Technical SEO: the practical guide

What is technical SEO?

Technical SEO is the optimization of all technical factors of a website so that search engines, and increasingly AI systems, can crawl, index, and understand it without errors. It is not about the words on the page but everything beneath them: server speed, code structure, reachability of pages, clean signals to the crawler. If content is the offer, technical SEO is the road it actually arrives on.

SEO splits roughly into three areas. On-page SEO covers everything on the page itself: titles, headings, internal linking, and content. Off-page SEO is about signals from outside, mainly backlinks. Technical SEO is the foundation under both: without clean technology the best content fizzles out, because Google cannot capture it properly in the first place. That is exactly why a technical audit comes at the start of any serious SEO work, not at the end.

The areas of technical SEO

Technical SEO is not a single lever but a bundle of clearly separated areas. The following are the ones that decide between success and stagnation in almost every audit. We go through them in order, the same way we work through them in a real project: reachability first, then speed, then the finer signals.

Crawlability and indexing

Crawlability means the Googlebot can reach your pages and follow their links. Indexing means they then land in the search index. Both are the basic precondition: what is not crawled cannot rank. Typical brakes are orphan pages with no internal links, a click structure that is too deep, or accidentally set noindex tags.

In Google Search Console, the index coverage report shows exactly which pages are in and which were excluded and why. It is the first report we open in every audit. A detail many overlook: a disallow in robots.txt only prevents crawling, not necessarily indexing. To truly keep a page out of the index you need a noindex tag, not just an exclusion in robots.txt.

Load time and Core Web Vitals

Load time is one of the few technical factors Google has officially confirmed as a ranking signal, measured via the Core Web Vitals. These are three values: LCP for the loading experience, INP for the response to clicks, and CLS for visual stability. Important for 2026: INP officially replaced the old metric FID on March 12, 2024, because INP measures every interaction across the whole session instead of only the first. Good targets are an LCP under 2.5 seconds, an INP under 200 milliseconds, and a CLS under 0.1, each measured on real Chrome users.

On our FRAMEN project we built an elaborate, animated website and still reached a green PageSpeed Insights score. That is where clean engineering shows: elaborate design and fast load time are not mutually exclusive, they are a question of execution. We explain the deeper mechanics in our dedicated Core Web Vitals article.

Mobile optimization

Google has indexed mobile first for years. That means the mobile version of your site is the one that gets crawled and evaluated, not the desktop version. Since mobile devices now make up the majority of traffic, this is not a side note. If content is missing on the phone, buttons are too small, or the layout breaks, that is exactly what gets evaluated. A responsive design is therefore not optional but mandatory.

The check is simple: Lighthouse and PageSpeed Insights provide their own mobile score, and in Search Console mobile usability issues show up directly as errors. We additionally test on real devices, because emulators only roughly reflect the real touch targets and load times on weak mobile networks.

Structured data (schema)

Structured data is a code format, usually JSON-LD following the schema.org standard, with which you tell search engines the meaning of your content explicitly. You tell Google not just that there is a number, but that it is a price, a review star, or an FAQ entry. The result is rich results: star ratings, FAQ dropdowns, product data, or breadcrumbs directly in the search result, which can noticeably raise the click rate.

Relevant types for most websites are Organization, LocalBusiness, Article, Product, FAQPage, and BreadcrumbList. You can check them with Google’s Rich Results Test. Important: structured data is not a direct ranking factor. But it makes you more visible in the SERPs and, this becomes decisive in 2026, machine-readable for AI systems without destroying the presentation for humans.

HTTPS and security

HTTPS is the encrypted connection between browser and server and a confirmed, if light, ranking signal. More important is the trust aspect: browsers mark pages without a valid certificate as not secure, which immediately deters visitors. An SSL certificate is standard today and usually free via the host or Let’s Encrypt.

The most common mistakes are not the missing certificate but mixed content, meaning HTTPS pages that still load individual resources like images or scripts over HTTP, and missing redirects from the HTTP to the HTTPS version. You can check both quickly in the browser or with a crawler like Screaming Frog.

Sitemap and robots.txt

The XML sitemap is the map of your website for search engines: a list of all pages that should be indexed, often with the date of the last update. The robots.txt is the signpost file in the root directory that tells the crawler which areas not to enter. Both files work together and are checked first in every audit, because a mistake here has wide-reaching consequences.

The classic, expensive mistake is a robots.txt that accidentally blocks the whole directory or important resources, or a sitemap that still contains old, deleted URLs. You submit the sitemap in Google Search Console and see directly how many of the submitted URLs were actually indexed. A link to the sitemap also belongs in the robots.txt.

