Table of contents
Introduction
English version of an article first published in Polish on 28 April 2025. Read the Polish original.
Technical SEO is the foundation that all of your site's visibility in Google stands on. You can have the best content, the strongest links and a polished keyword strategy, but if Google cannot properly crawl, index and render the site, none of it is worth anything. Really, nothing.
In 18 years of working in SEO I have seen hundreds of websites lose traffic to technical problems while their owners had no idea it was happening. And I do not mean exotic edge cases that only a specialist would care about. I mean ordinary mistakes that keep coming back, site after site.
Below I go through the 10 most common technical SEO mistakes that really cost you rankings. For each one you will find:
- What the problem is and why it is dangerous
- How to diagnose it (tools and commands)
- How to fix it, step by step
- A real example from my own work
1. Crawl
Googlebot visits the page
2. Render
runs code and styles
3. Index
stores it for results
Who is this for? Site owners, marketers and developers who cannot work out why a site with good content does not rank. Do not worry: you do not need to be a programmer to diagnose most of these problems.
Blocking crawling in robots.txt
Why is it a problem?
robots.txt is the first thing Googlebot reaches for when it arrives at your site. Block important resources here by accident and Google simply will not see them.
And what Google does not see does not exist for the search engine. Full stop.
The scenarios I run into most often:
-
Disallow: /: the whole site blocked (a classic left over from the development stage) - blocked CSS/JS directories: Google has no way to render the page
- blocked image directories: the images do not make it into Google Images
- wildcard rules that catch far more than anyone intended
How to diagnose it
Start with the simplest check: type yourdomain.com/robots.txt into the browser and see what is in there.
Then look at Google Search Console → Settings → robots.txt. It shows exactly what Google itself sees, not what you assume it sees.
In Screaming Frog, run a crawl and look at the "Indexability" column. Pages cut off by robots.txt stand out immediately.
# Example of a correct robots.txt
User-agent: *
Allow: /
Disallow: /admin/
Disallow: /cart/
Disallow: /checkout/
Disallow: /search?
Sitemap: https://yourdomain.com/sitemap.xml How to fix it
- Check the current robots.txt file
- Make sure you are not blocking CSS, JS or images
- Remove the
Disallow: /rule if the site is ready to be indexed - Add the path to sitemap.xml
- Test it in Google Search Console → robots.txt Tester
From my own work: A client got in touch three months after a redesign, with organic traffic down by 90%. Can you guess what I found? The developer had left Disallow: / in robots.txt after moving the site from staging to production. The fix took three minutes. The loss: three months. It hurts to think about.
Missing or broken sitemap (sitemap.xml)
Why is it a problem?
A sitemap (sitemap.xml) is simply a list of the URLs you want in Google's index. Without it, Googlebot has to find its own way to your pages through internal links. In practice it looks like this: new or poorly linked pages reach the index with a long delay, or do not reach it at all.
Common slip-ups in sitemaps:
- URLs that return 404 or 301, or carry a
noindextag - URLs from a different domain or protocol (http instead of https)
- exceeding the limit of 50,000 URLs / 50 MB per file
- a sitemap that nobody has submitted in Google Search Console
How to diagnose it
Go to yourdomain.com/sitemap.xml and see whether it exists at all and whether the URLs in it are current. Then in Google Search Console → Sitemaps check the processing status and the number of discovered URLs.
How to fix it
- Generate the sitemap automatically, with an SEO plugin, a framework or a generator
- Keep only canonical, indexable URLs that return 200 OK
- Split large sites into thematic sitemaps and tie them together with a sitemap index
- Point to the sitemap in
robots.txtand submit it in Search Console
A sitemap is not a dumping ground for everything you have. It is a short list of the pages you really want to see in Google.
From my own work: An online shop had 12,000 URLs in its sitemap and barely 800 in the index. It turned out the generator was putting in everything it came across: filter pages, baskets, internal search results. We cleaned the sitemap down to real categories and products, and within a few weeks indexing got moving. Sometimes less really is more.
Duplicate content and missing canonical tags
Why is it a problem?
Imagine the same content sitting under several URLs at once: with filter parameters, with and without www, with and without a trailing slash. Google then has to decide which version to show at all. And instead of one strong page you have several weaker ones: diluted link equity, cannibalisation, wasted crawl budget.
The canonical tag is your way of telling Google directly: "this is the right version". Without it, Google guesses for you, and it does not always guess right.
How to diagnose it
In Search Console → Pages, look out for the "Duplicate" statuses and "Alternate page with proper canonical tag". Screaming Frog, in turn, will catch duplicated titles and content as well as missing canonical tags.
