What you are actually buying
An audit has three layers, and most are strong on one and thin on the other two.
- Technical — can Google crawl, render and index the pages, and how fast are they on a phone. Tools do this well, which is why it dominates most audits.
- Content — what each page is written to target, whether that matches what it actually ranks for, and where two of your own pages are competing with each other. This needs somebody to read the pages.
- Authority — what links you have, what the sites ahead of you have, and whether that gap is bridgeable in a sensible timeframe. This needs an honest judgement, which is the part nobody selling ongoing work wants to write down.
It is easy to generate three hundred findings about missing alt text. Two hundred findings in a document is not an audit, it is a haystack. Ten findings ordered by impact against effort, each with a sentence explaining why it matters to this business, is something a person can actually work through — and it is a much harder thing to produce.
What audits cost, and why ours is not sold separately
Standalone audits are a normal, well-established product and there is nothing wrong with buying one — particularly if you have somebody in-house who will implement it. The structural problem is that the document is the deliverable, so the incentive is to make it impressive rather than to make the fixes happen. The fixes are the only part that changes anything.
Here it is the first month of a growth block from £250 a month. Same work, ranked the same way, except that the rest of the month is spent at the top of the list.
That structure also removes the awkward moment where an audit recommends six months of work from the person who wrote the audit. It is a real conflict, it is rarely disclosed, and the way to defuse it is not to pretend it does not exist but to make the fixing and the finding the same engagement at the same price.
The half you can genuinely do yourself, for nothing
A serious amount of the value in an audit comes from three free things, and anybody can do them in an afternoon.
- Open Google Search Console and read the Pages report. It tells you which of your pages Google has indexed and which it has not, and why. That is not an estimate — it is Google telling you directly.
- In the Performance report, sort your queries by impressions and look at the ones ranking between about eight and twenty. Those are the pages nearest to earning something, and they are almost always where the fastest wins are.
- Open your own site on your own phone, on mobile data, and try to buy or enquire. Time it. Most conversion problems announce themselves within ninety seconds of doing this honestly.
What you cannot easily do yourself is the comparison — reading your pages against the results that beat them and deciding which gap is closable. That is the judgement half, and it is what you are paying for when you pay for an audit.
Can ChatGPT or a tool do the audit?
Partly, and the split is clean. A model is good at reading a page and telling you what it is about, drafting a better title, spotting that two of your pages target the same thing, and explaining a technical finding in plain English. Those are real hours saved.
What it cannot do is see your data. It does not have your Search Console, it does not know which pages earn, and it will confidently describe a live result it has not looked at. An audit built from a model’s general knowledge of SEO is a summary of the internet’s opinions about SEO, not a reading of your site.
Crawlers have the opposite problem: they see everything and rank nothing. The combination that works is a crawl for coverage, Search Console for what Google actually did, and a person to decide the order. Nobody has automated the last one yet, which is fortunate.
On free audits
A free audit is a sales document, and that is fine as long as everybody knows it. It is generated by a crawler, it will find between forty and four hundred issues on any site including very good ones, and its purpose is to produce a phone call.
The tell is whether the findings are ranked and whether anybody looked. If the headline finding is “missing meta descriptions” on a site whose real problem is four pages competing for one term, nobody read it.
A related warning worth carrying into any external audit of any site: automated tools report what they can reach. A site behind a bot-challenge or a redirect will be reported as broken by a crawler that could not get in, and every finding underneath that will be nonsense. Open one of the flagged pages in a real browser before actioning anything.
What an audit cannot tell you
How long the fixes take to show. Technical fixes can move within weeks; content and authority work is months, and anybody offering a date is guessing.
The size of the prize, with any precision. Search volume figures are estimates, and click-through rate by position varies enormously by query — a result sitting under an AI Overview does not behave like one at the top of a plain list.
And whether the traffic will convert. That is conversion work, not audit work, and an audit that promises revenue from a crawl is selling something a crawl cannot deliver.