Canonicals and duplicate content

Duplicate content arises when the same content is reachable under several URLs, for example with and without www, with filter parameters, with and without a trailing slash, or in print versions. Google then does not know which version to rank and spreads the signals across several URLs instead of bundling them. The canonical tag solves this: it points to the one authoritative version of a page.

In practice the most common sources are online shops with filter and sort parameters, and pages reachable via both HTTP and HTTPS. Screaming Frog lists all duplicates and faulty canonicals at a glance. A clean baseline matters: every URL should have exactly one self-referencing canonical so the signals do not dilute.

HTTP status codes and redirects

Every request to your server is answered with a three-digit HTTP status code, and these codes are an underrated technical lever. 200 means everything is fine. 301 is the permanent redirect you use for moved pages to preserve their rankings. 302 is a temporary redirect. 404 means the page does not exist, and 410 that it was permanently removed. 500 is a server error that needs immediate attention, because it endangers reachability.

The most common mistake here is redirect chains, meaning several redirects one after another. They slow down crawling, waste crawl budget, and can cost rankings. Equally expensive are many broken internal links that lead to 404 pages. You check both regularly with a crawler like Screaming Frog or in Search Console and clean them up consistently.

URL structure, breadcrumbs, and internal linking

The URL is the first impression users and crawlers get of a page. Good URLs are short, descriptive, and hierarchical, with hyphens instead of underscores and the focus keyword in a prominent position. They reveal what is on the page before the click. A logical folder structure helps Google understand the setup of your website.

Breadcrumbs, the small path navigation at the top of a page, are a double win: they help users orient themselves and give Google, marked up as BreadcrumbList schema, a clear picture of the page hierarchy. Internal linking, finally, distributes authority between your pages and leads the crawler to deeper content. Important pages need many internal links, while orphan pages with no links are found poorly or not at all.

JavaScript rendering and SSR

Many modern websites assemble their content only in the browser via JavaScript. The Googlebot can render JavaScript, but does so in a second, delayed step and with a limited budget. If the content arrives late or faulty, it is captured worse or not at all. Server-side rendering (SSR) delivers the finished HTML content directly from the server, so crawlers see it immediately.

For this we rely in many projects on Next.js, a framework that ships SSR and static rendering as standard. The decisive advantage shows in the next area: AI crawlers are noticeably stricter here than Google, and this is exactly where a technically clean site separates from one that stays invisible in the AI era.

International SEO and hreflang

As soon as your website serves several languages or countries, a separate technical layer is added. The hreflang tag tells Google which language version is meant for which region, so German users get the German and Austrian users the Austrian version. If it is missing or wrongly linked, your own language versions compete with each other for the same ranking.

The second decision is the domain strategy. Country domains like .de and .at send a strong local signal but are demanding to run. Subdirectories like yoursite.com/en/ are easier to maintain and bundle domain authority in one place, which makes them the more pragmatic choice for most companies. What matters most is that the mapping of language and target region is technically clean and free of contradictions.

Which tools do you need for technical SEO?

You do not need an expensive suite to start. The most important tools are free or have a usable free tier. This selection covers the bulk of the work, from finding errors to measuring performance:

Technical SEO in the AI era

Here is the point most checklists do not yet capture: AI crawlers like GPTBot (OpenAI) and ClaudeBot (Anthropic) do not render JavaScript. They read the raw HTML the server delivers, like a browser from 2010. If your content is only created in the browser via JavaScript, ChatGPT and Claude see an empty page. They cannot cite you, because they do not even know what is on your page.

That turns server-side rendering from an SEO advantage into the entry ticket for AI search. Added to that, the old technical virtues gain new importance: clean structured data, fast load times, and a layout that puts the core statement up front, because a large share of AI citations comes from the upper part of a page. Whoever wants to become visible in AI answers cannot retrofit that at the content level, it starts at the technical base of the website. This is exactly the bridge between technical SEO and Generative Engine Optimization.

This is precisely where we, as a studio, work differently than a pure SEO agency: we build the technical base clean from the start instead of repairing it afterward. SSR, performance, and machine-readable structure are part of how we build, not a later optimization project.

Your first steps

Technical SEO looks overwhelming as a checklist but is logical in order. First you check whether Google can even get in, then you clear the coarse errors, then you refine speed and signals. These five steps are the sequence with which we start every audit.

Frequently asked questions