Technical SEO Checklist (2026): 36 Checks for Any Website
36 checks on crawling, indexing, rendering, speed, mobile, security, structured data, international setup and AI crawler access. No keywords, no content.
- Updated
- 23 min read
By Szymon Kokot Published Updated 33 min read
SEO Checklist (2026): 54 Checks You Can Run Yourself
An SEO checklist for a business website covers 54 checks in nine areas: technical SEO, on-page SEO, keywords, backlinks, competitors, trust, local presence, user experience and AI search visibility. This checklist gives you each check in order, with the free tool to use, what a good result looks like and what to write down. It follows the same 54 points as the Illucrum Website SEO Audit, so it is the real method, not a shortened version of it.
In short: A website SEO checklist should test whether search engines can reach and index the site, whether each page says clearly what it is about, what the site already ranks for, how its links and content compare with competitors, whether visitors can trust and contact the business, and whether AI assistants can read and cite it. All 54 checks below run on free tools.
What does this SEO checklist cover?
This SEO checklist covers 54 checks in 9 sections. It is built for service businesses, consultancies, agencies, B2B firms and any other site that is not a software product, an online store or a mainly local business. Those three have their own checklists: SaaS, e-commerce and local SEO.
| # | Section | Checks | The question it answers |
|---|---|---|---|
| 1 | Technical SEO | 1 to 12 | Can search engines reach, load and index the site? |
| 2 | On-page SEO | 13 to 22 | Does each page say clearly what it is about? |
| 3 | Keyword opportunities | 23 to 28 | What does the site rank for, and where are the nearest gains? |
| 4 | Backlink profile | 29 to 33 | Which sites link to yours, and how healthy do those links look? |
| 5 | Competitor comparison | 34 to 36 | What do the sites that outrank you have that you don't? |
| 6 | Trust and E-E-A-T | 37 to 41 | Can a visitor see who is behind the business and believe it? |
| 7 | Local presence | 42 to 44 | Does the business show up for searches in its area? |
| 8 | UX and conversion | 45 to 48 | Can a visitor from search find their way and take the next step? |
| 9 | AI search visibility | 49 to 54 | Can AI assistants reach, read and cite the site? |
If the business has no office or service area it wants to be found in, mark section 7 as N/A. The checklist then has 51 checks.
What do you need before you start?
You need verified access to Google Search Console and Google Analytics 4 for the site, plus a handful of free tools. Without Search Console, about a quarter of the checks can only be estimated from public data. Without Analytics, check 47 can't be done at all.
| Tool | Used for | Cost |
|---|---|---|
| Google Search Console | Indexing, queries, positions, links, manual actions, Core Web Vitals | Free, needs a verified site |
| Google Analytics 4 | Engagement of visitors from organic search | Free, needs property access |
| PageSpeed Insights | Performance scores and Core Web Vitals | Free |
| Illucrum Bulk SEO Checker | Titles, descriptions, headings, canonicals, links, images, word counts and schema for up to 25 URLs at a time, exported to CSV | Free, no account |
| Chrome DevTools | Mobile view, network requests, redirects, headers, console warnings | Free, built into Chrome |
| Rich Results Test and Schema Markup Validator | Structured data | Free |
| SSL Labs Server Test | Certificate and HTTPS configuration grade | Free |
| A backlink tool | Referring domains and anchors for your site and competitors | Free tiers exist; use the same tool for every domain |
| A desktop crawler such as Screaming Frog | Site-wide broken links and redirect chains on sites with more than a few dozen pages | Free up to 500 URLs |
Four decisions come before the first check:
- Open a spreadsheet with four columns: check number, finding, evidence (a URL, a number or a screenshot) and severity. Fill it as you go.
- Pick a page sample and use it for every check: the homepage, your two or three most commercially important pages, the contact page and one article.
- List 5 to 10 target queries, the non-branded searches you want to be found for, taken from Search Console and from how customers describe what you sell.
- Name three search competitors. Search your target queries in a private window. The three domains that appear most often in the top five results are your search competitors, even if they aren't the businesses you think of as competitors.
The Bulk SEO Checker saves most of the copying in sections 1 and 2. Paste your sample pages, or up to 25 URLs from your sitemap, and it returns one row per page with 51 columns. It reads the HTML your server sends and doesn't run JavaScript, which is also how many crawlers see a page.
