Most SEO audits are a list of 1,500 errors and a file nobody opens after the first month. I work differently: I group findings by root cause, rank them by what they’re worth to your business, and give every recommendation a test a crawler can run once the fix ships.
The short version
SEO audit services are an outside review of how search engines and AI answer engines crawl, index, and rank your site. The review covers technical health, content, links, and AI search readiness, and ends in a prioritized list of fixes your team can ship and verify.
When a full audit is the wrong move
If one known event explains the drop – an accidental noindex, a robots.txt change – I check that cause directly. And if your team has only a few hours per sprint, a long report outruns what ships, so I scope the audit to one template or one symptom first.
Where the method comes from
I read three audits of the same large website, written by three different teams over six years. Each one was thorough. And each one flagged the same problems, because the recommendations from the previous audit never shipped.
The analysis wasn’t the issue. The format was: lists of URLs instead of causes, priorities set by error severity instead of business value, and recommendations nobody could check off as done. Every one of those audits broke at least two of the three rules below.
Method
“1,500 pages have an error” is a week of work for an intern. “One pattern in the blog post template produces 1,500 instances” is one ticket for a developer. Same information, but only the second version gets implemented.
A page with 1,000 visits and 27 leads matters more than a page with 10,000 visits and none. I calculate priority from organic traffic, business value, scale, and confidence, weighed against effort. If I don’t have lead data, I say so plainly instead of pretending traffic is a stand-in for it.
If a recrawl can’t verify it, it isn’t a recommendation – it’s a hypothesis. Those can be valuable too, but they go in a separate section, not on the list of commitments.
UntestableImprove internal linking.
TestableEvery service page reachable within 3 clicks of the homepage; zero in-content links pointing to a 404 or a 301.
What I check
Six areas. The basic audit covers the technical foundation; the full audit adds content and links.
Which pages search engines can reach, which they index, and what doesn’t match between the crawl, the sitemap, and Search Console. Every issue is split into indexable and non-indexable pages.
Status codes, redirect chains, canonicals, directives, and hreflang – plus what changes once JavaScript renders the page. Checked per template, not just on the homepage.
Titles, headings, and structured data per template, and how internal links distribute authority: click depth, orphan pages, and links that point to 404s or redirects.
Whether AI answer engines can reach and cite you: crawler access per bot in robots.txt and at the firewall, and content present in the raw HTML. The full audit adds entity structured data and how directly key pages answer questions.
Whether pages match the searches they target, where pages compete for the same query, and which thin pages drag a section down. Each finding comes with an action: update, merge, expand, or remove.
Where your authority comes from and where it leaks: referring domains, anchor patterns, risky links, and pages that hold links but no longer serve a purpose.
What you get
I deliver results as a presentation where each finding gets its own card. I lead with what nobody has flagged before. Every card points to a file with the full list of URLs, so your team can start work right away.
/blog/post/ → canonical: /blog/post/?utm_source=…/blog/post/ → canonical: /blog/post/P1-canonical-with-parameter.csvExample data.
Process
Business goals and KPIs, scope, access, previous audits, and what came of them. Anything I can’t check gets written down up front as a limit on the conclusions.
Two passes: a wide one without rendering, and a narrow one with JavaScript for a representative of each template. I measure the shape of the site first and only then exclude sections.
How many pages are indexable, how many are orphaned, and what doesn’t match between the crawl, the sitemap, and Search Console. I split every issue into indexable and non-indexable pages, because those are two different decisions.
A presentation of finding cards, grouped by root cause and ordered by priority, plus files with full URL lists for the team that implements the fixes.
I save a snapshot of the site on audit day. After changes ship, I recrawl, and every recommendation gets a result: PASS, FAIL, or REVIEW.
Sign-off
Every recommendation has an acceptance criterion, so after each sprint a recrawl shows how much is really done – not “in progress”, but verified.
Below, two indicators that usually move first: pages with template errors and pages Google indexes.
URLs with template errors
−84%
Indexed pages
+31%
Illustrative example – it shows the sign-off process, not a specific client’s data.
Scope
Basic
“Is it technically ready?”
A new product, a site before launch, a quick second opinion.
TimeUsually within a week.
You getA card-format presentation (10–20 slides) with full data.
