Technical SEO: definition, levers and priorities

What is technical SEO?

Technical SEO covers the optimisation of a site’s infrastructure, rather than its copy or its inbound links. Server, HTML code, URLs, directives to robots, JavaScript, speed: the visitor sees none of this, but Googlebot runs into it on every visit.

In the usual split of SEO into three pillars, technical work comes before content and popularity. The SEO guide presents these three pillars; this page details the first one and links to a guide for each lever.

Why does technical SEO come first?

Google processes a page in three stages: crawling, indexing, then serving in the results (Google, “In-depth guide to how Google Search works”, updated on December 18, 2025). Each stage depends on the previous one. Technical SEO acts on the first two, where a single error makes all the editorial work useless.

One setting can be enough to block everything:

• a Disallow: / directive inherited from the staging site;
• a noindex applied to an entire template;
• a canonical tag pointing to the home page.

These errors do not show on screen. You find them in Search Console or in a crawl.

The order of work therefore follows four questions: can Google crawl the page? Render it? Index it? Serve it quickly? As long as the answer to the first is no, working on the fourth brings nothing.

What are the levers of technical SEO?

Seven levers cover most of technical SEO. Each one has its own guide on the site.

Crawling: robots.txt and links

The robots.txt file tells robots which areas not to crawl. Google points out that it is not a way to remove a page from the index: it manages crawling, not indexing (introduction to robots.txt). Syntax errors and accidental blocking are covered in the guide on robots.txt. As for links, only <a href> elements are followed reliably.

Discovery: XML sitemap

The sitemap lists the URLs you want crawled, with their modification date. It helps discovery without guaranteeing indexing. It should only contain pages that return a 200 code and are indexable and canonical. The method is described in the XML sitemap guide.

Indexing: noindex, canonical and duplicates

When several URLs show the same content, Google picks one of them as canonical. Redirects, the rel=canonical tag and the sitemap steer that choice without imposing it (Google, consolidating duplicate URLs). The guide on the canonical URL details the signals, and the one on Google indexing the exclusion reasons in Search Console.

JavaScript rendering

Google processes JavaScript pages in three phases: crawling, rendering, indexing (JavaScript SEO basics). Content or a link that only appears after the script runs is seen later, sometimes badly. The reference test: are the main content and the navigation in the initial HTML? See the JavaScript SEO guide.

Performance and Core Web Vitals

Google measures the loading experience on real visits, at the 75th percentile, separately on mobile and desktop (web.dev, “Web Vitals”). INP replaced FID in March 2024.

Metric	What it measures	Good	Poor
LCP	Rendering of the largest visible element	2.5 s or less	more than 4 s
INP	Responsiveness to all interactions	200 ms or less	more than 500 ms
CLS	Visual stability of the layout	0.1 or less	more than 0.25

Performance mainly serves the user and crawling. A good score does not make up for less relevant content. Measurement method and fixes: the Core Web Vitals guide.

Mobile and HTTPS

Google indexes and ranks pages based on their mobile version (Google, mobile-first indexing). Content, a link or structured data present on desktop but missing on mobile is therefore invisible to it. HTTPS is the norm: 301 redirect from HTTP, internal links in HTTPS, no mixed content.

Structured data

Schema.org markup describes the content of a page in a machine-readable format. It does not push a page up, but it makes it eligible for some rich results and clarifies entities. It must describe what is visible on the page: see the structured data guide.

Where should you start, depending on your type of site?

Technical priorities change with the size and technology of the site. The following table is a starting point. It summarises our practice, not a Google rule.

Type of site	Common risks	First checks
Brochure site (under 500 pages)	Staging site indexed, pages set to noindex by mistake, slowness	Page indexing report, robots.txt, Core Web Vitals
E-commerce	Facets and parameters that multiply URLs, duplicates, pagination	Canonicals, filter handling, sitemap, depth of product pages
Editorial or media site	Archives and tags indexed, duplicate content, freshness	Indexed pages with no traffic, internal linking of new content
JavaScript application	Content and links missing from the initial HTML	Comparison of raw HTML and rendered HTML, URL Inspection
Redesign or migration in progress	Lost URLs, missing or chained redirects	301 redirect plan, testing before go-live

Which tools for technical SEO?

• Google Search Console: indexing, Core Web Vitals, inspection of a URL and of the version Google crawled, robots.txt report, crawl stats.
• A crawler (Screaming Frog, Sitebulb, Oncrawl): HTTP codes, depth, canonicals, tags, orphan pages, at site scale.
• PageSpeed Insights: field data from visitors and a simulated Lighthouse test for a URL.
• The Rich Results Test and the Schema.org validator: checking structured data.
• Server logs: the only record of what Googlebot actually visits, useful beyond a few thousand pages.

None of these tools replaces interpretation. A crawler flags hundreds of alerts, many of which carry no weight for your site. The value comes from cross-checking sources and from prioritisation.

Technical SEO and AI engines: what changes

For Google, the answer is simple: no additional technical requirement is needed to appear in AI Overviews or AI Mode. The page must be indexed and eligible for a snippet (Google, “AI features and your website”, updated on December 10, 2025). Classic technical SEO therefore remains the foundation.

For other assistants, two points come on top. AI crawlers have their own user agents, which robots.txt can allow or block. And some CDNs block these crawlers by default, whatever the robots.txt says. The guide on AI crawlers explains how to check the actual setting.

Answer engines, for their part, mostly read text. Important content shown in an image, a badly structured PDF or a late-loading JavaScript component is less likely to be picked up.

How can you tell if your technical SEO is healthy?

Five signals are enough for a first check, without any paid tool:

• the number of indexed pages is close to the number of pages you want to see in Google;
• no strategic page appears under “Crawled – currently not indexed” or “Excluded by ‘noindex’ tag”;
• the Core Web Vitals report flags no “poor” URL on mobile;
• inspecting a key page shows, in the crawled version, the same content as the one displayed in the browser;
• no internal link leads to a 404 error or a redirect.

If one of these signals is red, or before a redesign, a full diagnosis is needed. Our technical audit method describes its steps, tools and checklist.