What technical SEO covers
Technical SEO is the work of making a site reliably crawlable, indexable and understandable: clean URL structure, correct canonicals, accurate robots and sitemap rules, valid structured data, server-rendered content, and pages that load fast and stably.
Why it comes before everything else
Content strategy assumes your pages can be reached. If a template ships with a canonical pointing at the homepage, or a release moves rendering client-side, the best content in your market will not save you. Technical problems are cheap to fix and expensive to ignore.
This is even more true for generative engines. Many retrieval pipelines do not execute JavaScript, so a page that only renders client-side is not partially visible to them — it is absent.
The pipeline a page has to survive
- Discoverable
- Crawlable
- Renderable
- Indexable
- Understandable
- Retrievable
- 01DiscoverableLinked internally and present in a current sitemap. Orphan pages are frequently never crawled at all.
- 02CrawlableNot blocked by robots rules, not behind a redirect chain, returning a 200.
- 03RenderableContent present in the server response, not assembled after hydration.
- 04IndexableNo conflicting noindex, a self-referencing canonical, no duplicate competing for the same URL.
- 05UnderstandableSemantic HTML, ordered headings, accurate structured data, unambiguous entities.
- 06RetrievableFast, stable, and clean enough that a passage can be extracted without losing meaning.
Canonicals and duplication
Duplication is rarely deliberate. It arrives through parameters, pagination, trailing-slash variants, www and non-www serving the same content, and staging hosts that got indexed. Each variant splits the signals that should have accumulated on one URL.
- One canonical host, with every other variant permanently redirected to it.
- A self-referencing canonical on every indexable page.
- One consistent trailing-slash convention across the whole site.
- Parameters that do not change content excluded from indexing.
- Preview and staging deployments explicitly noindexed.
Structured data that is true
Markup is a description of what is on the page, not an opportunity. Ratings for reviews you do not have, or offers for prices you do not publish, are a manual-action risk and an obvious credibility problem. The useful types are the ones describing what genuinely exists: Organization, WebSite, Article, BreadcrumbList, FAQPage and Product where it applies.
Performance as a search signal
Core Web Vitals matter as a tie-breaker and as a user-experience floor, not as a substitute for relevance. The practical targets are LCP under 2.5 seconds, CLS under 0.1 and INP under 200 milliseconds — and the most common cause of failing them is shipping more JavaScript than the page needs.
Watching for regressions
Most technical damage is introduced by a deploy, not by neglect. Continuous crawling means a canonical that changed, a robots rule that tightened, or a template that stopped server-rendering shows up as a finding — while the release is still fresh in someone’s mind.