What is technical SEO? Basics, evaluation areas and priorities

Google describes crawling, indexing, and serving as fundamental stages. Rendering is particularly relevant when content or links are generated through JavaScript.

Crawling: Discover and retrieve URLs

Crawler discover URLs via known pages, links, and sitemaps and then fetch them from the server. DNS issues, timeouts, server errors, robots rules, or non-crawlable links can hinder this step.

Rendering: process the browser view

On JavaScript sites, a rendering system additionally processes the resulting HTML. Blocked, erroneous, or resources loaded only after user interaction may be missing.

Indexing: Content and signals categorize

After a page is retrieved, the search system analyzes content and signals. It also checks for duplicates and selects a canonical version. A technically accessible page does not have to be indexed.

For diagnoses, separation is essential: „not found,“ „not rendered,“ and „not indexed“ have different causes.

A clear website architecture guides users and crawlers from navigation and topic hubs to important detail pages. Reliable links consist of an HTML-<a></a>-element with accessible hrefRein visual click areas, search forms or JavaScript actions without crawlable target URLs are no equivalent substitute.

Technically verified are above all:

• Accessibility of important pages via internal links
• Click depth and distribution of links between templates
• orphaned pages without incoming internal links
• Filter, sort, calendar, and parameter URLs
• Referral chains and broken destinations
• Breadcrumbs, Hubs, and contextual anchor texts

An XML sitemap supports discovery but does not replace meaningful internal linking. How link architecture, anchor texts, and topic clusters are planned is explained in the guide Internal linking for SEO.

Crawl-budget is primarily relevant for very large or rapidly changing websites. For smaller sites, clear links, up-to-date sitemaps, stable servers, and fewer unnecessary URL variants are often more important.

robots.txt and noindex resolve different tasks:

• robots.txt: Allows or prohibits crawling specific paths for named crawlers.
• Meta Robots: contains page-specific rules such as noindex in HTML.
• X-Robots-Tag: Provides comparable rules in the HTTP header, for example for PDFs.
• Access protection: prevents public access, for example in internal test systems.

A per robots.txt blocked URL may still be known. Additionally, the crawler may have noindex do not read on the blocked page. Therefore remove robots.txt URLs not securely obtained from search results.

Confidential content requires genuine authentication; robot rules are not a data protection mechanism. Temporary staging blocks must not be carried over at go-live.

The HTTP status describes what happened with a request. For technical SEO, the following cases are particularly relevant:

• 200 OK for successfully delivered content
• 301 or 308 for lasting moves to professionally appropriate destinations
• 302 or 307 for actually temporary forwarding arrangements
• 304 Not Modified for reuse of an unaltered resource upon a matching cache request
• 404 Not Found or 410 Gone, if a content has been removed without an appropriate replacement
• 429 Too Many Requests or 503 Service Unavailable only as a temporary capacity or availability issue

A error page with status 200 can be considered a soft 404. Permanently removed URLs should not generally redirect to the homepage; redirect targets must align with factual accuracy.

During migrations, old and new URLs are systematically mapped, direct redirects are tested, and internal links are updated to point to the final destinations. The complete preparation describes the SEO Relaunch Guidance.

Google can execute JavaScript, but it does have limitations. Therefore, important content, links, titles, canonicals, and robots signals should not depend on fragile events or unnecessary requests.

A JavaScript check compares at least:

• HTTP response and original source code
• rendered HTML
• visible mobile display
• Network and console errors
• Links before and after rendering
• dynamically set metadata and structured data

Server-side or pre-rendering can help, but is not a blanket requirement. URL checks and render tests on representative page types show the actual result.

Technical SEO Performance Binds servers, frontend, and real user experience. The Core Web Vitals measure with Largest Contentful Paint (LCP), Interaction to Next Paint (INP), and Cumulative Layout Shift (CLS) loading performance, responsiveness, and visual stability.

Google specifies target values of LCP within 2.5 seconds, INP below 200 milliseconds, and CLS below 0.1. Field data reflects real-world use; lab data helps with reproducible troubleshooting.

Common levers are:

• Server response, caching and CDN configuration
• Image dimensions, formats, and prioritization of the primary image
• Fonts and render-blocking resources
• Scope and execution time of JavaScript
• Layout reservation for media and embedded elements

A green score guarantees no ranking. Users still benefit from stable and fast-reacting pages.

Structured data describe page types, properties, and relationships in a standardized format. They can provide search systems with additional hints and enable appropriate displays. The markup, however, must accurately represent the visible content.

Technically relevant are:

• a suitable main type instead of random schema collection
• complete required properties for supported search functions
• stable IDs for organizations, people, and other central entities
• consistent URLs, names, images, and relationships
• no staging, old-domain, or placeholder values
• Review with Rich Results Test and URL Check
• Monitoring after template or plugin updates

Even valid Markup does not guarantee a Rich Result and does not replace visible facts or quality.

International websites require distinct language and optionally regional variants. hreflang Connects corresponding pages together, but neither replaces translation nor a meaningful URL structure.

Every variant must reference all associated alternatives including itself. Backlinks must be complete. Canonicals typically point to the canonical page of the same language; a canonical for all translations pointing to only one language version may undermine the desired international targeting.

Additionally, language switchers, redirects, mobile parity, sitemap clusters, and x-default, if a neutral selection or fallback page exists.

Generative search and response systems require accessible, interpretable, and preferably unambiguous information. Technical SEO can improve the prerequisites for this:

• stable canonical URLs and clear page responsibilities
• crawlbare links between topics, authors, services, and sources
• visible, precise facts in understandable HTML
• consistent personnel and organizational signals
• correct structured Markup
• reachable sources, images and documents
• intentionally configured access rules for different crawlers

These measures facilitate machine processing, but guarantee neither inclusion in a specific index nor mention, linking, or citation. GEO additionally requires helpful content, verifiable evidence, clear entities, and external verification.

An audit begins with goals, key page types, platforms, releases, and known symptoms. Then follow samples and automated checks of relevant URL groups.

A feasible prioritization evaluates:

1. Reach: How many and which pages are affected?
2. Impact: Does the error prevent crawling or indexing, or does it merely reduce the technical quality?
3. Business value: Does it concern important services, categories, products, or conversion paths?
4. Security: Is the cause proven or initially only a hypothesis?
5. Effort and risk: How complex is the implementation, and what side effects are possible?
6. Verifiability: Are there clear acceptance criteria and monitoring after release?

A SEO Audit Combines technical audits with content, search intent, and internal architecture. This way, critical blockers are addressed first and cosmetic tool warnings are not equated with business-relevant issues.

• Crawling, indexing, and ranking as one and the same
• robots.txt to remove already known URLs
• noindex place on URLs and simultaneously block their crawling
• Canonicals, sitemaps, and internal links to different variants configure
• redirect all 404 pages to the homepage
• check only desktop source code instead of mobile rendered HTML
• Measure Core Web Vitals only on the homepage or exclusively in the lab
• output structured markup that doesn’t fit the visible content
• implement hreflang without complete backreferences
• after releases, no further review and event documentation to be performed

When important pages are not indexed, JavaScript and canonical tags are unclear, or a major release is imminent, a focused diagnosis creates a reliable prioritization for implementation.