Section 1: How do you check technical SEO? (checks 1 to 12)
Technical SEO checks establish whether search engines can reach, load and index the site. They come first because a well-written page that Google can't crawl or index has no chance to rank.
1. Page speed (mobile and desktop)
Test page speed by running each sample page through PageSpeed Insights twice, once for mobile and once for desktop, and recording both Performance scores. Judge the site on the mobile score of its weakest key page, because mobile is almost always the lower figure and Google indexes the mobile version of a site.
Good result: a mobile Performance score of 90 or above on every key page.
Record: mobile and desktop scores per page, and the top three opportunities PageSpeed lists.
2. Core Web Vitals (LCP, INP, CLS)
Core Web Vitals are Google's three measures of real visitor experience: Largest Contentful Paint (loading), Interaction to Next Paint (responsiveness) and Cumulative Layout Shift (visual stability). Read them from the field data at the top of each PageSpeed report, which comes from real Chrome users. If there is no field data, the site has too little traffic for Google's dataset; use the lab data and note it. Then open Search Console, Core Web Vitals, for the site-wide view.
Good result: LCP of 2.5 seconds or less, INP of 200 milliseconds or less and CLS of 0.1 or less, as defined by Google.
Record: the three values per page, whether they are field or lab data, and any URL groups Search Console marks as Poor.
3. Mobile usability
Check mobile usability by hand, because Google retired its Mobile-Friendly Test in 2023. Open each sample page in DevTools device mode at 390 pixels wide, then on a real phone. Look for sideways scrolling, text too small to read without zooming, and buttons, links or form fields too small or close together to tap. Confirm the viewport tag is in the page source: <meta name="viewport" content="width=device-width, initial-scale=1">.
Good result: every sample page reads and works on a phone without zooming or sideways scrolling.
Record: per page, whether the layout holds, with screenshots of anything broken.
4. HTTPS and SSL certificate
Confirm the site loads over HTTPS and that the http:// version redirects to https:// in a single hop, using the DevTools Network tab. Check the certificate's expiry date, run the domain through SSL Labs for a configuration grade, and open the DevTools Console on each sample page to catch mixed content warnings (files loaded over plain HTTP).
Good result: HTTPS on every page, one-hop redirect from HTTP, SSL Labs grade A or A+, no mixed content.
Record: redirect behavior, certificate expiry date, SSL Labs grade and any mixed content URLs.
5. XML sitemap
The XML sitemap is the file that lists the pages you want search engines to find, usually at /sitemap.xml. Confirm it returns status 200 and is valid XML. Paste its URLs into the Bulk SEO Checker in batches of 25: every listed URL should return 200, carry a canonical pointing to itself and have no noindex. Then confirm the sitemap leaves out thank-you pages, admin pages and parameter URLs, and that it is submitted in Search Console under Sitemaps with the status Success.
Good result: a valid sitemap that lists only live, canonical, indexable URLs and is submitted in Search Console.
Record: URL count, URLs that redirect, error or carry noindex, and the Search Console status.
6. Robots.txt
Robots.txt is the file at /robots.txt that tells crawlers which parts of a site they may visit. Open it in a browser and read all of it. Confirm it returns 200. Look for Disallow: / under User-agent: *, which blocks the whole site, and for rules that block key sections or the CSS and JavaScript files needed to display pages. Confirm a Sitemap: line points to the live sitemap. Rules for AI crawlers are covered in check 54.
Good result: robots.txt returns 200, blocks nothing important and declares the sitemap.
Record: the status code, the full file, any rule that blocks content, and whether the sitemap is declared.
7. Canonical tags
A canonical tag (<link rel="canonical">) tells search engines which address is the main version of a page. The Bulk SEO Checker shows the canonical of every URL you paste; each should be an absolute HTTPS URL pointing to the page itself. Then test the duplicate paths of one key page: with and without a trailing slash, with and without www, on http, and with ?utm_source=test added. Each variant should redirect to the main URL or carry its canonical.
Good result: self-referencing absolute canonicals on every page, and every duplicate path consolidating to one URL.
Record: pages with missing, relative or wrong canonicals, and any duplicate path that returns 200 with no canonical.
8. Structured data
Structured data is code (usually JSON-LD) that describes a page to search engines in a fixed vocabulary. Run the homepage and one page of each type through the Rich Results Test and the Schema Markup Validator, and record the types found, such as Organization, Service, Article and BreadcrumbList. Confirm the markup matches what the page visibly says. Don't count FAQ markup as a benefit: Google stopped showing FAQ rich results on May 7, 2026.