<!-- Canonical tag in the <head> section -->
<link rel="canonical" href="https://yourdomain.com/product" /> How to fix it
- Set a self-referencing canonical on every canonical page
- Choose one version of the domain (https with or without www) and 301-redirect the rest
- Unify URLs with parameters: point the canonical to the base version
- Do not mix canonical with
noindexon the same page, because these are contradictory signals
Slow page loading (Core Web Vitals)
Why is it a problem?
Speed is a ranking factor but, honestly, it is even more of a conversion factor. A slow site loses customers before Google even comes into play. Core Web Vitals describe the real user experience in three numbers:
- LCP (Largest Contentful Paint): the time it takes the main content to load, target < 2.5 s
- INP (Interaction to Next Paint): responsiveness, target < 200 ms
- CLS (Cumulative Layout Shift): layout stability, target < 0.1
How to diagnose it
Use PageSpeed Insights (lab and field data), the Core Web Vitals report in Search Console and the Lighthouse tab in Chrome DevTools. My advice: look mainly at the field data. It shows how real users experience the site, not how a lab does.
How to fix it
- Compress images and serve them in WebP/AVIF, add
width/height - Turn on lazy loading for content below the fold
- Minify and defer unnecessary JavaScript, limit third-party scripts
- Turn on caching and a CDN, reserve space for elements (to eliminate CLS)
No mobile responsiveness
Why is it a problem?
Google evaluates sites on a mobile-first basis, which means it looks first and foremost at the mobile version. If the text is unreadable on a phone, the buttons are tiny and the content spills off the screen, you lose both rankings and customers. And we are talking about more than half of all traffic.
For Google, your site is above all the one on the phone. Desktop is the add-on today, not the other way round.
How to diagnose it
Open Chrome DevTools in device mode, run the Mobile-Friendly Test and look at the "Mobile Usability" report in Search Console. And most importantly, pick up your phone and tap around the site like a normal person. No tool can replace that.
How to fix it
- Use a responsive layout (RWD) instead of a separate mobile domain
- Add a correct
<meta name="viewport"> - Tap targets of at least 48×48 px, readable typography (≥16 px)
- The same content and structured data on mobile and desktop
Indexing errors (noindex, nofollow in the wrong places)
Why is it a problem?
The noindex tag is a clear message to Google: "do not show this page in the results". The trouble is that it can land on important pages by accident, after a migration or globally from the theme settings, and quietly sweep them out of the index. nofollow, in turn, shuts off the flow of link equity.
One stray noindex can do more damage than ten weak articles. And it is silent: nothing blows up, the traffic simply disappears. How to diagnose it
In Search Console, look for the status "Excluded by 'noindex' tag", and in Screaming Frog use the Directives → Noindex filter. Do not forget the X-Robots-Tag HTTP header. It works exactly like the meta tag and is easy to miss, because it is not visible in the page code.
How to fix it
- Remove
noindexfrom pages that are meant to rank - Keep it deliberately only on the basket, login, thank-you pages and the like
- Check the CMS setting that discourages search engines from indexing the site
- After the fix, use "URL Inspection → Request indexing"
Redirect chains and 301/302 errors
Why is it a problem?
Every redirect is an extra hop and a little lost link equity. Now imagine a chain: A → B → C → D. The page slows down, crawling gets harder and the equity thins out along the way. On top of that, using a 302 (temporary) where a 301 (permanent) belongs means Google does not pass full equity to the new URL.
How to diagnose it
Screaming Frog → the Redirect Chains report lays out all the chains and loops for you in plain view. It is also worth looking at the response codes in the server logs or in a tool such as Redirect Path.
How to fix it
- Shorten chains to a single hop, from A straight to D
- Use 301 for permanent changes and keep 302 for what is truly temporary
- Update internal links so that they point straight at the final URL
- Get rid of redirect loops and redirects that lead to a 404
From my own work: After two migrations in a row, one site had chains of four or five hops on its old URLs. Each new team had added its own redirect on top of the previous ones. Flattening everything to a single hop and updating the internal links sped up crawling so much that Google started visiting new pages noticeably more often.
No structured data (Schema Markup)
Why is it a problem?
Structured data (Schema.org) helps Google work out the context of your content, and it unlocks rich snippets: review stars, prices, FAQ, breadcrumbs, recipes. Sounds trivial? Yet they make the difference: the result in the SERP becomes larger and more clickable, so at the same position you get a higher CTR.
<!-- JSON-LD example for an article -->
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "Article",
"headline": "Technical SEO: 10 mistakes",
"author": { "@type": "Person", "name": "Sebastian Baryś" }
}
</script> How to fix it
- Match schema types to the content (Article, Product, FAQPage, LocalBusiness…)
- Implement them in JSON-LD, the format Google recommends
- Check the result in the Rich Results Test and in the Schema Markup Validator
- Do not mark up content that is not on the page, because that is a straight road to a penalty
JavaScript rendering problems
Why is it a problem?
