Numbers come from checks
The score is built from observable signals: speed data, rendered page checks, page structure, trust markers and copy patterns.
Public methodology
How we decide whether a website is ready to do its job.
Clunky AI scores the six basics every business website should nail. The goal is not to shame a site. The goal is to show what is working, what is getting in the way, and what to fix first.
Score frame
Website Readiness Score
A 0-100 score answering one plain question: is this website ready to do its job for a founder, buyer or small team?
6
basics
100
points
0
raw IPs stored
AI explains, it does not score
Narrative commentary can make the report easier to read, but the numbers are not invented by a language model.
Every scan is versioned
When the method improves, new scans use the new version. Historic reports keep the method they were scored with.
The report must be useful
A low score should tell a busy team what to fix next, not just make them feel judged.
The six basics
Six direct questions, weighted by how much they affect a working business site.
Each category earns points from captured checks. Higher is better. The full report shows the question, the score and the strongest proof behind it.
Basic 1
Does it load fast enough?
Performance & Technical
20%
weight
What we check
Mobile loading speed, Core Web Vitals, HTTPS, SSL and HTTP/2.
Why it matters
A site can look expensive and still lose buyers if the first useful screen arrives too slowly.
What you see
The report shows the loading and transport checks that affected the score.
Basic 2
Will anyone find it?
SEO & Discoverability
20%
weight
What we check
Titles, descriptions, headings, structured data, crawl signals, internal links and page depth.
Why it matters
If search engines and AI answer systems cannot understand the page, the business is harder to find.
What you see
The report calls out weak discoverability signals instead of hiding them in a black-box grade.
Basic 3
Can everyone actually use it?
Accessibility
15%
weight
What we check
Rendered-page accessibility checks covering contrast, landmarks, labels, heading order and common blockers.
Why it matters
Accessibility problems are not polish issues. They stop people using the site and create avoidable legal risk.
What you see
Repeated instances of the same issue are capped in the score, but the report still shows affected elements.
Basic 4
Does it feel trustworthy?
Trust + visual polish
15%
weight
What we check
Privacy, cookies, terms, contact, secure connection, mixed-content and trust-marker checks.
Why it matters
Buyers hesitate when a site feels anonymous, unfinished or hard to verify.
What you see
We check whether the basics are present and reachable, not just whether the words appear somewhere.
Basic 5
Does it help people take action?
Conversion + layout
15%
weight
What we check
Above-fold calls to action, form usability, CTA contrast, layout clarity and action density.
Why it matters
A useful website should make the next step obvious without shouting at every visitor.
What you see
The scan looks at the rendered page, so hidden or buried calls to action do not get undue credit.
Basic 6
Does the copy feel credible?
Copy Credibility
15%
weight
What we check
Copy specificity, sentence rhythm, repeated phrasing, generic AI-shaped wording and sample size.
Why it matters
Visitors can feel when a page was assembled from vague claims. Specific, human copy builds confidence faster.
What you see
This is a probabilistic copy-quality signal, not a claim about who wrote the copy.
How a scan runs
We scan the site a customer actually gets, not an idealised version of it.
A report should survive scrutiny. That means the score needs to be traceable back to page checks, not a vibes-based opinion about the design.
- 1
Fetch the public page and follow redirects.
- 2
Render it like a real visitor would see it.
- 3
Run the six basics against the captured page signals.
- 4
Store the score, version, findings and hashed IP only.
- 5
Generate plain-English commentary and fix prompts from the issues found.
Score bands
The labels are plain on purpose.
We use direct language because vague language is how website problems survive. A site can be good and still have clunky parts. A broken score means the basics need attention before more traffic is sent to it.
85-100
Sharp
Rare. The site is doing the basics well and can be used as a benchmark.
70-84
Good
A decent baseline. There are still fixable weak spots worth tightening.
55-69
Clunky
The danger zone for many SMB sites: visible friction, but usually very fixable.
40-54
Very clunky
Important basics are failing and likely costing trust, attention or leads.
0-39
Broken
The site needs urgent work before it can reliably do its job.
Fix prompts
The useful part is what happens after the score.
When we can identify a specific issue and a likely website builder, the report creates scoped prompts you can paste back into that builder. The prompt references the issue found in the scan and asks for a narrow fix, not a full-site regeneration.
After applying the prompt, you can re-scan the site and compare the delta. That before-and-after result helps us learn which builders actually fix their own mistakes.
Boundaries
What the score is not.
- It is not a guarantee of revenue, rankings or legal compliance.
- It is not a paid placement system. Rated entities cannot buy a better score.
- It is not an AI detector. Copy Credibility is a copy-quality signal.
- It is not a replacement for human judgement on brand, strategy or product-market fit.
Ready to see the checks for your site?
Run the scan, read the findings, then decide whether the site is sharp enough for the job you need it to do.