Good result: relevant types on every key page type, no errors, and nothing in the markup that the page doesn't show.
Record: types found per page type, errors, warnings and page types with no structured data.
9. Index coverage
Index coverage is the count of pages Google has stored and can show in results. In Search Console, open Indexing, then Pages, and record the indexed and not indexed totals and the reasons listed, especially "Crawled, currently not indexed", "Discovered, currently not indexed", "Soft 404" and "Excluded by noindex tag". Run URL Inspection on each sample page. Without Search Console, search site:yourdomain.com/page-path for each key page instead.
Good result: every key page indexed, and every excluded page excluded on purpose.
Record: indexed and not indexed counts, the reason breakdown and any key page that isn't indexed.
10. Redirect chains and broken links
A redirect chain is a redirect that passes through more than one step, for example A to B to C. On a small site, export every internal link from the Links table of the Bulk SEO Checker and paste the link targets back in: anything that errors or redirects is a finding. On a larger site, a desktop crawler lists broken links, chains and loops in one pass. Confirm permanent moves use 301 or 308 redirects, and request a page that doesn't exist to confirm the site returns a real 404 status rather than a 200.
Good result: no broken internal links, no chains or loops, and a real 404 for missing pages.
Record: broken URLs with the pages linking to them, each chain with its steps, and the status code of a missing page.
11. Duplicate and thin pages
Duplicate pages repeat each other; thin pages say too little to be worth ranking. Sort the Bulk SEO Checker export by title and by meta description to find duplicates, then sort by word count and open the shortest pages. Also check whether tag archives, author archives or parameter URLs are indexed with a site: search. Judge thin pages against their job: a 200-word contact page is fine, a 200-word service page is not.
Good result: every indexed page has a distinct purpose and enough unique content to serve it.
Record: pages sharing titles or descriptions, pages that are thin for their purpose, and indexed archive or parameter URLs.
12. Hreflang (multilingual sites)
Hreflang tags tell search engines which language or country version of a page to show each visitor. If the site has one language for one market, mark this check N/A. Otherwise, view the source of a page that exists in several languages and check that the language and region codes are valid, that each version links to all the others and to itself, that an x-default exists, and that every target URL returns 200.
Good result: valid codes, reciprocal links between versions, a self-reference and an x-default on every version.
Record: the hreflang set per sampled page, missing return links, invalid codes and broken targets.
Section 2: How do you check on-page SEO? (checks 13 to 22)
On-page SEO checks look at the signals on each page that tell search engines and visitors what the page is about. Most on-page problems come from a template, so one fix usually repairs many pages at once.
13. Title tags
The title tag is the page title shown as the headline of a search result. From the Bulk SEO Checker export, check every page has a title, no two pages share one, and each title names the page's topic before the brand. Aim for roughly 50 to 60 characters so it isn't cut off; the checker counts characters, while Google cuts titles by pixel width, so treat the number as a guide. Google rewrote about a third of title tags in an Ahrefs study, most often when they were vague or stuffed.
Good result: unique, descriptive titles of about 50 to 60 characters, with the main topic first.
Record: missing, duplicate, overlong and brand-only titles.
14. Meta descriptions
The meta description is the short summary that often appears under a search result. Check the export for missing and duplicate descriptions on key pages, and read the ones that exist. A description is not a ranking factor and Google rewrites most of them, so write it for the click on your key pages rather than rewriting every page on the site.
Good result: a unique, specific description of about 140 to 160 characters on every key page.
Record: missing and duplicate descriptions on key pages, and any that are generic.
15. Heading structure (H1 to H3)
Headings are the outline of a page: one H1 for the topic, H2s for sections, H3s for subsections. The Headings table of the Bulk SEO Checker lists every heading in order. Check each sample page has one H1 that names its topic in plain words, and that H2s and H3s follow in a logical order. Read the H2s on their own; "Our approach" and "Why choose us" tell a reader and a search engine nothing.
Good result: one descriptive H1 per page and an H2 outline that makes sense read alone.
Record: pages with no H1 or several H1s, skipped levels, and slogan H1s that don't name the topic.
