What Is a Technical SEO Audit?
A diagnosis rather than a treatment, which this industry sells as though it were both. Written for somebody holding a proposal and trying to work out what they would actually receive, what it would cost to act on, then whether they need one at all.
An Audit Is A Diagnosis
An audit tells you what is wrong. It changes nothing by itself. Almost every disappointment with one comes from a business expecting the document to be the outcome rather than the beginning of the work.
What you receive. Findings.
An account of what was examined, what was found and what it means. That is a genuinely useful thing to have and it is not a change to your website.
What you do not receive. Anything fixed.
Unless implementation was explicitly part of what you bought, the site is exactly as it was on the day the document arrives. Every finding is still there.
Why the confusion is so common. The proposal blurs it.
Audits are sold with language about improving a site, which describes what acting on the findings might achieve rather than what the audit does. Those are separate purchases.
The medical comparison, which holds. A diagnosis is not treatment.
Nobody expects to leave a consultation cured because a condition was identified. The identification is valuable, necessary and entirely distinct from the treatment.
What to establish before buying. Which one you are buying.
Ask directly whether the price includes making the changes. If not, ask what implementation would cost. That single question resolves most of the disappointment described on this page.
Reporting Is Not Fixing
A two hundred page audit delivered and never acted on has achieved nothing. That is extremely common. A business should budget for the work rather than for the document.
Why this happens so often. The document feels like progress.
Something substantial arrives, it is read, then it goes into a folder. The sense of having addressed the problem is genuine and nothing about the site has changed.
Why it suits a supplier. It is the easiest thing to sell.
An audit is the easiest deliverable in this industry to produce and the easiest to charge for while changing nothing. It is definable, deliverable and complete on the day it lands.
What that means for the buyer. The money is in the wrong place.
A business with a fixed budget spending most of it on a diagnosis has little left for treatment. The proportions should usually run the other way.
Why we hold this position everywhere. Consistency.
The same argument appears in our material on measuring performance and on hiring an agency. It is not a technical observation, it is a commercial one about how this industry bills.
What good looks like. A plan with a cost attached.
Findings, prioritised, with an estimate of what fixing the important ones involves. Anything without that second half is half a service, which is block nine.
The Automated Export Problem
A crawler can produce hundreds of pages of findings in minutes. That output is frequently sold as an audit. Telling the difference requires no technical knowledge at all.
What the tools produce. Everything, unsorted.
Software examines a site and lists every deviation from an ideal. It is thorough, fast and entirely undiscriminating, because it has no view about which findings matter to your business.
Why the output looks impressive. Volume.
Two hundred pages of findings suggests two hundred pages of work was done. The document is long because the software found a lot, which is not the same thing.
The test. Did anybody prioritise it.
Look for a statement of what to do first and what to ignore. If everything is presented at equal weight, nobody has applied judgement and you are holding an export with a cover page.
The second test. Does it mention your business.
A real audit references what you sell, who you sell to and which pages matter commercially. A generated one could describe any site in your sector, else any site at all.
Why the tools are still useful. They are the right starting point.
Nobody examines a site of any size by hand, so this output is where a genuine audit begins. The value is in what somebody does with it afterwards.
What A Real One Contains
Seven areas, where the value is in the judgement rather than the coverage. Any competent audit examines all of these, which is why the list is not what distinguishes a good one.
Crawling and indexing. Whether pages are found and kept.
Which pages are reachable, which are stored and where the gap between those numbers comes from. Crawling and indexing covers the sequence this examines.
Your instructions to crawlers. What the site is asking for.
Whether anything is being blocked deliberately or otherwise, plus whether those decisions still make sense. This is where the most damaging single mistake in this section hides.
Addresses and duplication. How many versions exist.
Whether the same content sits at several addresses and whether the site has stated which version it regards as definitive.
Status codes and redirects. What the server is answering.
Which requests fail, which redirect and whether permanent moves are announced as permanent. Status codes covers what actually needs action.
Performance, mobile and structured data. The remaining three.
How quickly pages respond, whether the mobile version contains what the desktop one does, then whether descriptions of pages are accurate.
Why the list is not the point. Everybody covers it.
Any audit worth the name examines all seven. What separates them is whether somebody then decided which findings matter, which is block five.
Prioritisation Is The Deliverable
A hundred findings with no order is not usable. A real audit says what to do first and what to ignore. The ignore list matters as much as the fix list.
Why an unordered list fails. It cannot be started.
Faced with a hundred items of apparently equal weight, most businesses do nothing. Not through negligence. It is because there is no obvious first step and no way to judge one.
What ordering requires. Knowing the business.
Which pages earn money, which sections matter, what the site is for. Software cannot supply that and it is the entire difference between findings and advice.
Why the ignore list matters. It is permission.
Telling a business which findings are not worth acting on removes work rather than adding it. That is unusual in a deliverable people are paying for. It is also the part they benefit from most.
What a good ordering looks like. Severity against effort.
What is genuinely costing you something, weighed against what fixing it involves. A serious problem that takes an hour outranks a moderate one that takes a fortnight.
The question to ask of any audit. What should we do on Monday.
A document that cannot answer that in one sentence has not been prioritised, whatever else it contains.
Tools
Four categories contribute different things and no single one sees everything. We are describing what each contributes rather than naming products, which is a position we hold across this whole section.
A crawler. Walks the site as software would.
