50+ SEO Checks for Auditing a Business Website in 2026

The first question is not whether the pages are perfectly optimized. It is whether Google can reach them, understand which versions are intended, and include the right pages in its index.

This is where many audits become unnecessarily confusing. A crawler may find a URL because it follows links inside the website, while Google may exclude that URL because of an index instruction, a canonical conflict, a rendering problem, or a server response. Therefore, your crawler’s URL count and Google’s indexed-page count are not expected to tell the same story.

Google Search Console is the starting point because its Page Indexing report shows the URLs Google knows about, while the URL Inspection tool can show information about an individual indexed URL, test its live version, and display the rendered page Google receives.

• 1. Review robots.txt for unintended blocks. Confirm that important page paths, scripts, stylesheets, APIs, and resources are not being blocked. It is worth remembering that robots.txt controls crawling rather than guaranteed removal from the index.

• 2. Check for accidental noindex instructions. Inspect both HTML robots meta tags and X-Robots-Tag response headers. A blocked URL may remain indexed because Google cannot crawl it to see the noindex instruction, which is why Google treats blocking crawling and blocking indexing as different controls.

• 3. Compare expected pages with Search Console’s indexing data. Group exclusions by reason and template. Five obsolete filter URLs marked “Crawled, currently not indexed” are not equivalent to the same status appearing across the main service pages.

• 4. Audit the XML sitemap as a statement of preference. It should contain canonical, indexable URLs that return successful responses and deserve discovery. Redirected, duplicate, broken, parameter-based, or intentionally excluded URLs do not belong there. Google also makes clear that submitting a sitemap is a hint, not a guarantee.

• 5. Verify HTTP response codes. Important pages should not return soft 404s, server errors, redirect responses, or misleading 200 OK responses for content that is actually missing.

• 6. Find redirect chains, loops, and irrelevant destinations. A permanently moved page should normally reach its closest replacement through a direct server-side redirect. Google recommends permanent redirects such as 301 or 308 when a page has genuinely moved in its redirect guidance.

• 7. Compare declared and Google-selected canonicals. Check whether Google is indexing the version the business intended. Canonical tags, sitemap entries, internal links, redirects, and hreflang references should support the same choice rather than contradicting one another.

• 8. Audit old URLs after redesigns, migrations, and platform changes. Valuable pages should redirect to their closest genuine replacements. Sending every retired URL to the homepage rarely preserves the original meaning or gives the visitor a useful destination.

• 9. Consolidate protocol and hostname variants. HTTP, HTTPS, www, and non-www versions should resolve consistently to the preferred site, while canonical tags, internal links, and sitemap URLs should reinforce that decision.

• 10. Check whether the hosting environment blocks or delays search crawlers. A URL can be allowed in robots.txt and still fail when a CDN, firewall, security challenge, timeout, or overloaded server interferes with the request.

For example, imagine that a service template adds noindex whenever its optional testimonial field is empty. A tool may report eleven excluded pages, but that number alone does not explain the problem. The useful finding is that three of those pages previously generated enquiries, all eleven share the same template condition, and the repaired output must be confirmed in the rendered HTML before recrawling is requested.

This is also why indexing controls should not be changed in bulk simply because an audit platform produced a large warning count. First confirm the affected template, the intended page state, and what the business could lose if the decision is wrong.

Once the right pages can be found and indexed, the next question follows naturally: does the website itself treat those pages as important?

Being indexable does not mean being easy to discover. A page may exist in the sitemap and still receive almost no internal support, sit several clicks away from the homepage, or remain disconnected from the content that proves its relevance.

Internal links matter because they help search engines find pages and understand their relationships. They also guide customers from a question toward evidence, an offer, and a decision. Therefore, site structure should not be judged only by whether every URL can technically be reached.

• 11. Measure click depth to commercially important pages. If a major service, category, or product requires four or five decisions from the homepage, the site is quietly treating it like a minor page.

• 12. Identify orphan pages. Find indexable pages with no crawlable internal links pointing to them. A sitemap entry may expose the URL, but it does not make the page part of a meaningful customer or topic journey.

• 13. Evaluate the navigation and footer by business priority. Check whether they promote the pages that explain, sell, and convert, rather than giving most of their space to policy pages, old archives, and categories with little value.

• 14. Repair broken and redirected internal links. Internal links should normally point directly to the current destination. Repeated internal redirects slow navigation, waste crawl activity, and make later migrations harder to understand.

• 15. Check breadcrumb accuracy. Breadcrumbs should reflect the real hierarchy and give users a useful path upward. They should not invent relationships simply to place more keywords on the page.

