Snoika
Technical SEO·How-to guide

How to run a technical SEO audit: a step-by-step process for teams

A technical SEO audit checks whether search engines can reach the pages you want found and render them for indexing. It also checks whether those pages load well for real users. You start with a full crawl and end with a fix list ranked by impact. A site under 500 URLs can be audited in a day with free tools. Larger sites need a paid crawler and about a week of analysis.

Portrait of Artem LozinskyArtem LozinskyOperations and product, Snoika · 9 min read
One person of mixed ages in a glass-walled meeting room with a whiteboard.

Run a technical SEO audit

Run the audit in a fixed order, because each step uses data from the step before it. The crawl gives you the list of URLs. Index coverage tells you which of those URLs Google kept. Rendering and speed checks explain why Google dropped or ranked down the pages that matter. Follow this sequence:

  1. Crawl the site and export every URL with its status code. Include each URL's canonical and indexability in the export.
  2. Compare the crawl with Google's index in Search Console.
  3. Check that JavaScript content shows up in the HTML Google renders.
  4. Measure Core Web Vitals from real-user data.
  5. Validate structured data on your key templates.
  6. Find the URLs that waste crawler visits.
  7. Write the findings into one report.
  8. Rank the fixes by impact and hand them to owners.

Before you start, gather these inputs:

  • Owner or full-user access to Google Search Console
  • A desktop or cloud crawler
  • Your XML sitemap URLs
  • Server log files covering at least 30 days, if your host provides them
  • A shared tracker such as a spreadsheet or Jira board
  • One developer contact who can confirm how the site is built

Choose your audit tools

Pick your crawler by site size and by how much JavaScript the site uses. A small brochure site can be audited with free tools. A JavaScript-heavy shop with 50,000 URLs needs a paid crawler that renders pages. Prices below are the 2025 to 2026 list prices.

Google Search Console

Price
Free
Crawl limit
Your verified property
JavaScript rendering
Shows Google's rendered HTML per URL
Suits
Every audit, as the source of index data

Screaming Frog SEO Spider (free)

Price
£0
Crawl limit
500 URLs per crawl
JavaScript rendering
Not included
Suits
Small sites and one-off checks

Screaming Frog SEO Spider (paid)

Price
£199 / $279 per user per year
Crawl limit
Limited by memory and storage
JavaScript rendering
Included
Suits
Freelancers and in-house teams on desktop

Sitebulb Lite / Pro

Price
£125 or £305 per user per year
Crawl limit
Desktop, no fixed cap
JavaScript rendering
Included
Suits
Teams that want prioritized hints and client-ready reports

Sitebulb Cloud

Price
From £95 per month
Crawl limit
Built for large sites
JavaScript rendering
Included
Suits
Agencies sharing audits across a team

Ahrefs Webmaster Tools Site Audit

Price
Free for verified sites
Crawl limit
Your verified sites
JavaScript rendering
Included
Suits
Owners who already use Ahrefs data

Screaming Frog's paid licence costs £199 per year per user, with lower per-seat prices for 5 or more licences. Sitebulb has no permanent free tier, only a 14-day trial, and its desktop Lite plan costs £125 a year according to a 2026 comparison by Indexed. Ahrefs offers its Site Audit free for sites you verify in Ahrefs Webmaster Tools. Teams that write their own fix tickets get the most out of Screaming Frog. Sitebulb's built-in prioritization saves hours when the report goes to a client or a manager.

Crawl the site first

Crawl the whole site before you open any other report, since every later step checks URLs against this list. Set the crawler's user agent to Googlebot Smartphone, because Google indexes the mobile version of your pages. Switch on JavaScript rendering if the site runs on React, Vue or a similar framework. Then add your sitemap URLs to the crawl so the tool can flag orphan pages that no internal link reaches.

Once the crawl finishes, export the URLs and sort them by status code. Pages returning 4xx or 5xx errors and redirect chains come first. Pages with canonicals pointing elsewhere share that priority. Ahrefs found a 3XX redirect on 95.2% of sites in its study of over 1 million domains. A single redirect is harmless. Chains of three or more hops and internal links that point to redirected URLs are worth fixing.

