How to Audit a Website for SEO: A Practical 2026 Guide

A useful SEO audit ends with a work queue, not a score. For each issue, record the affected URLs, evidence, proposed fix and how you will verify it worked.

This guide explains how to audit a website for SEO without turning it into a month-long project. Start with a focused review of your important pages, then expand wherever you find a recurring problem. A short audit can catch obvious blockers; it cannot prove that every URL on a large site is healthy.

What tools and pages do you need before you start?

Open Google Search Console, your analytics platform and PageSpeed Insights. Add a crawler such as Screaming Frog SEO Spider if you need to inspect titles, response codes and internal links across multiple pages. A browser alone is useful for spot checks, but not for finding every orphaned or broken URL.

Choose pages by business importance and template, not randomly. Include your homepage, a product or service page, a pricing or contact page, a leading article and a page whose organic traffic has declined. For a store, add a category and a product detail page.

Record each page’s intended search query and desired action. “Book a consultation” and “learn how payroll works” require different content and different measures of success.

Compare equivalent periods and account for seasonality. A drop during a quiet trading period does not automatically indicate a technical failure.

Can Google crawl your important pages?

Start with accessibility. A page cannot reliably earn search traffic if Google encounters errors or cannot discover its content.

Inspect the following in order:

1. Response codes: Important live pages should return HTTP 200. Investigate server errors and broken internal links. Keep genuine 404 or 410 responses for permanently removed content with no suitable replacement.
2. Redirects: Links should point directly to the final destination. Use permanent redirects for permanent moves, and avoid sending unrelated deleted pages to the homepage.
3. Robots rules: Check the robots.txt file for accidental blocking of public sections or resources needed to render pages.
4. Discovery: Make sure important pages have crawlable internal links and appear in a sitemap where appropriate.
5. Rendering: For JavaScript-heavy sites, inspect Google’s rendered page through URL Inspection. Check that primary copy and links exist in the rendered output.

HTTPS should work consistently, with alternate HTTP and hostname versions redirecting to your preferred version. Check for certificate problems and important resources still loaded over HTTP.

A robots.txt file is not mandatory for indexing. What matters is that any rules you publish do not block content you want crawled. Likewise, a sitemap helps discovery but does not guarantee indexing. Include preferred, indexable URLs rather than redirects, errors or pages deliberately marked noindex.

Are the right URLs eligible for indexing?

Crawling and indexing are different. Google may fetch a page without adding it to its index, and a robots-blocked URL can sometimes appear in search without its content being crawled.

Use Search Console’s URL Inspection for important pages. Check the indexed result first, then use the live test to investigate the current page. A successful live test does not guarantee indexing.

Look for accidental noindex directives in page metadata or HTTP headers. Do not block a page in robots.txt and expect Google to read its noindex directive: Google needs crawl access to see that instruction.

Inspect canonical declarations next. Each indexable page should have a clear preferred URL, usually expressed with a self-referencing canonical. Avoid conflicting canonical signals between HTML, headers, sitemaps and internal links. Canonicals are signals, not commands; Search Console can show when Google chooses a different URL.

Not every excluded URL needs fixing. Account pages, duplicate variants and retired content may be excluded correctly. Investigate the reason before removing safeguards.

Do titles, headings and snippets describe each page accurately?

Review the actual page, not just a spreadsheet warning. Every important indexable page needs a descriptive title that distinguishes it from other pages and matches its purpose.

For example, “Services” tells a searcher little. “Bookkeeping for Independent Restaurants” identifies both the service and the audience. Add a location only when the page genuinely serves that market.

There is no universal Google requirement that titles stay under 60 characters or descriptions under 160. Search display varies by device and available space, and Google may generate different title links or snippets. Treat length checks as editing prompts, not pass-or-fail ranking rules.

Write meta descriptions as accurate previews with a reason to click. They are not a direct ranking factor, and Google may use page text instead.

Use a clear primary heading and subordinate headings that explain the page’s structure. One obvious H1 is a useful convention, not a magical ranking requirement. Fix confusing hierarchy rather than chasing a tag count.

