The work that decides whether anything else counts.
Content cannot rank from a page a crawler never reaches, and it cannot be quoted by an answer engine that could not parse it. Technical SEO is the layer underneath everything, and it is where most sites quietly lose.
- Crawl, render, index
- In priority order
- Against live crawl data
Most sites are slower and thinner than their owners think.
A site looks fine in a browser because a browser is patient. A crawler is not. It has a budget, it abandons slow responses, and it reads what the server returns rather than what JavaScript eventually assembles.
The usual findings are the same across most small business sites. Hundreds of indexable URLs nobody intended, generated by filters and tag archives. Pages linked from nowhere. Redirect chains three hops deep. A sitemap listing URLs that redirect. Images at full camera resolution scaled down in CSS. And a render path that blocks on a stylesheet from a plugin used on one page.
None of these are visible to the person who owns the site. All of them are visible to the crawler, and now to the retrieval systems behind AI answers, which are considerably less forgiving about structure.
The first technical audit on a neglected site usually finds more upside than six months of content would.
It is also the cheapest work we do, because it is finite. You fix it once.
Crawl it the way a machine does, then fix what blocks it.
We crawl the site properly, render included, and compare what the crawler sees to what is in the index and what is in the sitemap. The three almost never agree, and the gaps between them are the findings.
Then we work in order of consequence rather than order of ease. A canonical pointing at the wrong URL costs more than a missing alt attribute, so it gets fixed first, even though the alt attribute would make the audit score go up faster.
Everything gets re-crawled after the fix. A technical recommendation nobody verified is a to-do list, not a result.
We will tell you which findings do not matter. Most audit tools flag hundreds of issues and weight them all the same.
Half of a typical report is noise, and acting on it wastes developer time that the real problems needed.
Deliverables.
Full technical crawl
Rendered crawl of the whole site, compared against index coverage and the XML sitemap, so you can see where the three disagree.
Indexation plan
A decision for every URL type: index, noindex, canonicalise or remove. Written down with the reasoning, not just applied.
Core Web Vitals work
Diagnosis against field data where you have it and lab data where you do not, then the specific fixes: image handling, render blocking, layout shift, interaction delay.
Structured data
Schema that describes what the page actually is, validated, and checked for the errors that stop it being eligible for anything.
Internal linking
Orphan pages found and linked, and link equity directed towards the pages that have a commercial job rather than spread evenly across everything.
Redirect and 404 cleanup
Chains collapsed to a single hop, loops removed, and genuine 404s either restored or redirected to something relevant rather than the homepage.
Retrieval readiness
The structural work that makes a page usable by an answer engine: clear headings, self-contained sections, specific factual statements, no content that only exists after a script runs.
Re-crawl and verification
The same crawl run again after the work, so the fixes are demonstrated rather than asserted.
How the work runs.
Crawl and baseline
Full rendered crawl, index coverage, sitemap comparison, Core Web Vitals field data where it exists.
Prioritised findings
Everything found, ranked by consequence, with the noise explicitly marked as noise so nobody spends a day on it.
Implementation
We make the changes where we have access, or hand your developer a specification precise enough to act on without a meeting.
Verification
Re-crawl, re-test, and a short written confirmation of what moved and what did not.
Monitoring
A crawl schedule so the same problems do not quietly return the next time a plugin is updated.
Straight answers.
Is technical SEO still relevant if AI is answering the questions?
More so. Answer engines are retrieval systems, and retrieval needs a parseable, crawlable, clearly structured source. A page that a crawler struggles with is a page that never enters the candidate set an AI picks from.
The difference is that the structural work now has two audiences rather than one, and neither of them is patient.
Can you do this without access to my developer?
Often, yes. Most of it is configuration, templates and server behaviour we can change directly. Where a fix needs a code change we cannot make, you get a specification detailed enough to be handed straight to whoever can.
How long before it shows in rankings?
Indexation changes can show within days. Speed and structure work usually takes a few weeks to be recrawled and reassessed. Anything that depends on the site being recrawled depends on how often you are crawled, which the work itself tends to improve.
Is this a one-off or ongoing?
Mostly a one-off, then light monitoring. Technical debt accumulates slowly, usually after plugin updates, redesigns and migrations. A crawl every quarter catches it before it costs anything.
What if the audit finds the site needs rebuilding?
Then we say so, and we will tell you roughly what that costs before you spend anything else on patching it. Some platforms cannot be optimised into being fast, and pretending otherwise just bills you for the attempt.
Where this connects.
Tell us what’s next.
Send over the problem and the numbers you have. We will tell you honestly whether we are the right people for it.