When key content or links appear only after JavaScript runs, which is everyday practice in SPAs built with React, Vue or Angular, Google has to go through a costly rendering stage. And the outcome varies: sometimes the content never reaches the index at all, and sometimes it gets there with a serious delay.
If content is visible only after JS has run, then for Google it is "for later" at best. And in SEO, "for later" often means "never".
How to diagnose it
In Search Console, go to URL Inspection → Test live URL → Screenshot / HTML and see what Google has actually rendered. A useful trick: turn off JavaScript in the browser and compare the raw source code with the rendered DOM. The differences will tell you everything.
How to fix it
- Use SSR or SSG (e.g. Astro, Next.js) for SEO-critical content
- Alternatively, dynamic rendering / prerendering for bots
- Links as real
<a href>elements, not onClick events - Do not block JS/CSS files in robots.txt
No HTTPS and security problems
Why is it a problem?
HTTPS is a confirmed ranking factor and simply the basis of trust. A site without an SSL certificate gets a "Not secure" label from the browser, which puts visitors off very effectively and pushes up the bounce rate. In addition, mixed content damages both the security and the appearance of the page.
How to diagnose it
Look at the padlock in the address bar and check that the certificate is valid. Mixed content warnings will show up in DevTools → Console. And make sure http redirects to https with a 301, a detail that is often overlooked.
How to fix it
- Install an SSL certificate. It is often free, for example from Let's Encrypt
- Force a 301 from http to https and unify the www / non-www version
- Switch all resources to https and say goodbye to mixed content
- Update the canonical tags, the sitemap and the address in Search Console
FAQ: the most common questions
What exactly is technical SEO?
It is the optimisation of the technical side of a website, the parts that decide whether a search engine can crawl, index and render it at all. That covers quite a lot: robots.txt, sitemaps, loading speed, HTTPS, structured data and clean code. Put simply, it is everything that determines whether Google can "see" and understand your site.
How often should I run a technical audit?
A full audit once a quarter is plenty, plus one after every major change: a migration, a redesign, a switch of CMS. Two things I keep under constant watch, though: indexing errors and Core Web Vitals. They can break overnight, so I check them in Search Console regularly, not once every three months.
Do I need to know how to code to fix this?
You can handle many of the fixes yourself, from the CMS or an SEO plugin: robots.txt, the sitemap, noindex, simple redirects. It gets harder with JavaScript rendering, Core Web Vitals and server-level redirects. That is where a developer, or someone who works with this every day, usually comes in useful.
Which technical mistakes are the most dangerous?
The ones that hurt fastest are those that throw a site out of the index: a faulty robots.txt that blocks the whole site and an accidental noindex. A small thing like that can wipe out organic traffic within days, and I have seen it more than once. The rest (speed, duplicates, schema) works more slowly, but over time it drags rankings down just as effectively.
Is technical SEO alone enough to rank high?
No. It is the foundation, not the whole house. Technical SEO removes barriers and lets Google assess the site fairly, but rankings also depend on content quality, user intent and the link profile. Treat it as a necessary condition: without it, everything else works at half power.
What can I use to check technical SEO for free?
You can cover the basics without spending anything: Google Search Console (indexing, Core Web Vitals, sitemaps), PageSpeed Insights (speed), the Rich Results Test (structured data) and Chrome DevTools with Lighthouse. The free version of Screaming Frog crawls up to 500 URLs, which is plenty for a smaller site.
Summary
Technical SEO is not a "fix it and forget it" job. It is a process that never stops and needs regular attention. Every change to the site (a new page, a CMS update, a redesign) can quietly introduce a fresh technical problem. So do not ask whether something will break, ask when you will catch it.
The cheapest technical mistake is the one you catch before Google has a chance to notice it.
Technical SEO checklist
Sources and further reading
Official documentation and tools worth basing technical decisions on:
Not sure which of these mistakes your site has?
Order a technical SEO audit. I will diagnose all the problems and prepare a prioritised list of fixes.
Order a technical SEO auditHow this text was written
I will not pretend it was written entirely by hand, by candlelight. For the research, the proofreading and the code of this page I used a language model. It is a good tool for digging through documentation and catching the typos you no longer see on the third read of your own text. The English version was also prepared with a model, from my Polish original.
What the model did not do: it did not choose the topic, form the arguments or decide what this data means for your site. That stays with me, as does the last word on what goes into the text and what comes out of it.
Where the model helped
- digging through sources and organising data
- proofreading and the English version of the text
- code, page layout and infographics
What stays with me
- the choice of topic and what is worth saying about it
- the arguments, the reading of the data and the conclusions
- fact-checking and the decision on what goes into the text