Also inspect image alt text and structured data. Alt text should describe meaningful images for people who cannot see them; decorative images generally need empty alt text. Where applicable, validate supported structured data with Google’s Rich Results Test. Passing validation does not guarantee a rich result.

Does the content satisfy the search behind the keyword?

Search the page’s main query and inspect the kinds of results Google shows. Are they tutorials, product categories, comparison pages or local businesses? This is a clue to intent, not permission to copy competitors.

A sales page targeting an informational question may struggle because it offers the wrong experience. Conversely, a long tutorial may not satisfy someone looking to buy a specific service.

Read the opening paragraph. It should establish the subject and help visitors know whether they are in the right place. Clear wording helps readers and search systems understand the page, but an unclear opening alone does not prove that Google cannot understand it.

Then check whether the page provides what a visitor needs to act: an actual process, product limitations, pricing context, examples, evidence or answers to common objections. Replace generic claims with details you can substantiate.

Assign each reviewed page a decision: keep, improve, merge or remove. Our content audit guide explains how to make that decision without deleting useful assets.

If two pages cover similar ground, inspect their queries before merging them. Overlap alone is not a problem. Merge when they serve the same purpose and a combined page would be better; preserve distinct pages when they answer different needs.

For health, finance or other high-stakes topics, check author qualifications, review processes and sources carefully. No amount of keyword editing compensates for unreliable advice.

Do internal links help readers and crawlers find useful pages?

Compare your crawl with your sitemap, Search Console exports and CMS page list. A crawler that follows links cannot discover a truly orphaned page by itself.

Find important URLs that receive few contextual links. Add links from relevant articles, category pages or service explanations using descriptive anchor text. “Restaurant bookkeeping services” explains a destination better than “click here.”

Repair broken links and update links that pass through redirects. Keep navigation usable on mobile, and avoid making important destinations accessible only through a site-search form or a JavaScript interaction without a crawlable link.

A useful test is to start at the homepage and try to reach a revenue-generating page. If the path is confusing for you, investigate the information architecture before adding more footer links.

Are speed and mobile usability getting in the way?

Use PageSpeed Insights on representative templates, not only the homepage. Its field data reflects real-user experience where available; its lab test is a controlled diagnostic run. Neither should be mistaken for a complete site assessment.

Google’s Core Web Vitals guidance defines “good” thresholds at the 75th percentile of page loads as follows:

If field data is unavailable, use lab diagnostics and manual testing, but do not claim that the page passes real-user Core Web Vitals. Our Core Web Vitals guide explains the distinction.

On a phone, test menus, forms, consent banners and calls to action. Check that mobile users receive the main content and that overlays do not prevent reading it. A perfect performance score is not required for rankings; prioritize delays and failures that affect actual use.

Which off-page and local signals should you check?

Review Search Console’s Links report for useful mentions and major changes, recognizing that it is not an exhaustive backlink database. Look for legitimate opportunities such as an unlinked partner mention or a broken link pointing to a moved resource.

Do not treat every unfamiliar backlink as harmful or disavow links just because an audit tool labels them “toxic.” Google advises that most sites do not need its disavow tool.

For eligible businesses serving customers in person, check the Google Business Profile: categories, hours, contact information, website destination and service details should be accurate. Ask customers for honest reviews without incentives or selectively requesting only positive feedback.

Relevant industry directories can help people discover a business, but there is no required number of listings. Use our local SEO playbook when local search is a meaningful acquisition channel.

What should you fix first, and how do you verify it?

Prioritize by business importance, scope and confidence. A template error affecting every product page normally matters more than an awkward description on an old announcement.

Give each task an owner and a specific completion test. “Improve SEO” is not a task; “remove the unintended noindex directive from published service pages” is.

SiteScoper’s free website audit can provide an initial scored review and ranked fixes without an account. Treat that as a starting point, not a replacement for Search Console evidence or a full crawl.

After shipping, verify technical changes immediately, then monitor indexing, query impressions, clicks and organic conversions over time. Annotate deployment dates and avoid attributing every movement to your changes. Revisit the audit after migrations, template changes or unexplained losses so the work queue stays tied to the site you actually run.