16. Image alt text
Alt text is the written description of an image, read by screen readers and search engines. The Images table of the Bulk SEO Checker separates a missing alt attribute from a deliberately empty one, which are different things. Meaningful images need a short, specific description; decorative images should have alt="". Flag file names used as alt text ("IMG_4471.jpg") and lists of keywords.
Good result: every meaningful image described in a short phrase, and decorative images marked empty.
Record: images with missing alt text, file names or keyword lists as alt text.
17. Content quality and depth
Content quality is whether a page fully answers what a visitor arriving from search came to find. Read each key page as a prospective customer would. Does it explain the process, what is included, typical timelines or how price is set? Does it contain anything specific, or could a competitor publish it unchanged? Word count is a symptom, not a target: the average top-10 result is about 1,447 words in a Backlinko study, but it found no direct link between word count and rankings.
Good result: key pages that answer the obvious questions with specifics a competitor couldn't copy.
Record: per key page, a one-line verdict and the questions it leaves unanswered.
18. Internal linking
Internal links are links from one page of your site to another. They tell search engines which pages matter and help visitors move through the site. Use the Links table of the Bulk SEO Checker to see every link on your sample pages with its anchor text. Check key pages are linked from related content, not only from the menu and footer, and that anchors describe the target ("website SEO audit") rather than saying "read more". Compare your sitemap with the pages you can reach by clicking; a page nothing links to is an orphan page.
Good result: key pages linked from related content with descriptive anchors, within three clicks of the homepage, and no orphan pages.
Record: links pointing to each key page, generic anchors and orphan pages.
19. URL structure
Check that page addresses are lowercase, use hyphens, describe the page and avoid unnecessary parameters and IDs, and that similar pages follow one pattern. The usual finding is inconsistency: some URLs with trailing slashes and some without, or dates in some paths. Changing working URLs has a cost, so recommend it only for serious problems, and always with 301 redirects.
Good result: short, readable URLs that follow one pattern per page type.
Record: the pattern per page type and every place it breaks.
20. Open Graph and social previews
Open Graph tags (og:title, og:description, og:image, og:url) control how a page looks when shared on LinkedIn, Facebook, WhatsApp and similar apps. The Bulk SEO Checker export includes every Open Graph and Twitter tag. Confirm each key page has them, that og:image loads, and that it is about 1,200 by 630 pixels. Paste one URL into LinkedIn Post Inspector to see the real preview.
Good result: complete tags and a correct preview image on every key page.
Record: missing tags per template, broken images and pages sharing one generic image.
21. Content freshness
Content freshness means pages are accurate today, not that they were edited recently. Check the visible dates on articles and key pages and the date of the most recent article. Look for outdated facts: old prices, past years written as current, services you no longer offer, staff who have left, statistics that have been replaced. Evergreen pages don't need new dates; the date should change because the content did.
Good result: no outdated facts on key pages, and dates that reflect real updates.
Record: the date of the latest article and every page with outdated information.
22. Keyword placement
Keyword placement checks whether each page uses the words its audience searches for in the places that matter. Match each sample page to its target query, then check the query, or a natural variant, appears in the title, the H1, the first paragraph, at least one H2 and the meta description. There is no ideal keyword density. The Words table of the Bulk SEO Checker shows the 20 most repeated words per page, which is a quick way to see whether a page is about what you think it is about.
Good result: the topic is clear in the title, H1, opening and headings, and the page still reads naturally.
Record: per page, the target query and which of the five positions it appears in, plus any page that reads as stuffed.
If working through this on your own site is more time than you want to spend, the Website SEO Audit covers the same 54 points and ends in a scored report with the fixes in order.
Section 3: How do you check keyword opportunities? (checks 23 to 28)
Keyword checks show what the site already ranks for, where the nearest gains are and whether each page has a clear job. Use the last three months of Search Console data, under Performance, Search results, for every check in this section.
23. Ranking keywords overview
Switch on all four metrics (clicks, impressions, CTR and average position) and filter out branded queries (see check 28). Count the non-branded queries with impressions and how many rank in the top 10. Then read the top 15 by impressions and ask whether they come from people who could buy from you, or are accidental matches.
Good result: 20 or more relevant non-branded queries, several of them on page one.
Record: the number of relevant non-branded queries, the number on page one, and the top 15 with position and CTR.
24. Target query positions and striking distance
Record the average position and the ranking page for each target query from your setup. Then filter all queries to a position greater than 3 and less than 21 and sort by impressions. These striking-distance queries already rank close to the top, so improving the page that ranks for them is usually the most efficient work available.