• 16. Find low-value archives absorbing internal authority. Tag pages, filtered URLs, dated archives, and duplicate category systems can collect thousands of links while important commercial pages receive very few.

• 17. Connect related pages through a deliberate topic path. A service page should lead naturally to relevant evidence, examples, FAQs, supporting guides, and the next commercial step. Internal linking should help the visitor continue a decision rather than merely circulate authority.

For example, a business may publish a detailed guide to industrial roof replacement while leaving it disconnected from the commercial roofing service page. The guide demonstrates expertise, and the service page presents the offer, yet neither helps the reader move from one to the other.

The answer is not to add twenty keyword-rich links. It is to create a clear path between the customer’s question, the evidence that answers it, and the service that resolves the problem.

However, even the clearest path will fail if the page at the end of it loads badly, hides its content, or breaks when the customer tries to use it. That brings the audit from structure to performance.

Performance auditing is often reduced to a single score. This is attractive because a score feels objective, but it can distract from the experience that real visitors and search crawlers actually receive.

A page may score well in a laboratory test while its price, product details, or form fail to render. Another page may miss a Core Web Vitals target while still being the strongest answer in its market. Therefore, the aim is not to pursue a perfect number. It is to identify technical conditions that hide content, delay important interactions, or cause customers to leave.

Google’s current “good” Core Web Vitals targets are an LCP of 2.5 seconds or less, an INP below 200 milliseconds, and a CLS below 0.1. These measurements are useful, but they still need to be interpreted by page type, traffic, and business function.

• 18. Test the templates that receive search traffic and generate revenue. Include service, product, category, article, location, lead form, and checkout templates rather than testing only the homepage.

• 19. Review Core Web Vitals as template-level patterns. Determine which shared component, asset, or implementation causes poor LCP, INP, or CLS instead of optimizing isolated URLs without understanding the pattern.

• 20. Compare visible content with Google’s rendered HTML. JavaScript interfaces can show customers a complete page while Google receives missing headings, links, prices, or article copy. Google recommends checking the rendered output through URL Inspection or the Rich Results Test in its JavaScript SEO guidance.

• 21. Test the complete journey on real phones. Navigation, accordions, forms, consent banners, filters, and calls to action must work without overlap or missing content. This matters even more because Google predominantly uses the mobile version of a site’s content for indexing and ranking.

• 22. Check interaction delay and layout movement. A page may appear quickly while buttons hesitate, filters freeze, or content shifts beneath the user’s finger. Test the actions people take after the initial load.

• 23. Find oversized or poorly delivered images and video. Review dimensions, compression, formats, responsive delivery, lazy loading, and whether the hero asset delays the largest visible content.

• 24. Audit third-party scripts. Chat tools, advertising pixels, heatmaps, testing platforms, embeds, and trackers may be unused, duplicated, or disproportionately expensive.

• 25. Check server reliability under realistic conditions. Review time to first byte, intermittent errors, overloaded hosting, CDN behavior, and slower connections. A recurring hosting problem can resemble an unpredictable SEO decline.

Consider a product page that receives a respectable speed score while its price, availability, reviews, and purchase button appear only after a JavaScript request that fails for crawlers. Removing another 100 milliseconds from the hero image would improve the score without solving the real problem.

This means the audit has to show which content is missing, why it fails, which templates share the implementation, and how the corrected version will be tested.

Once the page can be found, reached, rendered, and used, it still has to make a persuasive case. Technical health creates access, but it does not create relevance or demand.

A technically clean page can still fail because it answers the wrong question, repeats what every competitor says, or gives the visitor no credible reason to act.

This is where simplistic audits often turn content quality into a word count. Yet a short page can answer a narrow question completely, while a long page can avoid the decision the customer came to make. Therefore, content should be judged by purpose, evidence, usefulness, and the business outcome it supports.

Google’s people-first content guidance emphasizes original, useful, reliable information created for readers rather than pages assembled mainly to attract search traffic.

• 26. Review titles and main headings for clarity and differentiation. Each important page should communicate its subject and purpose without relying on boilerplate. Also compare the declared title with what Google actually displays in search results.

• 27. Match the page to the searcher’s real task. Someone looking for pricing, eligibility, compatibility, a local provider, or a troubleshooting answer needs a different page experience. Containing the phrase is not the same as satisfying the intention.

• 28. Identify thin, duplicated, and templated pages. Determine whether each page contributes unique information, evidence, products, locations, expertise, or decisions. Do not diagnose quality from word count alone.

