Detailed guide contents
Engineering-led guidance. This guide draws on Productionise's standards and practical engineering experience. Examples are generalised to protect confidentiality.
What technical SEO can and cannot do
Technical SEO removes avoidable barriers such as inaccessible content, duplicate URL signals, missing metadata and broken internal links. It also helps site owners observe how search engines see the site.
Rankings depend on many factors outside the code: relevance, competition, content quality, reputation and external authority signals such as genuine backlinks. Publishing more pages does not by itself create useful search performance.
- Make useful public content crawlable and indexable.
- Describe each page accurately and consistently.
- Consolidate duplicate URL signals.
- Measure search visibility over time without promising an outcome.
Rendered HTML, semantics and accessibility
Important content should appear in the HTML a crawler receives or in reliably rendered output. Heavy client-side rendering can delay or complicate discovery, so verify the built result rather than assuming a framework handles everything.
Use semantic landmarks, one clear primary heading, logical subheadings and descriptive links. These choices help people using assistive technology and give search systems clearer structure. Accessibility is valuable in its own right, not a keyword technique.
Titles, descriptions and canonical URLs
Give each indexable page a distinctive title that reflects its purpose and a concise description written for a prospective visitor. Search engines may rewrite either field, so they should be accurate rather than promotional filler.
A canonical URL identifies the preferred version when multiple URLs could show the same content. Use one HTTPS production origin consistently in metadata, internal links and the sitemap. A canonical is a signal, not a redirect; alternate hostnames may also need platform or DNS handling.
Sitemap.xml and robots.txt
A sitemap lists canonical, indexable URLs that actually exist. Generate it from the same route registry where practical, exclude error and placeholder pages, and avoid fabricated modification dates.
robots.txt tells compliant crawlers what they may fetch and can declare the sitemap location. It is not an access-control system: private information must be protected by authentication and architecture, not merely disallowed in robots.txt.
- Use absolute HTTPS canonical URLs.
- Include only live, indexable pages.
- Keep sitemap membership aligned with generated routes.
- Allow required scripts, styles and images.
- Do not expose a future article just to fill a category.
Structured data, internal links and images
Structured data can describe an organisation, website, service or genuine article when the visible page supports those claims. Use the most specific valid type, serialize JSON safely and avoid invented authors, ratings, prices or business details.
Internal links help readers move to relevant next steps and show how topics relate. Use descriptive anchor text and link only where it assists the reader; a site-wide web of forced links is not useful.
Optimise informative images to appropriate dimensions and formats and provide meaningful alternative text. Decorative images should have empty alternative text. Image work supports performance and accessibility as well as discovery.
Core Web Vitals and JavaScript
Core Web Vitals assess aspects of loading, responsiveness and visual stability using field and laboratory data. They are useful diagnostic signals, not a guarantee of rankings.
Reduce unnecessary JavaScript, size images correctly, reserve space for content and test real devices and networks. Static output can create a strong performance baseline for suitable sites, but architecture, implementation and third-party scripts determine the result.
- Measure before and after material changes.
- Prioritise the main content and customer journey.
- Avoid layout shifts from media and late-loading elements.
- Review third-party scripts for value, privacy and performance.
- Test the exported production HTML and assets.
Search Console, analytics and external authority
Google Search Console can confirm ownership, accept a sitemap and report crawling, indexing and search performance after data becomes available. Submission is not a promise that a page will be indexed.
Privacy-conscious analytics can measure aggregate usage and enquiry-path events. Load it only on the intended production host, avoid personal form values and sensitive query strings, and prevent duplicate automatic and manual page views.
Backlinks from relevant, trustworthy websites can act as an external authority signal. Earn them through genuinely useful work, partnerships and references; do not treat bulk link schemes as a technical shortcut.
- Verify the correct domain property without publishing verification values.
- Submit the full canonical sitemap URL.
- Review indexing reports after processing time.
- Configure one page-view mechanism.
- Keep personal and user-entered data out of analytics.
- Use observed data carefully; short histories do not support strong conclusions.
Should I care?
This matters if…
- You want search engines to discover and understand the public site.
- The same page can be reached through several hostnames or URL forms.
- Important content is hard to find through normal links, or Search Console reports technical problems.
It may not be urgent if…
- The website is intentionally private, password-protected or still an internal experiment.
- You are still deciding what customer questions the public site should answer.