What this is
An outside read of one website, delivered as a written report. You name up to five public pages; we review what those pages actually show, and send back a prioritised list of findings with the evidence for each, what to do about them, and how to tell whether the fix worked.
It is asynchronous. No call is required, no access to your systems is required, and nothing on your site is touched.
What is reviewed
- Messaging — does the page say what you do, for whom, and does it match
where the visitor came from
- Navigation and structure — can someone get from arrival to the action the
page exists for; do the buttons land where their labels promise
- Visible technical issues — what the served response shows: status codes
and redirects, canonical and robots directives, titles and descriptions, heading structure, structured data, obvious weight
- The visitor-to-purchase journey — the path to the point of commitment,
and where it asks for something it has not yet earned
Each finding carries the observation it rests on — a quote, a captured value, a response header — so you can check it yourself, and re-check it after a change. Where the next step is something Claude Code can do, the report includes the prompt.
What is not included
- No changes to your website. The report recommends; you or your developer
implement.
- No security testing. No scanning, no penetration testing, no certificate
of anything.
- No legal or accessibility certification. Accessibility issues are noted
where they also affect visitors or crawling; that is not a conformance audit.
- No guarantees. Not of rankings, traffic, sales, conversion rate, or
performance. Anyone selling those is selling something else.
- No unlimited revisions or ongoing support. One round of written
clarification questions after delivery.
- No access to your analytics, admin, hosting or any private system — and
none is requested. The form below never asks for a password, an API key, or payment details.
The limit worth understanding before you ask
A public page shows what it says, not what visitors do. Without your analytics, nobody can honestly tell you where people drop off or what share of them convert. This review reads the pages the way a first-time visitor and a search engine read them, and says where that experience contradicts what the page is trying to do.
If a finding would need data we do not have, the report says so under "not checked" rather than guessing.
What you receive
A report with: the scope and date, the pages examined, the findings — each with its evidence, a priority, why it matters, a recommended action and how to validate it — the limitations, and one round of questions.
There is no minimum number of findings. Where the pages are sound, the report is short. A padded report is worse than a brief one.
See one first
The sample report is a real review of this site, with real observations you can verify on the live pages — four findings, none invented, including two the site should act on. It is published in full rather than described:
Reviewing our own site is deliberate. The alternative — a report about a fictional client — would tell you nothing about whether the findings are real.
Status and how to ask
This page takes requests, not bookings, and takes no payment. Price, turnaround and availability are agreed in the reply, and a request commits you to nothing.
Send the form and you get one of two answers: an offer with the price, the agreed pages, the exclusions and when the report would arrive — or a decline with the reason, if the site or the goal is not something this service can honestly help with. The offer is the booking; nothing before it is.
If what you want is to run the review yourself, the method is published free: the SEO & Website Audit Toolkit and the Conversion & Revenue Optimization Toolkit are the packaged versions, and the free guides, the labs and the free tools cover the same ground at no cost. This service exists for people who would rather have someone else do it once.