SEO Audit Checklist 2026: 69 Checks, in the Order to Fix Them

Most SEO audit checklists are long lists with no order. You get 200 rows, each marked red or green, and no way to tell that one of them (a stray noindex) is costing you the whole page while forty others are cosmetic. This checklist is built the other way round. It has the 69 checks we run in Seotific's own audit, grouped into the order you should work through them, with the reason each one matters or doesn't.

Use it for a single page you want to rank, or for a whole site before a migration, a redesign or a client handover. You can download it as an Excel SEO audit checklist template and tick it off as you go.

What an SEO audit checklist covers in 2026

An SEO audit checks whether search engines can find a page, whether they can understand what it is about, and whether it deserves to rank for the query you care about. Those three questions map onto five groups of checks: access (crawling and indexing), relevance (title, H1, keyword placement), content (depth, structure, freshness), experience (speed and mobile) and machine readability (structured data and AI crawlers).

Three things changed the checklist in the last two years, and older checklists still miss them:

• FAQ and HowTo rich results are gone. Google removed HowTo rich results in September 2023, cut FAQ rich results to government and health sites the same year, and stopped showing FAQ rich results entirely on 7 May 2026. FAQ markup is no longer a "rich result" check. It can still help AI systems read the page, so it moved to the AI search group.
• INP replaced FID as a Core Web Vital in March 2024. A checklist that still says "FID under 100 ms" is measuring a metric Google no longer reports.
• AI search is now a reporting line in Search Console. In June 2026 Google added generative AI performance reports, so you can see impressions in AI Overviews and AI Mode next to normal results. The same month, Google's AI optimisation guide stated that llms.txt has no effect on Search. That settles a debate many 2025 checklists got wrong.

The core of the checklist did not change. Google's May 2026 core update (21 May to 2 June) rewarded the same things the December 2025 update did: pages that answer the query fully and are clearly about one thing. An audit is the fastest way to find the pages that aren't.

How the checks are weighted

Every check falls into one of five impact tiers. The tier decides both its position in the fix order and how many points it costs the page score in Seotific:

The weights are why you should never sort an audit by count. Forty missing alt attributes on decorative icons score 30 points less than one wrong canonical tag, and they should. The full method for turning a raw report into a ranked list is in how to prioritize SEO issues.

The 10-step audit workflow

Run the checks in this order. Each step assumes the one before it passed, which is why the order matters: there is no point rewriting a title on a page Google cannot index.

1. Confirm the page returns HTTP 200 and is indexable: no noindex in the meta robots tag or the X-Robots-Tag header, not blocked in robots.txt, and indexed in Search Console's URL Inspection.
2. Check the canonical tag points to the page itself (or deliberately to the version you want ranked), and that the URL in the canonical returns 200 without a redirect.
3. Pick one primary keyword for the page and confirm it appears in the title tag, the H1 and the first paragraph, in words a searcher would type.
4. Check the title fits in roughly 580 to 600 pixels (about 55 to 60 characters) and the meta description is unique and around 120 to 160 characters.
5. Render the page with JavaScript off and confirm the title, H1, canonical and main text are in the raw HTML, not injected later.
6. Count internal links pointing to the page: at least three contextual links from related, indexed pages, and a click depth of three or fewer from the homepage.
7. Compare the content with the pages that rank on page one for the keyword: the subtopics, questions and formats they cover that yours doesn't.
8. Check mobile usability and Core Web Vitals from real-user data: LCP 2.5 s or less, INP 200 ms or less, CLS 0.1 or less at the 75th percentile.
9. Validate structured data that matches the page type (Article, Product, Organization, LocalBusiness) with the Rich Results Test or Schema.org validator.
10. Check AI crawler access in robots.txt (OAI-SearchBot, Claude-SearchBot, PerplexityBot), then record every finding with its tier, the fix and an owner.

Steps 1 and 2 are where most "my page isn't ranking" problems end. In audits of new sites, the most common critical finding is a staging noindex that survived launch, followed by canonical tags hard-coded to the homepage by a theme.

Indexing and crawl access: the checks that block everything

These are the critical-tier checks. A failure here means nothing else on the page matters yet, because Google either cannot see it or has chosen not to keep it.

HTTP status. The page must return 200. A 3xx means you are auditing the wrong URL; audit the destination. A soft 404 (a 200 page that says "not found" or is nearly empty) shows up in Search Console as its own status and is worth fixing on templated sites.

