The SEO audit checklist: every check, in the right order

An SEO audit checklist works best in a fixed order: crawl access first, then indexing, then on-page elements, then speed, mobile rendering, structured data, and links. Each layer depends on the one before it. If Google can't crawl a page, no title tag rewrite will save it.

This guide is the full checklist for site owners and marketers auditing their own sites. Every check names the tool that answers it, most of them free, and the same order works on a 20-page brochure site or a 20,000-page store.

Key Takeaways

• Audit in dependency order: crawl access, indexing, on-page, performance, then links. Nothing further down the list can help a page that search engines never reach.
• The most damaging failures sit early in the checklist. One agency that publishes its 50-plus-item process says it finds broken sitemaps on roughly one in four sites it audits.
• Free tools cover most of the checks: Google Search Console, PageSpeed Insights, and single-page checkers. A paid crawler starts to earn its cost once a site passes a few hundred pages.
• Fix the worst problems first. A noindex tag left over from staging costs more traffic than every missing meta description on the site combined.

What a complete SEO audit checklist covers

A full audit runs through eight areas. Each one answers a distinct question, and each has a place a wrong answer tends to hide. The table below is the map; the rest of this article walks through the checks in order.

Audit area	The question it answers	Where failures usually hide	Where to check
Crawlability	Can search engines reach and read the page?	robots.txt blocks, server errors, content that only loads through JavaScript	robots.txt, Search Console crawl stats
Indexing	Is the page actually in Google's index?	leftover noindex tags, canonicals pointing at the wrong URL, soft 404s	the URL Inspection tool in Search Console
On-page	Do the title, headings, and copy match what people search?	missing or duplicate titles, multiple H1s, thin pages	a single-page meta checker
Performance	Does the page load fast enough to keep visitors?	slow largest contentful paint, sluggish interactions, layout shift	PageSpeed Insights
Mobile	Does the page render and work on a phone?	horizontal overflow, tiny tap targets, missing viewport tag	a mobile rendering test
Structured data	Can search engines parse the page's markup?	invalid JSON-LD, missing required properties	a schema validator
Internal linking	Can crawlers and visitors reach important pages in a few clicks?	orphan pages, broken links, pages buried four clicks deep	a site crawl
Off-page	Do other sites link to yours, and do those links look natural?	spammy anchor text, sudden link spikes	a backlink tool

Get access before you start

Two accounts make every check on this list faster: Google Search Console and your analytics platform. Search Console shows you what Google actually crawled and indexed rather than what you assume it did, and several checks below have no substitute for it. If the site is verified in Bing Webmaster Tools too, you get a second crawler's opinion for free.

Pull a baseline before you change anything. Export current rankings, organic traffic by page, and the indexed page count. An audit without a baseline can't prove anything improved.

Then decide how you'll gather evidence. Moz's audit checklist notes that a single URL can be checked by hand with free tools in an hour or two. That approach holds up to maybe a dozen pages. Past that, run a crawler across the whole site so the sitewide checks (broken links, redirect chains, orphan pages, duplicate titles) come from data rather than sampling. Moz's version ships as a downloadable spreadsheet with a Pass, Fail, or N/A column per item and an importance rating of High, Medium, or Low, which is a sensible format to copy even if you build your own.

Crawlability and indexing: the checks that gate everything else

Start here every time. These checks decide whether the rest of the audit matters, and they're where the expensive failures live.

Is the page in the index at all?

The quick test is to type site: followed by the exact URL into Google. It's imprecise, though: a site: query returns everything starting with that URL pattern. The better test is the URL Inspection tool in Search Console, which tells you whether the exact URL is indexed, when it was last crawled, and which canonical Google chose for it.

While you're there, confirm the page returns a 200 status code. A 4xx or 5xx response keeps it out of the index. A 3xx means the URL you're auditing isn't the one that will rank, so trace the redirect and audit the destination instead.

robots.txt and noindex directives

Open your robots.txt file and read every line. Sites block their own blog directories or entire sections without noticing, sometimes for months, and one wrong character turns an allow rule into a disallow. Google's own documentation at Search Central covers how its crawler interprets robots.txt rules and canonical signals, and it's worth reading before you edit the file.

Then check for noindex directives in two places: the meta robots tag in the HTML, and the X-Robots-Tag HTTP header. CMS updates and staging-to-production migrations are the usual source of a stray noindex, because the tag that correctly hid the staging site follows the code to production. URL Inspection reports both in one place.