• 29. Find pages competing for the same purpose. Compare queries, ranking URLs, headings, internal links, backlinks, and conversions before combining anything. Similar pages may be competing, or they may legitimately serve different audiences and stages.

• 30. Check factual and commercial freshness. Review prices, product availability, staff details, service areas, screenshots, statistics, dates, legal requirements, and recommendations that may have changed.

• 31. Review images and alternative text by function. Product images, charts, diagrams, and examples need descriptions that convey their meaning, while decorative images generally do not.

• 32. Require evidence for specific claims. Statistics should identify their source. Results should explain their conditions. Expert advice should identify the author or reviewer. Clear authorship helps readers assess accountability, but it should not be sold as a guaranteed ranking switch.

• 33. Audit the path from the visit to the business outcome. The offer, proof, next step, form, phone number, product action, or booking process should be visible and credible. Rankings do not compensate for a page that leaves the customer unsure what to do.

Suppose three pages rank intermittently for “commercial SEO audit.” One explains the service, one is a short blog post, and one is an old campaign landing page. The easy answer is to call this cannibalization and delete two pages.

However, that action would be premature. First determine which page converts, which one has backlinks, which queries each page serves, and whether the old page contains evidence worth preserving. The correct response may be consolidation, differentiation, a change in internal links, a redirect, or no change at all.

In other words, “low traffic,” “thin,” and “duplicate” are descriptions, not deletion instructions.

Once the visible page communicates clearly to people, its machine-readable description should support the same meaning rather than inventing a different one. This is where structured data enters the audit.

Structured data gives search engines explicit information about a page and can make supported content eligible for richer search appearances. Because it is machine-readable, it can look more powerful than it really is.

The temptation is to add as many schema types, service areas, questions, reviews, and entities as possible. Yet markup does not create a business fact. It can only describe one that genuinely exists and is supported by the visible page.

Google says structured data should remain visible, relevant, current, and non-misleading in its general structured data guidelines.

• 34. Confirm that organization and business details match the visible site. Names, URLs, logos, contact details, addresses, and external profile references should describe the same real entity.

• 35. Use schema types that fit the page rather than the keyword target. Do not label ordinary marketing copy as an FAQ, a service as a product with invented ratings, or a virtual office as a staffed branch.

• 36. Connect article authorship to a real author entity. The visible byline, author page, Article or BlogPosting markup, and genuine external profiles should refer to the same person without hiding or inventing credentials.

• 37. Keep FAQ and other structured content equivalent to what readers can access. The questions, answers, products, reviews, prices, and offers in JSON-LD must agree with the visible page.

• 38. Validate syntax, eligibility, and duplication. Use the Rich Results Test, Schema Markup Validator, Search Console reports, and source inspection. Valid JSON-LD can still use the wrong type, duplicate markup from a theme or plugin, or describe something unsupported.

It is also important not to turn structured data into a separate theory of AI visibility. For Google’s AI Overviews and AI Mode, there is no special schema type or separate technical admission process. Google’s current guidance for generative AI features says that foundational SEO practices remain relevant and that special AI markup or llms.txt is not required for Google Search.

This does not mean AI visibility is automatic. It means crawlable, indexable, useful content supported by accurate information remains the sensible starting point.

However, a business is not understood only through what it says about itself. Other websites also describe, cite, recommend, criticize, and associate with it. Therefore, the audit must now move beyond the website.

A backlink is not simply a number inside an SEO platform. It usually appears inside a page that says something about the business, its expertise, its products, or its reputation.

This means the audit should not stop at link totals or third-party authority scores. It needs to examine who created the reference, why it exists, what surrounds it, where it sends the reader, and whether the association is one the business would willingly approve.

Google uses links for discovery and as one of many relevance signals, while its spam policies prohibit practices intended to manipulate rankings. Therefore, source, context, purpose, and risk matter more than a large raw count.

• 39. Review the quality and relevance of referring sites. Determine whether they represent real organizations, audiences, editorial decisions, partnerships, directories, user contributions, or websites created mainly to sell links.

• 40. Compare link support for commercial and informational pages. An old article may attract many references while an important service or category page remains unsupported. The response may involve better internal linking, stronger evidence, digital PR, or a supporting resource worth citing.

• 41. Recover links lost through moved or deleted URLs. After redesigns and migrations, find externally linked pages that now return errors or redirect to irrelevant destinations. Restore the resource or redirect the URL to its closest legitimate replacement.

• 42. Find accurate unlinked brand mentions. A genuine article, association page, partner profile, or news story that names the business without linking may present a reasonable outreach opportunity.

