to see what each one receives. Accessibility is screened with axe-core in Chromium at desktop and phone widths. Your Google Business Profile, reviews, and competitors come from Google Maps and public web searches. AI answers are sampled by asking the same buyer questions on several assistants, more than once each, because answers vary.
Where the rules come from. Google Search Central documentation, Bing Webmaster Guidelines, each AI company’s published crawler documentation, the W3C’s accessibility guidelines (WCAG 2.1 AA), and FTC rules on reviews and endorsements. Industry studies are used only as supporting evidence and labeled as such.
Updates. Search and AI products change monthly. I review the official documentation every month and update this checklist when a rule changes. This page shows the version in use.
Version: checklist v3, November 2026. Scores from earlier versions aren’t directly comparable on every area; a re-score notes any change.