Follows links and records what it finds at every address. This is what produces the raw material and it sees your site the way something fetching it does.
The search engine's own reporting. What it actually did.
What was fetched, what was stored and what was excluded. Uniquely valuable because it reports actual behaviour rather than a simulation of it.
A performance measure. How it behaves for a person.
Loading and interaction measured under realistic conditions. A different question from whether the site is correctly built.
Something reading your own instructions. What you are asking for.
Checking the files and declarations your site publishes about itself, plus whether they agree with each other.
Why we name none of them. A consistent rule.
We never name the audit tool. We name others only where a page would be unusable without the name. This page is not, so the categories are what you need and the products are not.
Why no single tool is enough. Different vantage points.
A crawler sees what it can reach. The reporting sees what was actually done. Neither knows what the other knows, which is why an audit built on one source has blind spots.
Frequency
A full audit annually, plus after any significant change. Monthly audits are a way of billing rather than a need. Continuous monitoring is a different thing entirely.
Why annually is usually right. Sites change slowly.
The technical state of a site that has not been rebuilt does not shift much in twelve months. A second audit finding the same things is not a new diagnosis.
The other trigger. Significant change.
A rebuild, a platform move, a restructure or a migration. Those are the moments when technical problems are actually introduced. An audit afterwards is worth considerably more than one on a schedule.
Why monthly does not make sense. Nothing changed.
A monthly audit on a static site re-examines the same site and finds the same things. What is being purchased is the recurrence rather than the finding.
What monitoring actually is. A different service.
Something watching for problems appearing: pages failing, certificates lapsing, sections becoming unreachable. Genuinely valuable, ongoing, though not an audit.
Why the distinction matters commercially. They are priced alike.
Monitoring is largely automated and an audit is largely judgement. Being sold one as the other means paying for expertise you are not receiving.
The Checklist
A checklist is a memory aid for somebody who already knows what they are looking at. Handed to somebody who does not, it produces a list of unactioned items and a sense of failure.
What it does for an expert. Prevents omission.
Somebody who could examine all of this from memory still benefits from a reminder, because the risk is forgetting rather than not knowing. That is what checklists are for in every field.
What it does for everybody else. Produces anxiety.
A list of items you can neither assess nor act on. Each one reads as a possible problem and none can be resolved, which is worse than not having looked.
Why they circulate anyway. They are easy to publish.
A checklist looks generous and costs nothing to produce. It transfers the difficult part, which is judgement, to somebody who does not have it.
Why we are not publishing one. It would contradict this block.
Publishing a runnable checklist here would hand you exactly the artefact this paragraph describes as unhelpful. The pillars in this section explain what each item means instead.
What to use instead. The questions.
Each guide here ends with the questions a business owner can usefully ask. Those are answerable without technical knowledge, which a checklist is not.
What Happens Afterwards
Most findings need a developer. That is a separate cost. An audit delivered without any sense of what implementation involves is half a service.
Who does the work. Usually not the auditor.
Findings about templates, server responses and site structure require somebody who can change the site. That may be your developer, your platform provider or the agency. It is rarely whoever wrote the document.
What that costs. Frequently more than the audit.
Implementation is the larger number in most cases, which is exactly why the proportions in block two matter. A business budgeting only for the diagnosis has budgeted for the smaller half.
What makes it worse. Findings written for the wrong reader.
A document written for a business owner may not contain enough for a developer to act on. Somebody then has to translate it, which is additional work nobody priced.
What to ask for. Findings a developer can use.
Enough specificity that whoever implements can start without a further conversation. That is a reasonable thing to require and it is not always supplied.
The question that saves the most. What will this cost to fix.
Asked before commissioning rather than after. An auditor who cannot estimate that has not thought about the second half. Our material on hiring an agency covers judging that.
When You Do Not Need One
A small site on a modern hosted platform frequently has very little wrong technically. The problem is usually content or links. A business should not be sold an audit as a default first step.
Why modern platforms changed this. They handle it.
Sensible addresses, working sitemaps, secure connections and mobile layouts all arrive by default. The failures this section describes were far more common when everything was assembled by hand.
What a small site usually has wrong. Not much.
A handful of missing pages, a few slow images and some old content. Real, minor and not what is preventing the business from being found.
What is usually the actual problem. Content or links.
Not enough pages answering what customers search for, else nothing pointing at the site from anywhere else. Neither is technical and neither is fixed by an audit.
Why it gets sold first anyway. It is definable.
An audit has a clear scope, a clear deliverable and a clear end. Content and links are slower, harder to quantify and less comfortable to sell, so the technical work goes first.
When it genuinely is the right first step. Three cases.
A site that was rebuilt or migrated recently, a large site with thousands of pages, else a site where something changed and nobody knows why. Those are real diagnostic situations.
What we would tell you. Which of those you are.
If your site is small, modern and hosted, we would say so before quoting for anything. The full series is on the technical SEO guide.
Ask what it will
cost to fix, before
commissioning.
Implementation is usually the larger number, so a business budgeting only for the diagnosis has budgeted for the smaller half. An auditor who cannot estimate what fixing the important findings involves has not thought about the second half.
What you should receive:
If your site is small, modern and hosted, we will say so before quoting for anything.
Every guide.
One practice.
How search finds and stores a site, what your own files are telling it, addresses and duplication, status codes, mobile, hosting and what a real audit contains.