Check the crawl worked by comparing the URL counts from the crawl and the sitemap. For the third of the three counts, use the number of pages Google reports as indexed. If the crawl found far fewer URLs than the sitemap lists, the crawler was blocked or the internal links are broken.

Check index coverage

Check index coverage by comparing what you want indexed with what Google has indexed, then explain every gap. Search Console holds the data. The crawl from the previous step tells you which pages should be in the index.

Read the Page indexing report

Open Indexing > Pages in Search Console and look at the "Why pages aren't indexed" table. Each row is a reason, such as "Excluded by 'noindex' tag" or "Crawled - currently not indexed". Click a reason to see the affected URLs. Then export them. Redirected URLs and deliberate noindex pages belong in this table.

Inspect individual URLs

Use the URL Inspection tool on 10 to 20 of your most valuable pages. It tells you whether the URL is on Google and which canonical Google picked. Pay attention when Google's chosen canonical doesn't match the one you declared. That mismatch means Google read your duplicate signals differently from how you meant them.

Compare indexed pages with your sitemap

Submit each sitemap under Indexing > Sitemaps and filter the Page indexing report by sitemap. Pages that sit in your sitemap but aren't indexed are your priority list. Pay attention to "Discovered - currently not indexed", which points to crawl capacity or weak internal links. "Crawled - currently not indexed" points to thin or duplicate content.

Audit rendering issues

Audit rendering by comparing the raw HTML your server sends with the HTML Google sees after it runs JavaScript. If the main content or the internal links only appear after rendering, Google has to queue the page for a second pass before it can read them. Martin Splitt, a Developer Advocate at Google, put the risk plainly: "Even though Googlebot can render JavaScript, we don't want to rely on that."

Run the comparison in two places. In your crawler, export the original HTML and the rendered HTML for a sample of templates and compare word counts and link counts. In Search Console, URL Inspection > View crawled page shows the HTML Google actually rendered.

Look for these failures, which come up again and again:

  • Links built as buttons or `onclick` events without an `<a href>` tag
  • Product text or reviews loaded from an API after the page appears
  • Title, canonical or meta robots tags that JavaScript changes after load
  • Scripts or CSS files that robots.txt blocks

The check passes when the body text and internal links match in both versions, provided the head tags match too.

Measure page experience signals

Measure page experience with field data from real Chrome users, then use lab tools to find the cause. Google's web.dev team sets LCP within 2.5 seconds as the "good" threshold for one of the three Core Web Vitals. The other two thresholds are Interaction to Next Paint (INP) of 200 milliseconds or less and Cumulative Layout Shift (CLS) of 0.1 or less. Largest Contentful Paint (LCP) measures when the main content appears. A page passes only when it meets all three at the 75th percentile of visits, separately for mobile and desktop.

Start with the Core Web Vitals report in Search Console. It groups URLs by template, so one slow product template shows up as one problem instead of 4,000. Then open PageSpeed Insights for a sample URL from each failing group. Its Lighthouse section names the element that delays LCP or the script that blocks interactions.

Expect the results to vary by device. Mobile scores worse, and INP failures come from third-party scripts such as chat widgets and tag managers. Field data covers a rolling 28-day window, so a fix you ship today won't show as passed for about 4 weeks.

Validate structured data

Validate structured data by testing one URL per template, because markup is generated by templates and an error on one product page is an error on all of them. Paste each URL into Google's Rich Results Test to see which rich results the page qualifies for. Then check the Schema Markup Validator for errors that Google's tool ignores.

Check that the markup describes what's visible on the page. A price in the schema that differs from the price on screen breaks Google's guidelines and can cost you the rich result. Also check that required properties are present. Product markup without `offers` or reviews, for example, loses eligibility.

