SEO audit: how to do one step by step and what to check (with checklist)

An SEO audit is a structured review of whatever stops your website appearing and competing in Google. It takes six steps: crawling and indexing, URLs and titles, mobile experience, content, structured data and links. With Search Console, PageSpeed Insights and Lighthouse, all free, you can audit a small site in an afternoon.

To check it on your own site: SEO audit

The audit is the starting point of SEO as a whole: before you write more or chase links, it pays to know what is holding back the pages you already have.

What is an SEO audit, and what is it not?

An SEO audit is a technical and editorial diagnosis of a website that identifies what stops Google crawling it, understanding it or preferring it to other pages. The result is a prioritised list of problems, not a score or a promise of a position.

It is not rank tracking, which measures where you appear today, nor keyword research, which decides what to write about. It answers a different question: if someone searched for exactly what you offer, is there anything on your site that would stop Google showing it?

Google is clear on this: good results in the Core Web Vitals report or in third-party tools do not guarantee that your pages will rank at the top1. The audit removes the brakes; the rest depends on the competition for each query.

What do you need before you start?

Four free tools cover almost the whole audit:

• Google Search Console, with your site verified. It is the only one that shows what Google has actually indexed.
• PageSpeed Insights, which combines real-user data from the previous 28 days with a Lighthouse lab test2.
• Lighthouse, built into Chrome DevTools. It audits performance, accessibility, best practices and SEO3.
• Google's Rich Results Test, to validate structured data.

Before opening any of them, list your templates (home page, service, category, product, article) and audit one URL of each. A template problem repeats on every page that uses it.

Step 1 · Can Google crawl and index your site?

A page that is not in the index competes for nothing, so this step comes first. Start with the Page indexing report in Search Console. It shows how many URLs Google has indexed and why it left the others out: marked noindex, blocked by robots.txt, duplicate without a user-selected canonical, crawled but currently not indexed, 404s or soft 404s4. Do not panic if many are excluded: Google says you should not expect all your URLs to be indexed, only the canonical ones.

Three checks that often get skipped:

1. robots.txt is not for keeping pages out of Google. Google puts it this way: the file tells crawlers which URLs they can access, but it is not a mechanism for keeping a page out of Google. For that you need noindex or password protection5.
2. Search Console's URL Inspection tool tells you whether a page is indexed, which canonical URL Google has chosen and, with the live test, how its crawler sees the page6.
3. A sitemap helps, but small sites do not need one. Google considers a site small at about 500 pages or fewer with good internal linking, and points out that submitting a sitemap does not guarantee everything in it will be crawled7.

To see what your server returns, three terminal commands are enough (replace example.com with your domain):

# 1. Status code and headers: you want 200, not 302 or 404
curl -sI https://www.example.com/services/

# 2. Is there a noindex in the HTTP header? (it does not show in the page source)
curl -sI https://www.example.com/services/ | grep -i x-robots-tag

# 3. Do the domain variants all end up at a single URL?
curl -sIL http://example.com/ | grep -iE "^(HTTP|location)"

If the third command chains more than one redirect before the 200, or ends at a URL other than the one in your menu, you have a chain worth cutting down to a single hop. Redirects, canonicals and hreflang are covered in more depth in our guide to technical SEO.

Step 2 · Does each page have one URL and its own title?

When the same page responds at several URLs (with and without a trailing slash, with parameters, on http and https), Google picks one as canonical. Redirects and the rel="canonical" link are strong signals of which one you prefer; listing the URL in your sitemap is a weak one8. If URL Inspection shows a Google-selected canonical that is not yours, you have duplicates to resolve.

Google's SEO Starter Guide asks for a title that is unique to each page, clear and descriptive of its content, and a meta description that is short, unique and covers the most relevant points9. It also reminds you that Google Search does not use the meta keywords tag.

Lighthouse flags a missing meta description, but it does not judge whether it is any good10. That you check by reading: can you tell what is on the page from the title and description alone? How to write both, with examples, is in our guide to the title tag and meta description.

Step 3 · How is the page experience on mobile?

Google indexes and ranks the mobile version of your site, crawled with its smartphone agent11. So the mobile version must have the same main content, the same structured data and the same title as the desktop one. Compare the two for each template.

For speed and stability, the official Core Web Vitals thresholds are12:

They are measured at the 75th percentile of visits, separately for mobile and desktop. Look first at the field data in PageSpeed Insights, which comes from real visits over the previous 28 days. If a URL has too few visits, PageSpeed Insights shows data for the whole origin; if that is not enough either, you only get the lab test2. What each metric means and how to improve it is in our guide to Core Web Vitals.

