How to audit your site before you buy SEO tools

If you want to know how to do an SEO audit before buying software, start with evidence you can already access. In most cases, that means checking whether search engines can discover, crawl and index the pages that matter, whether those pages are genuinely useful for the searches they target, and whether people can move through the site without friction. You do not need a paid platform to answer those first questions well.

A good audit is not a hunt for as many issues as possible. It is a way to separate what you know from what you suspect. Some findings are direct evidence, such as a URL being excluded in Google Search Console. Some are interpretation, such as deciding whether that exclusion is a problem. Some are still only hypotheses, such as blaming a traffic drop on page speed before you have checked anything else. That distinction matters if you are deciding whether to buy tools, brief a developer, or ask for outside help.

What an SEO audit should tell you

An SEO audit should tell you three things.

First, whether your important pages are technically available to search engines and users. That includes discovery, crawling, indexing, status codes, canonicals, redirects, and whether the page can actually be used.

Second, whether those pages deserve to rank for the searches you care about. A page can be indexable and still fail because it answers the wrong question, repeats other pages, or gives visitors very little reason to trust it.

Third, what to do next, in order. An audit that ends with a long spreadsheet of warnings is not very useful if it does not distinguish urgent faults from minor housekeeping.

This is where it helps to separate evidence, interpretation and guesswork.

Evidence is something you can verify directly. Examples include a blocked URL in robots.txt, a page returning 404, or a URL shown as “Crawled - currently not indexed” in Google Search Console. Interpretation is your reading of that evidence. A “Discovered - currently not indexed” URL might be harmless if it is a low value filter page, or important if it is your main service page. Guesswork is when we jump beyond the evidence, for example deciding that Googlebot dislikes your whole site because one section is under-indexed.

When we explain how to do an SEO audit on The SEO Blog, we are really teaching how to assess evidence before buying a dashboard that may simply repackage what is already visible. If you want a companion checklist for avoiding expensive false alarms, read how to audit a site without buying the wrong fix.

Begin with indexing and crawl evidence

Start in Google Search Console, because it shows what Google has actually done, not what a third party crawler guesses might happen. Use the Page indexing report and inspect individual URLs that matter most, such as core service pages, key category pages, lead pages and recent articles. Google explains that excluded URLs need investigating by reason, and that not every excluded page should be indexed (Google Page indexing report).

Check three basic things.

One, are your important URLs known at all. If a page does not appear in Search Console and is not in your XML sitemap, that is a discovery problem. It may be orphaned, buried in navigation, or only reachable through an internal search or form.

Two, has Googlebot crawled the page. In the URL Inspection tool, look for the last crawl date and whether crawling is allowed. If the page is blocked by robots.txt, marked noindex, or redirected elsewhere, that is direct evidence of why it is not appearing.

Three, is the page indexed as the correct canonical URL. If Google has chosen a different canonical from the one you declared, inspect both versions. This often happens with duplicate pages, parameter variants, HTTP and HTTPS leftovers, or pages that differ too little to justify separate indexing.

Treat exclusions carefully. “Alternate page with proper canonical tag” is often fine. “Page with redirect” is expected if the redirect is intentional. “Not found (404)” may be normal for deleted pages, but not if internal links still point there. “Soft 404” can mean the page is thin, misleading, or functionally empty. “Crawled - currently not indexed” does not prove a penalty or technical fault. It tells you Google reached the page and did not index it at that point. That may reflect duplication, weak demand, weak content, or site quality concerns, but Search Console does not tell you which one.

Also check your XML sitemap. It should include canonical, indexable URLs that you actually want in search. If your sitemap is full of redirected URLs, tag pages, parameter pages or old content, it becomes less useful as a signal of priority.

If your business also cares about AI search visibility, keep the evidence types separate. A crawler visit from OAI-SearchBot or GPTBot is not the same thing as appearing in search results or being cited in an answer. OpenAI documents these crawler controls separately, and training controls are not the same as search access (OpenAI crawler documentation). Likewise, Google Search access is separate from Google-Extended, and Google does not require a special AI text file or special schema for its Search AI features (Google AI features).

For a wider explanation of what Search Console can and cannot prove, see what Search Console evidence can really tell you.

Check whether your important pages deserve to rank

Once you know the right pages can be crawled and indexed, ask a harder question. Do they deserve to rank?

Start with page purpose. Every important page should have a clear job. A service page should explain the service, who it is for, what problem it solves, and what the next step is. A product category should help the visitor compare options. An article should answer a real question better than a generic summary would.

Then check search intent. Search the terms you think the page should appear for and look at the current results. Are they guides, product pages, category pages, local business pages, or comparison pieces? If your page format does not match the broad intent shown in results, better title tags alone will not fix that.

Look for duplication in both content and purpose. Two pages do not have to repeat paragraphs word for word to compete with each other. If you have three pages all trying to rank for the same service in the same area, with only minor wording changes, Google may choose one and ignore the rest. In UK local search, that can matter because local visibility is not decided by on-page text alone. Google says relevance, distance and prominence all contribute to local ranking (Google local ranking). If one page is meant for local intent, make sure it genuinely adds local relevance instead of cloning another page.

Review headings, titles and links for clarity rather than keyword stuffing. Descriptive links help users and search engines understand what they will reach, and useful, accessible content remains the core recommendation in Google’s own starter guidance (Google SEO Starter Guide). “Read more” repeated twenty times is weak. “Compare our bookkeeping packages” or “See delivery areas in Suffolk” tells both people and crawlers much more.

