Technical SEO is the work that makes a site reachable, readable and quick enough for search engines to use. It doesn’t decide whether your content deserves to rank. It decides whether it gets the chance.
The gate model
The most useful way to hold technical SEO in your head is as a sequence of gates. A page has to be discovered, then crawled, then rendered, then indexed, before it can be ranked.
Technical SEO owns the first 4. Content and links fight over the fifth. Miss a gate and nothing downstream matters, which is why a brilliant article on an unreachable URL earns exactly nothing, and why diagnosis should always start by asking which gate failed.
| Gate | What it needs | Where it shows when it fails |
|---|---|---|
| Discovery | Internal links, sitemaps | URL absent from Search Console entirely |
| Crawling | Robots.txt access, healthy server | Discovered, currently not crawled |
| Rendering | Content present after JavaScript runs | Rendered HTML missing content in URL Inspection |
| Indexing | Something worth storing | Crawled, currently not indexed |
| Ranking | Relevance, quality, links | Indexed but no impressions |
What it covers
- Crawlability and rendering. Whether bots can reach and read your pages.
- Indexing control. Robots directives, canonicals, and what you allow into the index.
- Site architecture and internal linking. How authority and discovery flow through the site.
- Performance. Core Web Vitals, server response time, page weight.
- Mobile behaviour. Google indexes the mobile version, so the mobile page is the page.
- Structured data. Making entities machine-readable.
- Redirects and status codes. Migrations, chains, soft 404s.
- International targeting. Hreflang and market separation.
A worked example
A retailer with 40,000 products was certain they had a content problem: reasonable descriptions, almost no organic traffic, an agency recommending a rewrite.
Their category navigation rendered entirely client-side. The only internal links Googlebot could reliably follow were the 12 in the footer, so roughly 90% of the catalogue had never been crawled at all. The rewrite would have produced better descriptions on pages Google had never seen.
The fix was a server-rendered category navigation. That’s technical SEO earning its place: not a tweak to a page that ranks, but the reason 36,000 pages had no chance.
What actually deserves your time
Technical SEO attracts more busywork than any other part of the discipline, because audit tools generate hundreds of findings and most of them change nothing. A rough order of priority for a site that isn’t enormous:
- Anything blocking a gate. Pages not crawled, not indexed, or rendering empty. These are binary failures.
- Server speed. It affects crawling, ranking and conversion at once, and it’s usually the highest-leverage fix available.
- Internal linking. The lever you fully control, and the one most sites neglect.
- Mobile experience. Because it’s the version that counts.
- Structured data on templates. Cheap at scale once the template is right.
- Everything the audit tool flagged in amber. Last, and only if the above are done.
Common mistakes
- Treating the audit as the deliverable. A 200-issue crawl report is a to-do list, not an improvement. Rank findings by which gate they block.
- Chasing tool scores. Getting an audit tool to 100 out of 100 correlates with nothing. Google does not run that tool.
- Fixing problems you don’t have. Crawl budget work on a 300-page site is the classic.
- Leaving it to developers alone. The decisions that cause technical SEO problems are usually made in a framework choice months earlier, by people who were never told search mattered.
- Treating it as a one-off project. Sites break continuously. A deployment can undo a year of work in an afternoon.
Related
How Google Search works covers the full pipeline. Crawling and indexing covers the first gates in detail.