Sitemaps

An XML sitemap should exist, be referenced in robots.txt, be submitted in Search Console, and contain the URLs you actually want indexed, with none of the redirects, error pages, or noindexed URLs that dilute it. This check earns its place: Shortlist, an agency that publishes the 50-plus-item checklist it runs on client sites, says it still finds broken sitemaps on about one in four sites it audits. You can fetch and inspect any XML sitemap in the browser and read exactly what's in it before Google does.

Canonicals, soft 404s, and crawl budget

Every indexable page should carry a self-referencing canonical tag. When the canonical, the sitemap, and your internal links disagree about which URL is the real one, Google picks for you, and it doesn't always pick the one you'd want.

Watch for soft 404s: pages that return a 200 status code but show error text or empty results. Out-of-stock product pages are the most common case. Google treats them as errors, and they waste crawl attention. On larger sites, that attention has a name, crawl budget, the number of pages a search engine will fetch from your site in a given period. Shortlist gives the example of a site spending 60% of its crawl budget on filtered category pages nobody searches for while product pages waited in a queue.

Indexing blocker	Where it hides	What it does	How to find it
robots.txt disallow	one line in the robots.txt file	stops crawlers reading the content, which usually keeps the page out of results	read the file, then test the URL in Search Console
Leftover noindex	the meta robots tag or the X-Robots-Tag header	removes the page from the index entirely	URL Inspection in Search Console
Wrong canonical	the head of the page, pointing at a different URL	Google indexes the other page and ignores this one	page source, or any crawler's canonical report
Soft 404	a page returning 200 with error or empty content	Google classes the page as an error despite the status code	the page indexing report in Search Console
Redirect chain	old internal links passing through two or more hops	loses link equity at each hop and slows crawling	a redirect tracer on your key URLs
Orphan page	a page no internal link points to	crawled rarely, ranks rarely	compare the crawl list against the sitemap

On-page checks: titles, headings, and content

Once you know a page can rank, check whether it deserves to. Titles come first because they're the most visible lever: each page needs a unique title tag around 50 to 60 characters that leads with what the page is about, plus a meta description that earns the click. You can run a meta tag check on any URL and see the title, description, and heading structure the way a crawler sees them.

Headings are next. One H1 per page, and a hierarchy that descends without skipping levels, because the heading outline is how both screen readers and search engines skim the page. Then the content itself: thin pages with a few hundred generic words rarely rank against pages that answer the query fully, and images need alt text that describes what they show.

Two habits make this section faster. Before publishing anything new, work through a pre-publish on-page checklist so pages go live clean instead of getting fixed later. And for pages already live, score the on-page SEO of a URL to see which elements pass and which need work. For calibration on how granular these checks get: Backlinko's free checker runs 38 individual checks grouped into 8 areas on a single page, and scores the result out of 100.

Status codes, redirects, and HTTPS

Every HTTP URL should 301 to its HTTPS version in one hop. Crawl the site from the http:// version and watch what happens; a 302 where a 301 belongs, or a hop through two redirects, quietly leaks authority.

Redirect chains deserve their own pass. A link that goes through two or three redirects loses link equity at each hop, and chains accumulate silently over years of URL changes. Trace the redirect path of any URL and update internal links to point straight at the final destination.

Certificates fail quietly too. An SSL certificate that expires on a subdomain, an image CDN for instance, can throw browser security warnings for weeks before anyone notices, because the main domain still looks fine. Check that the certificate is valid, covers every subdomain in use, and isn't near expiry. Mixed content, HTTPS pages loading images or scripts over HTTP, belongs in the same sweep.

Speed and Core Web Vitals

Core Web Vitals are Google's three field metrics for how a page feels to load: largest contentful paint (LCP) for loading, interaction to next paint (INP) for responsiveness, and cumulative layout shift (CLS) for visual stability. The thresholds that matter: LCP within 2.5 seconds, and INP within 200 milliseconds. Shortlist puts the business case bluntly: a site that takes five seconds to load loses about half its visitors before they see anything.

Don't test every page. Run PageSpeed Insights on your most important templates, five or so, since a fix to a template fixes every page built on it. For each failing metric, the usual suspects are consistent:

• Slow LCP: oversized hero images, slow server response, render-blocking CSS and JavaScript.
• Poor INP: heavy JavaScript running when someone taps a filter, form, or accordion.
• Layout shift: images without width and height attributes, ads that load late and push content down, web fonts swapping in after render.