Indexability. Check three places for noindex: the meta robots tag, the X-Robots-Tag HTTP header, and plugins that add it per post type. Then check robots.txt does not block the URL. A blocked URL can still be indexed from links ("Indexed, though blocked by robots.txt") but Google cannot read its content, so it ranks badly. Google reads only the first 500 KiB of a robots.txt file, which matters on large sites with generated rules.

Index status. A page can pass every technical check and still not be indexed. Search Console's Page indexing report gives the reason. The two you will see most on a new site are crawled, currently not indexed (Google fetched it and decided not to keep it) and discovered, currently not indexed (Google knows the URL but hasn't fetched it yet). They have different fixes: the first is about page quality and duplication, the second about crawl demand and internal links.

Canonical tags. A canonical that points to another URL tells Google to fold this page into that one. It is a hint, not a directive, but if it points somewhere else the page will usually not rank for its own keyword. The usual causes and fixes are in canonical tag points to another page.

XML sitemap. List only canonical, indexable URLs that return 200. A sitemap file can hold 50,000 URLs or 50 MB uncompressed. Listing noindex or redirected URLs in a sitemap sends Google mixed signals and wastes the crawl you have.

This group has its own deep guide with 30-plus technical checks, log-file signals and Search Console statuses: the technical SEO audit checklist.

On-page relevance: does the page match the keyword?

Once a page can be indexed, the next question is whether it is clearly about the query. This is where a keyword-aware audit differs from a generic one. Without a target keyword, a tool can tell you the title is 52 characters long. With one, it can tell you the title never mentions what people search for.

The high-tier checks in this group:

• Keyword in the title tag, ideally near the start. The title is still one of the strongest on-page relevance signals and the headline people click.
• Keyword in the H1, in the words a visitor would use. "Our collection" is not a heading; "Men's trail running shoes" is.
• Keyword in the opening copy, in the first paragraph after the heading.
• One clear H1. Google handles several H1 tags without a penalty, but a missing H1 is a real gap. The detail is in multiple H1 tags and SEO.

Medium and low tier, worth fixing after the above:

• Title length. Google cuts title links at about 600 pixels on desktop, so "55 to 60 characters" is an approximation. A title of wide capital letters truncates sooner. See title tag length in pixels.
• Meta description. Not a ranking factor, and Google rewrites it often, but a unique one helps click-through. Length warnings are mostly noise; meta description too long explains when to bother.
• Keyword in the URL matters only for new pages. Never rename a live URL that already ranks just to add a keyword; the redirect costs more than the gain.
• Keyword density is not a check you need. There is no target percentage. Seotific flags density only when the keyword is missing entirely or repeated so often the copy reads as stuffed. More in keyword density and SEO.

Content quality and depth

Content checks are where audits are least precise, because "is this page good enough?" is a judgement. A checklist can still catch the measurable parts.

Opening answer. Does the first screen of text answer the query? Pages that bury the answer under 300 words of preamble lose both readers and AI Overview citations.

Content depth versus the ranking pages. Word count is not a ranking factor (Google has said so repeatedly, and does word count matter for SEO covers the evidence). What matters is whether the page covers what the top results cover. A product category page with 140 words of text, competing against pages that explain sizing, materials and price ranges, will lose for reasons a word counter cannot see. A competitor gap check compares headings, subtopics and structured data against the actual page-one results.

Structure. Steps in numbered lists, comparisons in tables, one idea per section. This is a medium-tier check because it makes passages extractable, which is how both featured snippets and AI answers quote a page.

Freshness. For time-sensitive topics, show an updated date and set dateModified in the JSON-LD. For evergreen product or service pages, a date is optional; Seotific rates freshness low unless the query is news-like.

Trust signals. On articles: an author or organisation, a date, and links to sources. Google's quality raters look for these, and they cost little to add.

Speed, mobile and page experience

Core Web Vitals are a confirmed ranking signal, but a small one. Fix them after relevance, not before, unless a page is so slow people leave.

Use real-user (field) data from the Chrome UX Report first; that is what Google uses. Lab scores from Lighthouse are for debugging. The thresholds at the 75th percentile of visits:

• Largest Contentful Paint: 2.5 seconds or less
• Interaction to Next Paint: 200 milliseconds or less
• Cumulative Layout Shift: 0.1 or less

The checks that usually move these numbers: serve large images as WebP or AVIF, set width and height on images, lazy-load anything below the first screen but load the hero image eagerly, and defer scripts in the <head>. Mobile: the page needs a responsive viewport tag, because Google has indexed the mobile version of every site since mobile-first indexing finished rolling out in 2024.