Good result: most target queries in the top three.
Record: each target query with its position and ranking page, and the striking-distance list.
25. Keyword mapping and cannibalization
Keyword cannibalization is when two or more of your pages compete for the same search and hold each other back. Click a target query in the Performance report, then open the Pages tab: if several pages share its impressions, or the ranking page keeps changing, they are competing. Also check each key page has one clear topic that no other page targets.
Good result: one page per important topic, with no pages splitting the same query.
Record: each query where pages compete, with the pages involved.
26. Search intent alignment
Search intent is what a searcher wants from a query: to buy, to learn or to compare. Search each target query in a private window and note what ranks: service pages, guides, comparisons or directories. Then compare with the page you point at the query. A service page competing against a results page full of guides is misaligned, and on-page work won't fix it.
Good result: each key page is the same kind of page that ranks for its query.
Record: per query, the dominant result type, your page type and whether they match.
27. Keyword gap vs competitors
A keyword gap is a set of searches where competitors rank and you don't. Search your target queries and 5 to 10 related service queries, and note where your three competitors rank and you are missing. Open their ranking pages and group the gaps into topics or page types rather than single keywords.
Good result: no major topic where competitors rank and you have no page.
Record: the gap topics, with the competitor page that ranks for each.
28. Branded vs non-branded split
Branded queries include your business name; non-branded queries don't. Filter queries with a regular expression that matches your brand name, common misspellings and founder names, record the clicks, then invert the filter. A site where most clicks are branded isn't finding new customers through search. For context, SparkToro and Datos found that slightly more than 44% of Google searches are branded.
Good result: more than half of search clicks from non-branded queries.
Record: branded clicks, non-branded clicks and the percentage split.
Section 4: How do you check the backlink profile? (checks 29 to 33)
Backlink checks look at the links from other websites to yours: how many, from where and how healthy they look. Record every figure with today's date. Search Console's Links report is a sample of Google's own data with no history, so this check produces a snapshot you can compare against later, not a trend.
29. Referring domains and link diversity
A referring domain is a website that links to yours at least once. In Search Console, open Links and record the top linking sites. Read the list: are links spread across real, relevant sites, or concentrated in one directory network, sitewide footer credits or a single partner? Record the share of links coming from the top five domains.
Good result: links from a spread of real, relevant sites, with no single source dominating.
Record: the number of linking sites, the top 20 with a note on what each is, and the top-five share.
30. Link quality and unnatural patterns
First open Security and Manual actions in Search Console. A manual action for unnatural links is the one case where link cleanup is urgent. Then scan the linking sites for signs of links someone built or bought: many exact-match commercial anchors from unrelated blogs, networks of similar sites, sitewide links with keyword anchors. Random spam you didn't create is normal, and Google says it ignores most of it; a disavow file is only worth it for a large pattern of artificial links that has caused, or is likely to cause, a manual action.
Good result: no manual action and no sign of built or bought links.
Record: manual action status and any suspicious pattern, with example domains.
31. Anchor text distribution
Anchor text is the clickable text of a link. In Search Console, open Links, then Top linking text, and sort the anchors into four groups: branded (your business name), bare URLs, generic ("here", "website") and exact-match commercial phrases ("accounting firm Boston"). A profile that was earned is mostly branded and URL anchors.
Good result: mostly branded and URL anchors, with few exact-match commercial phrases.
Record: the approximate share of each group.
32. Backlinked pages and broken link targets
Some pages other sites link to may no longer exist. In Search Console, open Links, then Top linked pages, and paste the URLs into the Bulk SEO Checker. Any that return 404, 410 or 5xx, or redirect to something irrelevant like the homepage, are wasting links. A 301 redirect from each broken URL to its closest live equivalent recovers them, which makes this the quickest link improvement most sites have.
Good result: every linked URL loads, or redirects once to a relevant page.
Record: broken or poorly redirected linked URLs, with the number of sites linking to each.
33. Competitor link gap
Run your domain and your three competitors through the same backlink tool on the same day and record each one's referring domains. Third-party figures are estimates, so compare them with each other, never as absolute numbers. Then look for domains that link to two or more competitors but not to you; those are the most realistic places to earn links.
Good result: referring domains within about 20% of the median competitor.
Record: referring domains per domain, the tool and date, and the list of gap domains grouped by type.