Image handling is the highest-yield fix on most sites. Serve modern formats such as WebP or AVIF, size images to their display dimensions, and lazy-load anything below the first screen. Then check caching: static assets should carry proper Cache-Control headers so repeat visits don't re-download every stylesheet.

Mobile checks

Google indexes the mobile version of your site, so a page that breaks on a phone is broken where rankings are decided. The audit here is short and concrete: the viewport meta tag is present, text is readable without zooming, nothing overflows horizontally at phone widths, and tap targets are large enough to hit with a thumb.

Test how a page renders on mobile rather than trusting how the desktop layout shrinks in your browser, since real rendering catches overflow and clipped content that responsive-mode previews miss. For the reasoning behind each check, and what Google actually evaluates since mobile-first indexing became the default, read this guide to what Google checks on mobile and how to pass.

Structured data checks

Structured data is the JSON-LD markup that tells search engines what a page is: an article, a product, a local business, an FAQ. The audit has three parts. First, the markup parses as valid JSON with correct @type values. Second, each type carries its required properties, because a Product without a price or an Article without a headline earns nothing. Third, the markup matches the visible page. Markup describing content the visitor can't see invites a manual action rather than a rich result.

Validate the JSON-LD on any page before and after you edit it. Missing markup is now a standard finding in mainstream crawlers too; Serpstat, for one, updated its site audit tool in July 2024 to detect missing JSON-LD markup and to crawl JavaScript-rendered pages.

Be realistic about the payoff. Some schema types produce visible rich results and some don't, so treat structured data as machine-readable labeling first and a snippet enhancement second.

Internal linking and site architecture

Internal links tell search engines which pages matter and what they're about. Four checks cover it:

1. Click depth. Your important pages should be reachable within three clicks of the homepage. Pages buried deeper get crawled less often and rank worse, and a crawler's click-depth report shows exactly which pages sit too far down.
2. Broken internal links. Every internal link that hits a 404 is a dead end for visitors and crawlers alike. Crawl the site and fix or redirect all of them.
3. Anchor text. A pricing page whose inbound links all say "click here" gets no signal from any of them. Anchors should describe the target page, in varied natural phrasing.
4. Link volume per page. A page carrying 300 or more links, usually thanks to a mega menu, dilutes the value each link passes. Trim navigation to what people use.

Orphan pages close the loop from the indexing section: if a page's only route in is the sitemap, its internal linking has failed. Doing this by hand across a large site is slow, which is why it helps to have internal link suggestions drawn from your own crawl rather than a spreadsheet of URLs you maintain manually.

Backlink checks

Off-page checks come last because you control them least. Three things are worth reviewing. Growth pattern first. A profile that grows steadily looks earned, while a sudden spike of hundreds of links from unrelated sites looks bought, whether it was or not. Then anchor distribution: a natural profile is mostly brand names and URLs, so one heavy with exact-match commercial anchors reads as manipulation. Then source quality, where a handful of links from real sites in your field outweighs hundreds from directories no human visits.

You need a backlink tool for this section since Search Console's link report is limited. Review your profile a few times a year, and always after any period of link building.

Auditing for AI search

Search results now include AI-generated answers, and AI assistants cite pages the same crawlers found. Two additions to the checklist cover it. First, access: check robots.txt for rules affecting AI crawlers, because a blanket block written years ago may now be excluding your site from AI answers along with the scrapers it was aimed at. Second, extractability: AI engines quote self-contained passages, so a page that answers its core question in a direct, liftable paragraph gets cited where a page that buries the answer doesn't. The structure this article has been recommending, clear headings, direct answers, valid schema, is the same structure AI extraction rewards. Tool vendors have followed the shift; Backlinko's free tool set now includes checkers for whether AI assistants recommend a brand and whether AI agents can read a site at all.

The tools that cover the checklist

You don't need a big budget. Here's what covers each part of the audit and what it costs.

Tool	What it covers	Cost	Best fit
Google Search Console	indexing status, sitemaps, crawl stats, manual actions	Free	every site; several checks in this list have no substitute for it
PageSpeed Insights	Core Web Vitals plus lab diagnostics per URL	Free	performance checks on your main templates
Single-page checkers (meta, schema, redirects, mobile)	one URL at a time, instantly, no signup	Free	spot checks and pre-publish reviews
Backlinko SEO Checker	38 checks across 8 areas on a single page, scored out of 100	Free, 3 checks a day	a quick scored snapshot of one page
Moz audit spreadsheet	the manual checklist with Pass/Fail status and High/Medium/Low priority per item	Free download	consultants who want a documented audit trail
Serpstat Site Audit	sitewide crawling; its July 2024 update added JavaScript rendering and missing JSON-LD detection	7-day trial, then paid	recurring sitewide audits
Desktop crawler	click depth, broken links, duplicate titles, and orphan pages at scale	varies by tool	any site past a few hundred pages

Turning your SEO audit checklist into a fix list

An audit that ends as a spreadsheet of 200 findings changes nothing. Two moves turn it into work that ships.

