What does an SEO audit include, and what a score widget leaves out

You pasted your address into a free checker, waited sixty seconds, and got a number. Now what? The number does not say what it looked at, whether it looked at one page or all of them, or which of your problems is the expensive one.

This page explains what a real SEO audit includes, area by area, in words that do not need a glossary. It also explains where a score out of 100 comes from, why two checkers can give the same site very different numbers, and what an audit genuinely cannot tell you. Everything here is checkable by hand, with free tools, by someone who has never done it before.

What an SEO audit includes

An SEO audit is a systematic check of a website against the things search engines need in order to find, understand and show its pages. A full one looks at every page, not the homepage alone, and ends with a list of specific problems rather than a single number.

A complete audit covers six areas:

1. Crawling and indexing. Can Google open your pages, and is it allowed to show them.
2. Speed on real devices. Does the page load fast enough on a phone on an ordinary connection.
3. Page structure and content. Does each page have its own title, description and headings, and do they describe what is on it.
4. Duplicate and competing pages. Do two of your pages say almost the same thing.
5. Structured data. Is the extra markup that lets search results show details such as ratings, prices or dates present and valid.
6. Links, internal and external. Do your own links work, and can a visitor reach every page from somewhere.

That is what a website SEO checker should be checking. Whether a given one checks all six, and tells you so, is the first thing to ask about it.

What each area actually checks

The third column is the important one. It is the same finding written the way it would matter to you, rather than the way a tool would label it.

Area	What is checked	What a typical problem looks like
Crawling and indexing	Whether pages are reachable and allowed to appear, and what is quietly blocking them	Google cannot open the page at all, or a line in a settings file is telling it to stay away
Speed on real devices	How long the main content takes to appear on a phone	The page takes six seconds to show anything, and people leave before it finishes
Page structure and content	Titles, descriptions, headings, and whether the page answers what people typed	The page has a title that repeats on nine other pages, so nothing in the search result tells anyone which to pick
Duplicate and competing pages	Near-identical pages and thin pages with little on them	Two pages say almost the same thing, Google picks one, and it is not always the one you would have chosen
Structured data	Presence and validity of the markup that describes what a page is	The search result shows no rating, price or date, because the markup is missing or broken
Links, internal and external	Broken links, and pages nothing links to	A link points to a page that no longer exists, or a page exists that nothing on your site points at

Speed is the one area with a published number to aim at. Google's guidance is that the largest piece of content on a page should appear within 2.5 seconds, measured at the 75th percentile of real page loads across mobile and desktop (web.dev, checked 9 September 2026). That is a target for real visitors on real devices, not for a test on your own laptop.

Technical checks in plain words

Three terms you will meet in any report, each explained without using another term:

• Indexing is Google having a copy of your page and being willing to show it. A page can exist, be perfectly good, and still not be indexed.
• A canonical is a line on a page saying "this other address is the real one". It exists so that near-identical pages do not compete with each other. Pointed at the wrong address, it hides a page you wanted shown.
• A redirect is an instruction that sends anyone asking for one address to another. Useful after a move, a problem when it points at something that no longer exists.

One more pair of words people mix up. A crawl is the machine part: something walks your links and collects what it finds on each page. An audit is the crawl plus the judgement on top of it, deciding which findings matter for your site and in what order. A crawl on its own gives you a spreadsheet. An audit gives you the first three things to do.

Why a score out of 100 is not an audit

A score is a summary, not a plan. It compresses everything a tool looked at into one number, and in doing so it throws away the only thing you needed: which problem to fix first.

A sixty second score widget	A full crawl
Usually looks at one page, often the homepage	Looks at every page it can reach
The weighting is chosen by whoever built it, and is rarely published	Reports each finding separately, so nothing is hidden inside an average
Does not tell you what to fix first	Puts findings in priority order
Names findings in tool language	Explains each finding in words you can act on
Leaves you with a number	Leaves you with a list

This is also the answer to why two checkers give the same site 62 and 84 on the same day. They are not disagreeing about your site. They are running different sets of checks and weighting them differently, and almost none of them publish the formula. Comparing two scores from two tools tells you about the tools, not about your website. Comparing one tool's score against itself over time is more useful, and even then only as a rough direction.

So what is a good SEO score? There is no number that answers that across tools, because each one sets its own scale and its own weighting. A score is useful against itself, on the same site, with the same tool, over time. Read as an absolute grade it means very little, whichever tool produced it.

None of this makes score widgets bad. A number is a fine way to notice that something is worth looking into. It is simply not the thing that tells you what to do next.

How findings get prioritised

A report with forty findings and no order is a report nobody finishes. Four levels, in the order they deserve your attention:

1. Google cannot see the page at all. Blocked, not indexed, returning an error. Nothing else you fix on that page matters until this does.
2. The page loads, but people leave before it finishes. Slow pages on phones, layouts that jump around while loading.
3. The page works, but says the wrong thing or says nothing. A title that does not match the query, a thin page, two pages competing with each other.
4. Cosmetic items that can wait. Missing alt text on decorative images, small markup warnings. Real, and not this quarter's problem.

The order matters more than the completeness of the list. A person who fixes the three items in level one has done more than a person who works alphabetically through forty. This is the same ordering that AI SEO content software applies programmatically, and the Kitto audit reports findings in exactly these bands rather than as a single figure.

What an audit cannot tell you

Being clear about the limits is part of being useful. An audit will not:

• Tell you whether your text persuades anyone. It can see that a page is readable. It cannot see whether it sells.
• Judge your expertise for you. Whether you genuinely know the subject you are writing about is a judgement a reader makes, not a check a crawler runs.
• Predict positions or traffic. No audit does this, and any tool that claims to is describing a hope, not a measurement.
• Replace watching what actually happened. That lives in your own Google Search Console, which records the queries where your site appeared and where it sat. Reading it is a separate job, and the Search Console opportunities guide walks through it step by step.

What to do after the audit

Five steps, in order.

1. Take the whole of level one first. Everything Google cannot see or cannot show. Nothing else competes with this.
2. Fix what you can fix in your own admin. Titles, descriptions, headings, duplicate pages, broken internal links. Most of this is editing, not coding, and most of it takes minutes per page.
3. Put the rest on a separate list for whoever handles your code. Server response times, redirect chains, template-level markup. Keeping them apart stops the list you can act on from feeling impossible.
4. Write down the date. You will want to know later what changed and when.
5. Come back in four to eight weeks and look at the same pages in Search Console. Impressions, clicks and average position, before and after, on your own URLs.

If you do not have a report to work from yet, you can run a free SEO audit on your own site. It needs no card and no account, it covers the six areas above, and the findings are yours to keep whether or not you do anything else with Kitto.