Technical SEO on a site that already exists: the fix list in priority order

Technical SEO checklists tend to be long, unordered and therefore not very useful. Everything on them is true, but they give you no way to tell the difference between the item that will change your traffic and the item that will change a number in an auditing tool.

This is the order I would work in on a site that is already built and already has content.

1. Can the pages be crawled and indexed at all

Nothing else matters until this is settled, and it is wrong more often than people expect.

Compare three numbers: how many pages you have published, how many Google has crawled, and how many it has indexed. Search Console gives you the second and third. The gaps between them are your entire diagnosis.

Published but not crawled means Google does not know the pages exist. Usually a missing or stale sitemap, or pages that nothing links to.

Crawled but not indexed means Google looked and decided not to bother. That is a content quality or duplication signal, and no amount of technical work fixes it. This is the finding people least want and most need.

Indexed but not the version you wanted means duplicates are competing with each other.

The specific things worth checking here: a robots.txt that is accidentally blocking something important, noindex tags left over from a staging environment, and a sitemap that lists URLs which redirect or return 404. A staging noindex shipped to production is one of the most common serious technical problems there is, and it is invisible unless you look.

2. One canonical URL per page

Most sites serve the same page at several addresses: with and without www, with and without a trailing slash, over HTTP and HTTPS, with tracking parameters attached. To a search engine those can be separate pages with duplicate content, splitting whatever authority the page has earned between them.

Pick one form, redirect everything else to it with a 301, and set a self-referencing canonical tag on every page. This is unglamorous and it is the highest-value hour in the list after indexation.

3. Titles and descriptions that are actually different

Auto-generated titles produce sites where forty pages share a title and every description is the same sentence. The title tag is still one of the clearest signals about what a page is for, and it is what a person reads before deciding whether to click.

Write them per page. Put the distinguishing thing first. Do not stack keywords: a title that reads like a list performs worse than one that reads like a description, because the click-through rate is part of how the page is judged.

4. Internal links, because they are the cheapest thing you own

Internal linking is the most underused lever in technical SEO. It tells search engines which of your pages matter, and it does not depend on anyone else's cooperation.

Two things to look for. Orphan pages, which nothing links to, so they are discovered only through the sitemap and treated as unimportant. And the pages that matter commercially, which should be linked to from the content that gets traffic. If you have an article that ranks and a service page that converts, the article should link to the service page in the body text, with anchor text that describes the destination.

5. Structured data, but only the kinds that apply

Structured data does not raise rankings by itself. It changes how a result is displayed, and a result that shows more information gets clicked more.

Worth adding: Organization on the site, Article on articles, FAQ where there is a genuine question and answer list, Product where you sell products, and BreadcrumbList where you have a hierarchy. Not worth doing: marking up content that is not visible on the page, which is a guidelines violation and can be penalised.

6. Core Web Vitals, last and honestly

Page speed is a ranking factor. It is a small one, and it is the item that receives the most attention relative to its effect, because it produces a satisfying score that goes up.

Fix it when the site is genuinely slow. A page taking six seconds to show its main content is losing visitors before it loses rankings, and that is reason enough. But a site that is already reasonably quick will gain more from one properly written page than from shaving another 200ms off a score.

What this looks like in practice

Indexation and canonicalisation first, because they can be worth a step change and are usually a day of work. Titles and internal links next, because they are steady compounding gains. Structured data and performance after that.

If you want the same order applied to your own site with the findings written out and prioritised rather than dumped as a tool export, that is what a technical SEO audit is, and it is a one-off piece of work with no retainer attached.