Audit for types Google no longer rewards, too. In August 2023, Google said FAQ rich results would only show for authoritative government and health sites and dropped HowTo rich results on desktop. Search Engine Journal reported that FAQ rich results stopped appearing on May 7 2026 for all sites. The markup does no harm, but it's no longer a reason to spend developer time.

Fix crawl budget waste

Fix crawl budget waste by finding the URLs Googlebot visits that you never want indexed, then closing them off. Google's own guide says crawl budget is mainly a concern for sites with 1 million+ unique pages that change weekly, or 10,000+ pages that change daily. A 300-page site stays below the limit. But the same waste on a smaller site still slows how fast new pages get crawled.

Find the waste in two sources. The Crawl stats report in Search Console (Settings > Crawl stats) shows requests by response code and file type. Server logs show the exact URLs Googlebot requested. Filter the logs by Googlebot's user agent and group the URLs by pattern. Parameters from faceted navigation and internal search results cause crawl waste. Session IDs and endless calendar pages cause it too.

Block parameter patterns that create no unique content in robots.txt. Return 404 or 410 for pages that are gone instead of redirecting them all to the homepage. Remove redirected URLs from the sitemap. The check passes when a growing share of Googlebot's requests in the logs hit your indexable pages.

Build your audit report

A technical SEO audit report should document the affected URLs and evidence for every issue. Each entry should explain the business impact and assign an owner to the proposed fix. The people who read it are developers and managers, and each group reads a different part. So the report needs two layers.

What goes in the executive summary

Keep the summary to one page. State the site's index health in numbers: pages you want indexed versus pages Google has indexed, and the share of URLs that pass Core Web Vitals. Then name the five issues that cost the most traffic and what fixing them requires. A manager can approve the work from this page alone.

What each issue entry needs

Every issue in the detailed section gets the same fields, so developers can turn it into a ticket:

  • The issue in one plain sentence
  • Number of URLs affected, with a link to the full export
  • Evidence such as a screenshot, a Search Console row or a log extract
  • The fix written as a technical instruction
  • Estimated effort in hours or story points
  • Owner and target date

Add a baseline at the end that records the crawl date and the tool version, along with the settings for that version. Without it, nobody can compare the next audit with this one. Maria Belkina, who ran Semrush's Site Audit product, made the case for repeat audits after her team studied 100,000 sites: "My goal was to prove the significance of monitoring technical SEO issues on a regular basis rather than on a yearly basis."

Prioritize fixes by impact

Prioritize fixes by how much organic traffic or revenue the affected URLs carry, multiplied by how badly the issue limits them, and then weigh that against effort. How common an issue is tells you little. Ahrefs found meta descriptions missing on 72.9% of sites, yet they don't affect rankings. A canonical tag pointing to a broken page turned up on only 2.6% of sites, and it's a serious problem wherever it appears.

Sort the issues into four tiers and work through them in order:

  1. Blocks indexing: noindex or robots.txt blocks on pages that earn traffic, 5xx errors, wrong canonicals.
  2. Hides content: links or main text that appear only after rendering.
  3. Splits signals: duplicate pages, redirect chains, crawl waste on large sites.
  4. Weakens ranking or clicks: failing Core Web Vitals, broken structured data, missing alt text.

Inside each tier, fix the issues that touch the most valuable templates first. A template fix that takes 2 hours and covers 3,000 product pages beats a manual fix on 40 blog posts. Pull clicks per URL from Search Console's Performance report and join them to your issue list. Then the ranking stops being a matter of opinion.

What to do next

Your next step is to re-crawl with the same settings 2 weeks after the first tier of fixes ships, and to compare the two crawls side by side. That comparison shows which fixes worked and which tickets got closed without a working change. After that, schedule a monthly crawl so new releases can't quietly undo the work. Buyers now find you through AI assistants as well as Google's index, so check whether pages that pass a technical SEO audit appear in those assistants' answers. Snoika audits websites and tracks brand visibility in Google and in AI assistants, with human review on top. To see where your brand stands, get a free AI visibility report.

FAQ

See how AI talks about your brand