Ask whether the page is genuinely useful to the reader. Can someone make progress after landing on it, or do they still need to leave and search again? Signs of weak utility include vague claims, no examples, no explanation of scope, missing prices where buyers expect them, no contact route, no evidence of expertise, and a layout that hides the main answer beneath banners or filler.

This is also the stage to check whether the page has been written for a tool rather than for a reader. If the copy sounds like it exists only to mention a phrase, it usually underperforms in practice. A sound audit often concludes that the main issue is not software, but weak page purpose.

Review internal links, navigation and accessibility basics

A site can have good pages and still underperform if users and crawlers cannot move through it clearly.

Start with navigation. Can a visitor reach your key commercial pages from the main menu or from obvious section pages? Are important pages buried four or five clicks deep for no good reason? If a page matters to the business, it should not depend on a footer link alone.

Then review internal links in context. Category pages should link down to subcategories and important detail pages. Articles should link to genuinely relevant next steps. Service pages should link to related services, case examples, FAQs or contact routes where those help the visitor. Anchor text should describe the destination naturally. This is one of the simplest checks in any process for learning how to do an SEO audit properly, because poor internal linking is common and often fixable without new tools.

Look for orphaned pages, pages with no internal links pointing to them from crawlable HTML links. XML sitemap inclusion is not a substitute for a normal site path. If users would struggle to discover a page by browsing, crawlers may treat it as lower priority too.

Accessibility basics matter here because they overlap with usability. The W3C’s design tips are practical and worth using as a checklist for clear controls, readable contrast, understandable navigation and sensible page structure (W3C accessibility design tips). You do not need to become an accessibility specialist to spot obvious problems. Check whether buttons look clickable, whether link text makes sense out of context, whether headings are used in a logical order, and whether mobile users can operate menus without frustration.

Also pay attention to page design choices that interrupt the task. Intrusive pop-ups, hard-to-close banners, carousels that push content down, and forms that ask too much too soon can all reduce the usefulness of a page even when the text is good.

If you run an ecommerce site, the same principles apply with more moving parts. Faceted navigation, stock pages, product variants and filter URLs create extra crawl and duplication questions. We cover that separately in how to audit an online shop without missing the basics.

Test performance without turning scores into promises

Performance should be tested, but not mythologised.

Use Core Web Vitals and real page behaviour as evidence, not as a promise of rankings or revenue. Core Web Vitals describe loading, responsiveness and visual stability, and poor results can indicate a frustrating experience, especially on mobile connections (web.dev on Core Web Vitals). They do not prove why a page is not ranking, and a perfect score does not guarantee better search visibility.

In practice, check whether key templates behave badly. Do pages jump as images or banners load. Does the main content appear quickly. Can users interact promptly with menus, forms and filters. Are there obvious delays caused by heavy scripts, video embeds, cookie tools or third party widgets.

Look at field data where available, because lab tests and live user experience are not the same. Search Console’s Core Web Vitals report can show patterns across groups of pages. If one template type is consistently poor, that is more useful than obsessing over a single homepage score.

Interpret results cautiously. Poor Largest Contentful Paint may point to oversized images, server delays, render-blocking resources or bulky themes. Layout shift may come from missing image dimensions, injected banners or late-loading fonts. Interaction problems may be tied to JavaScript execution. But until a developer checks the implementation, these are still likely causes, not confirmed diagnoses.

This is also where expensive tools can distract people. A page can fail a synthetic score and still be commercially effective. Another can pass and still confuse visitors. Use performance testing to identify behaviour worth fixing, not to chase vanity numbers.

Turn the audit into a practical action list

A useful audit ends with decisions.

Group findings by severity, confidence and effort.

Severity asks, if this is true, how much does it matter. A noindex on your main service page is high severity. Missing alt text on a decorative image is usually lower severity in SEO terms, though it may still matter for accessibility.

Confidence asks how sure you are. A 404 shown in your browser and in Search Console is high confidence. A suspicion that a page feels “thin” is lower confidence until you compare it with the search intent and competing results.

Effort asks what it will take to fix. Updating internal links may be low effort. Reworking a duplicated site structure or replacing a bloated theme may be high effort.

A simple action list often works better than a long audit deck:

• Fix now: issues with high severity and high confidence
• Investigate next: high severity but lower confidence issues
• Improve when practical: medium value items with clear fixes
• Monitor: patterns that need more data before action

For example:

• Main service page excluded by noindex, fix now.
• Important location page indexed under the wrong canonical, investigate and correct.
• Blog posts with weak descriptive links, improve when practical.
• Category templates showing poor Core Web Vitals across the board, investigate with development input.
• Several low value tag pages excluded from indexing, monitor or ignore if intentional.

This is the point at which you can decide whether you actually need paid tools. If your audit shows straightforward indexing, content and internal linking issues, more software may not be the bottleneck. If your site is large, has complex templates, multiple markets, or serious log analysis questions, then specialist tools may be justified because they save time and expose patterns you cannot easily see manually.

Before spending money, ask what decision the tool will help you make that you cannot already make with your current evidence. That one question prevents a lot of waste.

If you want another practical route through the same decision, read a practical site audit you can run before buying tools. We also publish wider guidance on AI search and visibility where crawler access, citations and search appearance need to be kept separate.

The main lesson is simple. Learn to inspect your own evidence first. When you know what your pages are for, whether they are indexed, and whether people can actually use them, you are in a much stronger position to judge any tool, any audit, and any recommendation you are offered.