Technical SEO Audit: What Actually Matters (And What Is Theatre)

A technical SEO audit has two failure modes. It is either a list of 400 issues, most of which do not matter, or it is a single paragraph that says "your site is fine" and misses the one thing that is broken. Both are common. Neither helps.

This is the version that works: the checks that change rankings, in the order to run them, with a note on what to ignore.

1. Can Google reach and index the pages that matter

Everything else is secondary. If a money page cannot be crawled or is blocked from indexing, no amount of content work will fix it.

• robots.txt: does it block anything it should not? A common failure is a leftover Disallow from a staging environment.
• Noindex tags: are they on pages that should be indexed, or missing from pages that should not be? Check the page source, not the rendered DOM.
• Canonical tags: does each page declare itself as the canonical, or point somewhere else? Self-referencing canonicals are the safe default.
• XML sitemap: does it exist, is it submitted, and does it contain only URLs that return 200 and are indexable?
• HTTP status: are important URLs returning 404, 301 or 500? A 404 on a page Google already ranks throws away its authority.

2. Is the site fast enough to not lose the visit

Core Web Vitals are a tiebreaker, not a ranking factor in the way people mean. They matter most where the competition is close, and where the page is slow enough that visitors leave before it finishes loading.

• Largest Contentful Paint under 2.5 seconds. The usual culprit is an oversized hero image or a render-blocking script.
• Interaction to Next Paint under 200 milliseconds. Heavy JavaScript on the main thread is the common cause.
• Cumulative Layout Shift under 0.1. Ads, late-loading fonts and images without dimensions are the usual suspects.
• Measure on a real mid-range phone, not a desktop with a fast connection. The lab score is not the user's experience.

3. Is the site understandable to machines

This is where most audits stop too early or go too far. The goal is not to mark up everything; it is to remove ambiguity about what the page is about.

• One H1 that says what the page is about. Multiple H1s are not a catastrophe, but one clear one is better.
• A title tag and meta description that match the page's actual content, written for a human reader.
• Structured data that states the facts: Organization on the homepage, Service on service pages, Article on articles, FAQPage where there are real questions. Validate it, because invalid markup is worse than none.
• Internal links that use descriptive anchor text and connect related pages. A page with no internal links pointing to it is hard to find and hard to rank.

4. Duplicate and thin content

The point here is not length. It is whether two URLs on your site compete for the same query, or whether a page exists that says nothing a competitor does not also say.

• Near-duplicate pages: two URLs with the same content and the same intent. Canonicalise or merge them.
• Doorway pages: pages created only to rank for a keyword, with no genuine content behind them. These are a spam-policy violation, not a tactic.
• Thin pages: pages with so little substance that they cannot answer a question. Either build them out or remove them.
• Parameter and filter URLs: tracking parameters, sort options and session IDs that create infinite URL variants. Canonicalise or block them.

What to ignore

Most of what an automated audit reports does not move rankings. Spending time on it is the reason audits feel unproductive.

• Meta keyword tags. Google has ignored them for over a decade.
• Image alt text on decorative images. Alt text matters for informative images and accessibility, not for decoration.
• Minor HTML validation errors. Search engines parse broken HTML fine.
• Social meta tags beyond Open Graph and Twitter card basics. Nice to have, not a ranking signal.
• A score out of 100 from an audit tool. The score measures the tool's checklist, not your traffic.

The order to run it

1. Crawlability and indexability first. Fix anything that stops Google reaching or indexing a page.
2. Redirects and 404s on URLs that already have authority or impressions.
3. Core Web Vitals on the pages that get traffic.
4. Structured data on the pages that should be rich results.
5. Duplicate and thin content last. It matters, but only after the basics work.

A technical audit is a diagnosis, not a treatment. The value is in the order in which you fix things, and in stopping when the fix stops paying.