Structured data and AI search readiness

Structured data no longer earns as many visual rich results as it did (FAQ and HowTo are gone), but it still makes a page eligible for product, review, breadcrumb and organisation features, and it gives AI systems a clean description of what the page is.

The checks:

• Schema matches the page. Article for articles, Product for product pages, LocalBusiness only for local businesses, Organization on the homepage with name, logo and sameAs links.
• Required properties are present for any rich result type you use. Re-test in Google's Rich Results Test after changes.
• AI crawler access. robots.txt rules for OAI-SearchBot (ChatGPT search), Claude-SearchBot, PerplexityBot and Google's crawlers. Blocking a training crawler like GPTBot is a business decision; blocking a search crawler removes you from that product's answers.
• llms.txt is optional. No major search engine or assistant has confirmed using it, and Google says it has no effect on Search.

You can test all of these on any site with the free AI search readiness checker, and the schema markup for AI search guide covers which types are worth adding.

Platform-specific checklists

The same 69 checks apply on every platform, but each CMS breaks them in its own way. The fixes are different enough to need their own pages:

• WordPress: plugin conflicts that output two canonical tags or two meta descriptions, attachment pages, and tag archives. See the WordPress SEO audit checklist.
• Shopify: duplicate product URLs under /collections/, the fixed /products/ and /collections/ URL structure, and theme apps that slow the page. See the Shopify SEO audit checklist.
• Webflow: the webflow.io staging subdomain getting indexed, CMS collection pages with empty fields, and missing alt text on CMS images. See the Webflow SEO audit checklist.
• Next.js and React: client-side rendering that leaves the title and H1 out of the HTML, missing metadata in the App Router, and robots.ts mistakes. See the Next.js SEO checklist.

Which audit approach fits which job

There is no single right way to run an audit. The method depends on how many pages you have and what you need to hand over at the end.

Many teams use two of these: a suite for keyword and link research, and a page-level audit to decide what to change on each page.

Common SEO audit mistakes and how to avoid them

Fixing by count instead of impact. Wrong: "We have 312 issues; let's start with the 140 missing alt attributes because that's the biggest number." Right: "We have one canonical pointing at the homepage on 40 product pages. That is the only critical issue; it goes first, and the alt text goes in the backlog."

Auditing without a keyword. Wrong: auditing /services/ and reporting "title 48 characters, H1 present, meta description present: pass." Right: auditing /services/ against "commercial roof repair" and reporting that neither the title nor the H1 contains it, so the page is not about the query it is meant to rank for.

Treating every warning as an error. Wrong: rewriting 600 meta descriptions because a tool flagged them at 163 characters. Right: rewriting the descriptions on the 20 pages with high impressions and low click-through in Search Console, where a better snippet can actually change traffic. Detail in are SEO audit warnings important.

Renaming URLs for keywords. Wrong: changing /blog/post-123 to /blog/keyword-rich-slug on a page that already ranks, with a redirect. Right: using keyword URLs for new pages, and leaving ranking URLs alone unless the site is being restructured anyway.

Auditing once. Wrong: a one-off audit in January and nothing until the next problem. Right: a re-audit after each release and a monthly crawl, so a template change that adds noindex to 300 pages is caught in days, not at the next traffic drop.

Tools and resources for an SEO audit

You need fewer tools than most checklists suggest.

Google Search Console is the only source of truth for indexing. The Page indexing report and URL Inspection tell you what Google decided; no third-party tool can. The new generative AI performance report shows impressions in AI Overviews and AI Mode.

PageSpeed Insights gives real-user Core Web Vitals from the Chrome UX Report alongside a Lighthouse lab run.

Rich Results Test and the Schema.org validator check structured data. The first tells you whether Google can use it; the second whether it is valid at all.

A crawler (Screaming Frog's free version handles 500 URLs) finds sitewide problems: duplicate titles, orphan pages, redirect chains, broken internal links.

Seotific runs the page-level checks in one pass with a target keyword, renders JavaScript in a real browser, weights each finding by impact and writes replacement titles, descriptions, headings and structured data. Whole-site crawls and scheduled re-audits are on the paid plans, and reports export to PDF and Excel with your own branding if you audit for clients. If the audit is for a client, the SEO audit report template covers how to present it.

The free Excel checklist has all 69 checks with a status column for each.

Frequently asked questions