Group findings, then order by severity

Group identical findings into one issue. Ninety pages missing meta descriptions is one line of work, one template or one batch edit, and listing it as ninety rows only buries the finding that matters. Then order the groups by damage, and let severity override the order you found things in:

1. Crawl and index blockers: robots.txt mistakes, stray noindex tags, broken canonicals. These suppress pages entirely.
2. Sitewide technical faults: redirect chains, certificate problems, template-level speed failures. One fix, many pages.
3. Page-level on-page work: titles, headings, content depth on the pages that earn revenue.
4. Polish: alt text, minor schema warnings, individual slow images.

Moz's importance ratings of High, Medium, and Low exist for the same reason: an unweighted checklist treats a missing favicon and a site-wide noindex as equals, and they are not.

Set a cadence and re-verify

Recheck any page after you change it, and put your money pages on a monthly cycle; Backlinko's guidance on its own checker is to recheck at least once a month for key pages, because issues creep back in with every deploy. A deep sitewide audit belongs after redesigns, migrations, and platform changes, plus a scheduled pass a few times a year.

When time is short, triage beats postponing: a focused pass can find what's broken in about 60 minutes by running only the gate checks above on your top pages. And if the audit lands on your desk as a deliverable someone else produced, here's how to read an SEO report without a technical background so you can tell the findings that cost traffic from the ones that pad the page count.

FAQs

How long does an SEO audit take?

A single page can be checked by hand with free tools in an hour or two, which is how Moz frames its own checklist. A sitewide audit scales with page count: the crawl itself might run an hour on a large site, and interpreting the findings usually takes longer than collecting them. The 60-minute triage version covers only the crawl, index, and speed gates on your top pages.

How often should you run an SEO audit checklist?

Recheck any page you've just changed, and check the pages that earn you money about monthly, which matches Backlinko's advice for its own checker. Run the full sitewide checklist after a redesign, migration, or platform change, and on a schedule a few times a year even when nothing seems wrong, because that's exactly when noindex tags and broken sitemaps go unnoticed.

Can you do an SEO audit yourself, or do you need an agency?

Most of the checklist is doable yourself with free tools: Search Console for indexing, PageSpeed Insights for speed, and single-page checkers for meta tags, schema, redirects, and mobile rendering. Bring in help for the heavier work: server log analysis, international hreflang setups, large migrations, or a penalty investigation.

What's the difference between a technical SEO audit and a full SEO audit?

A technical audit covers the machinery: crawling, indexing, status codes, speed, rendering, and structured data. A full audit adds two layers on top. One is content (titles, headings, depth, internal links) and the other is off-page (your backlink profile and competitors). The technical half comes first because content fixes can't help pages search engines never see.

What's a good SEO score?

Free checkers typically score out of 100; Backlinko's, for example, runs 38 checks and scores the page 0 to 100. Treat the number as a progress marker rather than a ranking prediction. Two pages can share a score for very different reasons, so the issue list underneath the score is where the value is, and re-running the check after fixes is how you confirm the work landed.

See what a finished audit looks like

Working an SEO audit checklist by hand teaches you the site; seeing the output as a scored, prioritized document shows you what to aim for. Browse a sample SEO report to see how the checks above look once they're graded and ordered for a real site. Then talk to the team about running the full audit on your own domain.

Resources / References

1. Search Central

Sources

• Free SEO Checker: Score Any Website in 60 Seconds
• The Technical SEO (& Beyond) Site Audit Checklist
• Advanced SEO Audit Checklist For All Stages Of The ...
• SEO Audit & Checklist Template
• The Technical SEO Audit Checklist: 50+ things we actually ...

Disclaimer: This article is general information about rules and figures that applied when it was written, not personalised financial, tax or investment advice. It does not take account of your income, obligations, risk tolerance or goals, and rates and thresholds change. Verify the current position with the relevant authority, and speak to a licensed adviser or a qualified accountant before acting.