• 43. Investigate sudden or unusual backlink patterns. Examine timing, source type, anchor text, target pages, and context before assuming an attack or uploading a disavow file. A large number alone does not establish harm.

• 44. Audit what each important placement communicates. Check the surrounding claim, destination, authorship, disclosure, and neighboring businesses. A measurable backlink can still misrepresent the brand or create an association the company would reject.

For example, a redesign may retire a research page that earned links from trade associations and universities. The old URL then redirects to the homepage. The links still appear in backlink tools, but visitors reach a generic page that does not answer the original reference.

A useful audit identifies the lost topic, verifies the strongest links, recommends restoring or replacing the resource, maps the redirect correctly, and explains how recovery will be monitored.

For a local business, these external signals do more than support the website. They also help establish where the business operates, how customers experience it, and whether the public information is consistent. That makes local visibility the next part of the same investigation.

Local SEO is sometimes treated as a separate collection of citations, city pages, and map settings. In reality, it is the point where the website, the Google Business Profile, customer reviews, local references, and the physical or service-area operation are expected to describe the same business.

Google says local results are mainly influenced by relevance, distance, and prominence. Complete and accurate profile information helps Google understand the business and helps customers decide whether it fits their needs. Its local ranking guidance also makes clear that a business cannot simply pay Google for a better local position.

• 45. Review the Google Business Profile for accuracy and completeness. Check the name, categories, address or service-area setup, phone number, hours, website URL, appointment link, services, attributes, photos, and verification state.

• 46. Evaluate location and service-area pages for real local value. Each page should contain information that genuinely changes with the location, such as availability, staff, projects, regulations, neighborhoods, logistics, proof, or customer questions. Replacing one city name inside repeated copy is not a local strategy.

• 47. Check business information across important sources. Resolve material differences in the name, address, phone number, website, hours, and location status across the site, Google Business Profile, major directories, industry platforms, and authoritative local sources.

• 48. Review reputation and local prominence. Compare review volume, recency, topics, responses, ratings, local links, community references, and competitor positioning. Do not reduce review acquisition to a numbers race or use prohibited incentives.

Consider a company that publishes thirty city pages while operating from one verified location and showing no location-specific evidence. The pages may look optimized, but customers and search engines receive little reason to trust the claimed relevance.

The better approach is to document where the business genuinely serves, explain how the service is delivered there, connect each page to real evidence, and avoid pretending that schema or repeated city names create a physical presence.

At this stage, the audit has examined whether the business can be found, understood, trusted, and used. Yet none of those findings can be prioritized properly if the measurement system does not show what produces traffic, enquiries, or revenue.

Measurement is often left until the end of an audit, but it influences every conclusion that comes before it. Without reliable data, a valuable page may be removed because it appears inactive, a tracking failure may be mistaken for a traffic decline, or branded demand may be presented as new organic growth.

Search Console and Google Analytics measure different parts of the journey, so their numbers should not be expected to match exactly. Google recommends using the two together to connect search discovery with onsite behavior and conversions in its Search Console and Analytics guidance.

• 49. Confirm access, ownership, and configuration. Verify the correct Search Console properties, Analytics property and data stream, consent behavior, filters, internal traffic treatment, time zone, and connections with advertising or reporting systems.

• 50. Analyze performance by query, page, country, device, and search appearance. One site-wide traffic graph can hide the difference between a declining commercial page and a growing article that produces no enquiries.

• 51. Track outcomes that matter to the business. Test forms, qualified calls, bookings, purchases, revenue, downloads, and other meaningful actions. Successful actions should fire once, while failed attempts should not count as conversions.

• 52. Separate branded discovery from non-branded demand. Growth from people already searching for the company name is valuable, but it does not prove that the website is reaching more people who did not know the business.

• 53. Annotate changes that can explain later movement. Record redesigns, migrations, releases, major content changes, tracking changes, promotions, outages, manual actions, and significant campaigns.

• 54. Turn findings into an owned action and verification plan. Each accepted recommendation needs a responsible person, implementation sequence, risk note, expected result, validation method, and review date.

For example, traffic may fall by 28 percent after a site release, making an algorithm update appear responsible. A segmented review could then show that rankings and Search Console clicks remained stable while Analytics sessions declined because the new consent configuration stopped recording a large share of visitors.

That is not an SEO recovery project. It is a measurement repair.

In other words, the audit is only complete when it can distinguish a search problem from a website problem, a conversion problem, a reputation problem, or a reporting problem.