Section 5: How do you check competitors? (checks 34 to 36)
Competitor checks compare your site with the three domains that win your target searches. Do them in a private window so your own search history doesn't change the results.
34. Authority comparison
Authority scores such as Domain Strength, Domain Rating or Authority Score are third-party estimates of a site's overall strength in search; Google doesn't use them. Record the score for your site and each competitor from the same tool on the same day. The number is only useful as calibration: it tells you whether you are competing with sites of similar weight or with sites years ahead of you.
Good result: within about 10 points of the strongest competitor on the same scale.
Record: each score, the tool and date, and the gap to the strongest competitor.
35. Page-type comparison
List the kinds of pages each competitor has: individual service pages, industry pages, location pages, case studies, pricing, FAQs, guides, comparisons. Compare with your own list. A page type that recurs across competitors and is missing from your site is usually the clearest opportunity in the whole checklist, because one page covering every service can't rank for each service separately.
Good result: you have every page type that most competitors rely on.
Record: a page-type inventory for each competitor and for you, with the gaps marked.
36. Content gap vs competitor pages
For two or three key services, open the competitor page that ranks best for the target query and read it next to yours. Note what it covers that yours doesn't: pricing or how price is set, process, timelines, what's included, objections, FAQs, proof. This is a reading exercise, not a word count, and it usually produces the most useful list in the audit.
Good result: your pages answer everything the best competitor pages answer.
Record: per key page, the specific sections and facts the competitor page has and yours lacks.
Section 6: How do you check trust and E-E-A-T? (checks 37 to 41)
E-E-A-T stands for experience, expertise, authoritativeness and trustworthiness, the qualities Google's quality rater guidelines ask about. Trust checks look for evidence a visitor can verify, not for design. They matter most for advice-led businesses such as legal, financial, health and consulting firms.
37. About page and business identity
Find the About page and check it is linked from the menu or footer. It should name the people or company behind the business, say what the business does, for whom, where and since when, and show real photos rather than stock images. Check the claims can be verified: named founders with LinkedIn profiles, company registration details, memberships.
Good result: an About page that names real people and gives details a visitor can check.
Record: whether the page exists, who it names and which claims can be verified.
38. Authorship and expertise
If the site publishes no articles, guides or advice, mark this check N/A. Otherwise, check each piece has a named human author, not "Admin" or "Team", and that the byline leads to a bio stating relevant experience in specifics: roles, years, qualifications, published work. For advice on health, money, law or safety, check the author or a named reviewer is qualified to give it.
Good result: named authors with specific, checkable bios on every article.
Record: the share of articles with a named author and the state of the bios.
39. Proof of work
Proof of work is evidence of what the business has actually done. Look for case studies that say what the situation was, what was done and what changed; testimonials with a real name and company; third-party reviews on Google, Clutch, Trustpilot or a sector platform; and credentials with a link to the register where one exists. Only show client logos with permission.
Good result: specific case studies or examples plus reviews or credentials a visitor can verify.
Record: each type of proof found, and review ratings, counts and latest dates.
40. Contact and company details
Check a contact page or contact details can be reached from every page, the business offers more than one way to get in touch, and the contact form works. Check the legal company name, registered address and registration or tax number are shown where the law requires it (common in the EU) and wherever they help trust. Check the site says what happens after someone gets in touch.
Good result: easy to reach from every page, complete company details and a clear next step.
Record: contact methods, company details shown, and the result of a test form submission.
41. Legal and policy pages
Check a privacy policy is linked from every page if the site collects any personal data, including through forms or analytics, and that terms and cookie information exist where they apply. If there is a cookie banner, open DevTools, reload the page before accepting, and watch the Network tab: in the EU, analytics and advertising tags shouldn't fire before consent. This is a hygiene check, not legal advice.
Good result: required pages present and linked, and a cookie banner that works as it says.
Record: pages found, and any tag that fires before consent.
Section 7: How do you check local presence? (checks 42 to 44)
Local presence checks apply only if the business has an office or a service area where it wants to be found; otherwise mark all three N/A. This is a light version of the local layer. For businesses that depend on local search, the local SEO checklist covers it in 54 checks.
42. Google Business Profile
Search the business name and find its Google Business Profile, the listing shown in Google Maps and in the map results. Check it is claimed (no "Own this business?" link), the name matches the real business name with no added keywords, the primary category fits the main service, and the address, hours, phone and website are correct. Record the rating, review count, date of the latest review and whether the owner replies.
