echnical SEO should be a build requirement, not a mysterious add-on after launch. It is the work that lets search engines fetch, render, understand and revisit the right pages while keeping private or duplicate surfaces out of the index.
It cannot guarantee rankings: demand, competition, reputation, links and content quality still matter. It can prevent avoidable losses and make editorial work eligible to perform. A proposal should name deliverables, covered templates and acceptance evidence.
Public marketing, service, legal and article pages should normally return meaningful HTML without a session. Static generation or incremental regeneration can reduce server work and improve consistency. Authenticated dashboards, token links and private documents should retain request-time controls and must not become public merely for speed.
tsx
// Public article segment: cacheable HTML with daily refresh export const revalidate = 86_400; // Private routes need authentication and request-time checks. // robots.txt is not an access-control mechanism.
- A robots.txt file that permits public assets and references the canonical sitemap.
- XML sitemaps containing only canonical, indexable, successful URLs.
- Page-level robots metadata or X-Robots-Tag for crawlable pages that must not appear in search.
- Authentication, authorisation, expiring tokens and appropriate 401, 403 or 404 responses for confidential content.
- Canonical URLs based on the production origin, with one hostname and protocol.
- Intentional handling for filtered, paginated, duplicate and retired URLs.
Robots directives are indexing instructions, not locks. Blocking a confidential URL in robots.txt may prevent a crawler seeing its noindex instruction, while links can still reveal the URL. Private information requires real access control.
Every indexable template should produce a useful title, description, canonical and social preview. Metadata should reflect the page's intent rather than repeat one generic phrase. The system should prevent staging domains, missing images and duplicate canonicals reaching production.
- Unique titles that put the topic before repeated branding.
- Descriptions written for the search result, not keyword stuffing.
- Open Graph and social-card data with stable absolute image URLs.
- Correct language, viewport and icon metadata.
- Article dates and authorship where maintained and visible.
Organisation, website, breadcrumb, service and article markup can clarify meaning, but it must be valid and supported by visible content. Schema is not a place to invent reviews, prices or claims. Test the rendered production output because valid TypeScript can still emit contradictory JSON-LD.
- Permanent one-hop redirects for changed URLs, based on an approved map.
- A real 404 response for missing resources, with a helpful human page.
- No loops, long chains or mass redirection of unrelated pages to the homepage.
- Consistent trailing slash, casing and hostname behaviour.
- Preservation of campaign parameters where measurement needs them.
A crawler and buyer should reach important pages through ordinary links. Navigation, breadcrumbs, topic hubs and contextual article links are stronger than an orphaned sitemap entry. Labels should explain destinations instead of repeating “read more”.
Set budgets for images, fonts, scripts and client JavaScript, then measure representative templates on real mobile conditions. Core Web Vitals are useful signals, but the commercial goal is a page that becomes useful quickly and stays responsive. Ask which third-party scripts count and who approves new ones.
- Search Console and analytics ownership belongs to the client organisation.
- Production events are documented and tested without unnecessary personal data.
- A pre-launch crawl is compared with the old site when URLs or content change.
- Key templates are checked for status, canonical, metadata, schema, headings and links.
- The launch plan includes monitoring, rollback criteria and a named owner for fixes.
Technical SEO is maintainable only when the next person can understand it. Handover should cover domain and Search Console access, sitemap and robots locations, redirect rules, metadata conventions, schema sources, performance budgets, monitoring, deployment access and safe publishing instructions.
- Public priority pages return meaningful, cacheable HTML.
- Private pages require real access and return noindex where appropriate.
- Canonicals, metadata and social cards are correct across indexable templates.
- Sitemaps, robots rules, redirects and 404s match the approved inventory.
- Structured data validates and matches visible facts.
- Representative pages meet agreed performance and accessibility budgets.
- Search, analytics and hosting accounts are client-owned and documented.
TechnicalSEOiscompleteonlywhenthelaunchcanbeinspected,therulescanbemaintainedandtheclientownstheevidence.