The Diagnostic: what’s wrong, in the order I’d fix it
“We know something’s off. We just don’t know what.”
That’s roughly where most people are when they get in touch. Traffic slid and nobody can say why. A developer did some work, or said they did. Someone ran a free audit tool and it came back with a wall of red warnings and no sense at all of which ones matter. You don’t need more findings. You need someone to tell you which five things to do first, and which forty to ignore.
That’s what this is. A fixed sequence, run the same way every time, on your site.
The four areas, every time
Same order whether you’re a five-page site or a catalogue with thousands of URLs. The depth changes; the sequence doesn’t.
Technical health
Crawl and index status, redirect chains, canonicals that disagree with each other, the sitemap against the robots rules, render behaviour, and Core Web Vitals measured on the templates people actually land on rather than the home page alone.
On-page and content
Titles, headings and internal links read against what each page is genuinely trying to rank for. Thin pages. Duplicated intent. Pages quietly competing with each other for the same query and splitting the result between them.
Link signals
What the backlink profile is actually made of, where it came from, how much of it is still live, and whether any of it is the sort of thing that turns into a liability later.
Generative-engine visibility
Whether your pages answer a question in a form a model can lift cleanly: an answer near the top, facts that are structured, schema that validates, and a description of who you are that stays consistent from one page to the next.
What lands in your inbox
One document. Numbered findings, ordered by what I’d do first — not by some severity score a crawler assigned.
Each finding says three things: what’s wrong, what it costs to put right, and whether putting it right is likely to matter. Some items will say leave this alone. Those are usually the ones another audit would have tried to sell you.
It’s written so your developer can act on it without me in the room: the file, the rule, the expected result. If you’d rather I did the work, the quote comes off the same list and you’ve already seen every number on it.
No portal. No login. No recurring report you stop opening after the second month.
How it runs
You send the URL
Along with anything you already know — the redesign that went badly, the migration nobody documented, the agency you left and why. Context saves me time, which saves you money.
I run the four areas
A crawl, then manual review of the real page types, then the link data, then the generative-visibility pass. This part is hands-on. No tool writes the report; tools just gather what goes into it.
You get the report
Numbered, priced, in priority order. Yours to keep and yours to shop around with if you want a second opinion on the quote.
We talk it through, if that’s useful
Some people want an hour to go through it. Plenty of others take the document and get on with it themselves, and that’s a perfectly good outcome — it’s written to work that way.