Good result: a claimed, accurate, complete profile with recent reviews and replies.
Record: each field's state, the primary category, and the review figures.
43. NAP consistency and local schema
NAP stands for name, address and phone number. Collect every place they appear on the site (footer, contact page, About page, structured data) and compare them character by character with the Business Profile. Then check for LocalBusiness markup, or a more specific type, with name, address, telephone, url and openingHoursSpecification, and confirm the values match the profile.
Good result: identical NAP everywhere, matching the profile, and valid local markup.
Record: every NAP instance that differs, and the local markup fields present or missing.
44. Local search visibility
Local results depend on where the searcher is, so set a search location first: in Chrome DevTools, open Sensors and set a custom location in the business's area, or use a free local rank checker. Search three to five "service + city" and "service near me" queries and record whether the business appears in the map pack (the three map results) and in the regular results. In Search Console, filter queries containing the city name.
Good result: in the map pack for most core local searches.
Record: each query with map pack and organic positions, the location method and the date.
Section 8: How do you check UX and conversion? (checks 45 to 48)
UX and conversion checks establish whether visitors who arrive from search can find their way and take the next step. Rankings that don't lead to inquiries are an expensive hobby.
45. Calls to action
On each sample page, find the main call to action (the button or link you want a visitor to use) and check it is visible without scrolling on desktop and on mobile. Read the wording: "Request a quote" says what happens, "Submit" and "Learn more" don't. Count competing buttons in the same view; three different offers side by side usually means no offer.
Good result: one clear, specific call to action above the fold, repeated on long pages.
Record: per page, the main action, its exact wording, its position and the number of competing actions.
46. Navigation and site structure
Count the top-level menu items; five to seven, labeled in the words customers use, is a sensible range. Count the clicks from the homepage to each key page (two or fewer) and to the contact page (one). Test the menu on a phone separately. Where the site has a real hierarchy, check breadcrumbs are shown.
Good result: a short, clear menu, key pages within two clicks and a mobile menu that works.
Record: menu item count, clicks to each key page and any mobile menu problem.
47. Engagement (GA4)
If you have no Google Analytics access, mark this check N/A. In GA4, open Reports, Engagement, Landing page, set the last three months and filter to the Organic Search channel. Record the engagement rate and average engagement time for each key landing page. GA4 counts a session as engaged if it lasts more than 10 seconds, includes a key event or views two or more pages. Compare each page with the site's own average rather than an industry benchmark.
Good result: no key landing page far below the site's own average engagement rate.
Record: engagement rate and average engagement time per key landing page, organic traffic only.
48. Intrusive pop-ups and interstitials
Open each key page on a phone from a Google result, as a new visitor would, in a private window so cookies don't hide pop-ups. Note anything that covers the content on arrival or shortly after. A standard cookie notice is fine. A full-screen newsletter or discount overlay shown to someone arriving from search is something Google's page experience guidance names as a problem.
Good result: nothing covers the content on arrival except a consent notice.
Record: which pages show overlays, when they appear and how much of the screen they cover.
Section 9: How do you check AI search visibility? (checks 49 to 54)
AI search visibility checks establish whether Google's AI Overviews and assistants such as ChatGPT, Perplexity and Gemini can reach the site, understand it and cite it. Run every test logged out or in a temporary chat, and record the date, because answers vary between runs and change over time. The GEO checklist covers this subject in 51 checks.
49. Google AI Overview presence
An AI Overview is the AI-generated summary Google shows above some search results. Search each target query in a private window and note whether an AI Overview appears. Where it does, expand the sources and record which domains are cited. Queries where an Overview appears and cites only competitors are your priority list. In a Pew Research Center study, 18% of Google searches in March 2025 produced an AI summary, so many queries won't show one.
Good result: your site is cited in AI Overviews for several target queries (or no Overviews appear).
Record: per query, whether an Overview appears, the cited domains and whether you are among them.
50. AI assistant visibility
Ask ChatGPT, Perplexity and Gemini three or four buying questions in your category, such as "best [service] for [need]" or "recommend a [service] provider in [region]". Then ask each one "What is [business name] and what does it do?" Record whether the business is recommended, whether the description is accurate, and which sources each assistant cites. The cited sources show where the recommendation is being decided.
Good result: recommended by two or more assistants and described accurately when named.