Round off the step with Google's page experience questions: HTTPS, content that displays well on mobile, no excessive ads and no intrusive interstitials1.

Step 4 · Does your content answer better than what already ranks?

The editorial side is audited with the questions Google publishes for assessing content. The two most useful: does the content provide original information, research or analysis, and does it provide substantial value compared with other pages in the search results13?

The second question forces you to look outside. Search for the main query of each page and read the top three results. If they all give price ranges, timescales and examples and your page does not, that is the gap. We set out the method in our SEO competitor analysis.

Check for internal duplicates too. Google makes clear that duplicate content is not a violation of its spam policies, although it can make for a poor user experience9. The real problem is usually something else: two of your own pages competing for the same query, with neither answering it fully.

Step 5 · Does your structured data say the same as the page?

Two Google rules shape this part: do not mark up content that readers cannot see, and correct markup does not guarantee that any rich result will appear14.

Run each template through the Rich Results Test. Look for errors, but above all for mismatches: a price in the markup that differs from the one on the page, an address that differs from your Google Business Profile, or a rating that is not shown anywhere. Which types to use and how to write the JSON-LD is explained in our guide to structured data.

Step 6 · How are your links, inside and out?

Internal links are entirely under your control. Google recommends descriptive anchor text that says where the link leads9. Check that the pages you care about most receive links from the menu or from other pages, that none is left orphaned and that no link leads to a 404. Finding and fixing the last of these is covered in our guide to broken links.

For external links, the Links report in Search Console shows the sites that link to you most and the most common anchor text. It is a sample, not the full list15. What you are looking for are odd signals: hundreds of links from unrelated sites, or identical anchor text containing your keyword. Buying links to rank is something Google defines as link spam16.

The complete checklist

How should you prioritise what you find?

Blockers first, then whatever drags you down. A forgotten noindex or a badly written robots.txt can shut out a whole section, however good it is; a weak title only makes a result a little less appealing. Within each level, sort by the number of pages affected: a template problem matters more than a one-off.

This audit also does some of the work for AI search in advance. Google says that to appear as a link in AI Overviews or AI Mode, a page must be indexed and eligible to be shown with a snippet, with no additional technical requirements17. What the AI side adds, such as access for each crawler or whether a passage makes sense on its own, is covered in the AI visibility audit.

Can an SEO audit be automated?

Largely, yes. Declaration of interest: this blog belongs to SmoothSeen, a tool that automates this review. Its SEO audit analyses the URL you give it and returns two scores, one for Web ranking and one for AI ranking, a prioritised action plan and a before-and-after view when you analyse again. The paid plans add a comparison with up to three competitors and the report as a PDF or a shareable link, with your own branding on the Agency plan.

It also checks things this checklist leaves out and that are often forgotten, such as HTTP security headers or the domain's SPF and DMARC records. What no tool replaces is the reading in step 4.

Frequently asked questions

How often should you run an SEO audit?

Run a full one when you start working on a site's SEO and again after every major change: a redesign, a domain move, a new template or a new content management system. In between, checking the indexing and Core Web Vitals reports in Search Console once a month is enough to catch new problems before they spread across the site.

Can you do an SEO audit for free?

Yes. Search Console, PageSpeed Insights, Lighthouse and the Rich Results Test are all free and cover indexing, speed, structured data and the links Google knows about. What it costs is time: a few hours for a small site and a good deal more for a large online shop with thousands of product and category pages.

If Lighthouse gives me 100 for SEO, is my site well ranked?

Not necessarily. Lighthouse's SEO category checks basic requirements, such as whether a meta description exists or the status code is correct, but it does not judge their quality or compare your page with the ones competing for the same search. A perfect score means there are no basic technical errors, not that you will reach the top.

What is the difference between an SEO audit and an AI visibility audit?

They share a foundation, because Google uses its normal index for AI Overviews and AI Mode. An AI audit adds checks specific to assistants: whether robots.txt lets in crawlers such as OAI-SearchBot or PerplexityBot, whether a single passage of the page makes sense without the rest, and whether figures and authorship can be verified.

What to do next

Open the Page indexing report in Search Console and note any important URLs that are missing from the index. Then work through the checklist with one URL per template. To see how each block fits into the bigger picture, carry on with our guide to what SEO is.