Full
“Why aren’t we ranking, and what should we do this year?”
A new strategy, a dispute over the diagnosis, a weekly review cycle.
TimeSeveral weeks, with a weekly review.
You getA diagnosis, priorities from the value model, and an implementation plan verified by recrawl.
If the question grows mid-project, I say so and we revisit the scope. I don’t quietly add layers.
Options
Each route works for some sites and fails for others. Here’s how I’d choose if I were on your side of the table.
Works when
A small site and a first look. It flags missing titles, broken links, and obvious speed problems in minutes.
Watch out for
It can’t weigh trade-offs or tell you which of 50 warnings matters.
Works when
Your team knows the product, the history, and the codebase better than any outsider.
Watch out for
The audit competes with daily work and often stops halfway. A team doing the same thing for years also stops seeing new problems – it looks where it always looks.
Works when
A large site with several stakeholders that needs technical, content, and links covered in one project.
Watch out for
Higher cost and the risk of a templated report. Ask for a sample deliverable before you sign.
Works when
A migration, a traffic drop with no clear cause, JavaScript rendering, or a site where findings need to turn into tickets.
Watch out for
One person’s capacity – scope and timeline are agreed up front.
Start with the risk. A migration, a traffic drop, or a revenue-critical section justifies a specialist. A small brochure site rarely does – a free tool and an hour of your time will do.
Cost
There’s no single price, because two audits of a similar size can take very different amounts of work. These are the factors that move the quote:
A 50,000-page store built on six templates can take less work than a 2,000-page site where every section is built differently.
Sites that build content with JavaScript need a second, rendered crawl and a comparison with the raw HTML.
Each language version adds hreflang checks and its own search data to review.
Search Console, GA4, and server logs widen what I can conclude – and the time it takes to analyze them.
Basic or full, and which areas: content and links each add their own analysis.
Whether the quote includes a recrawl after your team ships the fixes.
Audits start at PLN 2,500. I quote after a short call and a first look at the site, for an agreed scope. You know what’s included before anything starts.
Honestly
Observation measured
255 URLs in one section of the site are non-indexable.
Hypothesis likely
That section is eating crawl budget.
Root cause confirmed
A generator of industry × technology × country combinations creates a URL space nobody capped.
A hypothesis never goes into the report as a fact. I treat tool estimates the same way: “you need 45 links to reach page one” is a model’s output, not a measurement, and I label it as one.
So ChatGPT, Perplexity, Gemini, and AI Overviews can find, understand, and cite you – measured, not promised.
MoreContentA decision for every URL: keep, update, merge, or remove – with cannibalization, decay, and a redirect map.
MoreMigrationA new platform, domain, or structure without losing traffic: a baseline, a 301 map, and checks before and after the switch.
MoreAudits start at PLN 2,500. The final price depends on the work, not the page count: the number of templates, JavaScript rendering, markets, data access, and scope. I quote after a short call and a first look at the site, for an agreed scope.
A basic audit usually fits within a week. A full audit of a large site takes several weeks, typically with a weekly review. We agree on the scope and timeline before we start.
It’s worth it when the findings get implemented. An audit pays back when it exposes a fault that blocks traffic, such as indexing errors after a migration. For a small site with no known problems, a free tool is often enough.
The site URL and your business goal. Access to Search Console, GA4, and server logs broadens the conclusions. If you can’t share it, the report states what can’t be determined without it.
No. The audit ends with a task list and an acceptance criterion for each task. Your team implements the changes, or we scope that work separately, and once it ships I can verify the result with a recrawl.
An SEO audit checks how search engines and AI engines reach, read, and rank the site – the full audit includes content against search intent. A content audit goes further: a page-by-page inventory with a keep, merge, or remove decision for every URL. That’s a separate service.
Not on its own. ChatGPT can’t crawl your whole site or read your Search Console data unless you export it, and it can report problems that don’t exist. It’s useful for grouping issues from a real crawl – as long as every finding is checked against the source data.
Quite the opposite. I start by reading previous audits and checking what came of them. Repeating someone else’s unimplemented recommendation without that context changes nothing.
Tell me what the question is about. I’ll reply with a suggested scope – basic or full – and what I’ll need from you.
Ask about an audit