Record: per assistant and question, whether you appear, the description and the sources cited.
51. Answer-ready content structure
AI systems split pages into passages and quote the passages that answer a question best. For each key page, find the question its heading implies and check the complete answer appears in the first 100 words. Check each section still makes sense read on its own, headings say literally what their section covers, comparisons sit in tables and processes in numbered lists, and FAQ answers are complete on the page.
Good result: key pages answer their core question early, in sections that stand on their own.
Record: per page, where the direct answer sits and which sections depend on earlier text.
52. Citable specifics and original data
Compare the concrete, checkable facts on each key page (prices or price ranges, timelines, numbers, named methods, examples) with the generic claims ("tailored solutions", "proven results"). Test each generic claim: could a competitor put the same sentence on their site unchanged? If so, it gives an assistant nothing to cite. Note any original material only your business has, such as your own data or a documented method.
Good result: pages full of specific facts, with some material that exists nowhere else.
Record: per page, the number of specific facts and the generic claims worth replacing.
53. Entity markup (Organization and sameAs)
Entity markup tells machines who the business is. Check the homepage for Organization markup, or a more specific type, with name, url, logo, description and contact details. Check its sameAs property lists the business's real profiles elsewhere: LinkedIn, the Business Profile, directories, review platforms. sameAs is the field most often left empty, and it is what connects the site to its reputation on other sites.
Good result: complete organization markup with sameAs links to real profiles.
Record: the type used, the fields present and the full sameAs list.
54. AI crawler access
AI crawlers fall into three classes, and blocking the wrong class removes a site from AI answers. Check robots.txt for each one, including the wildcard rule that applies to any crawler without its own.
| Crawler class | What it does | Examples |
|---|---|---|
| Retrieval | Builds the index an assistant searches | OAI-SearchBot, Claude-SearchBot, PerplexityBot, Bingbot |
| User-triggered | Reads a page when a person asks the assistant to | ChatGPT-User, Claude-User, Perplexity-User |
| Training | Collects content for training models | GPTBot, ClaudeBot, Google-Extended, CCBot |
Retrieval and user-triggered access decide visibility. Blocking training crawlers is a separate choice with little effect on citations; Google-Extended, for example, doesn't affect Google Search. A firewall or CDN can also block AI crawlers whatever robots.txt says. You can note whether an llms.txt file exists, but don't count its absence against the site: Google's John Mueller called it "purely speculative" in June 2026.
Good result: retrieval and user-triggered crawlers allowed, and a deliberate choice about training crawlers.
Record: the rule for each crawler, whether each rule was intended, and whether llms.txt exists.
Frequently asked questions
How long does a website SEO audit take to do yourself?
A first pass through this checklist takes four to eight hours on a small site if you already know the tools, spread over a couple of days because crawls and speed tests involve waiting. The technical and on-page sections take longest. The AI search section takes under an hour and often produces the most surprising findings.
Can I do an SEO audit with free tools?
Yes. Every check in this checklist runs on free tools: Google Search Console, Google Analytics 4, PageSpeed Insights, Chrome DevTools, the Illucrum Bulk SEO Checker, free validators and the free tier of a backlink tool. Paid tools save time on large sites and give more backlink data, but they don't reveal a category of problem the free tools miss. The SEO audit vs DIY comparison covers when doing it yourself makes sense.
How often should a website be audited for SEO?
Audit a business website fully once a year, and after any major change: a redesign, a platform or CMS migration, a domain change or a hosting move. Between audits, a monthly look at the Search Console Pages and Core Web Vitals reports catches most technical problems as they appear.
What is the difference between this checklist and a technical SEO checklist?
The technical SEO checklist has 36 checks and covers crawling, rendering, indexing, speed, security and AI crawler access in more depth, and nothing else. This checklist covers the technical layer in 12 checks, then adds on-page SEO, keywords, backlinks, competitors, trust, local presence, conversion and AI visibility.
Should I fix everything the checklist finds?
No. Fix anything that stops search engines or AI crawlers from reaching or indexing the site first, then other serious problems, then quick fixes that take less than a day, then the rest. A finding on a page with no traffic and no commercial value can wait.
Fifty-four checks produce a document, and the document is the point: findings with evidence attached can be put in order and handed to whoever does the work. If you'd rather have them gathered, scored and prioritized for you, the Website SEO Audit covers the same 54 points at a fixed price listed on the pricing page.