Martin Stepanek 🏳️‍🌈

Not active here anymore, find me on LinkedIn ⤵️

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

If your e-commerce product pages don't have proper canonical tags, Google will index the version with tracking parameters, session IDs, or filter combinations. Not your clean URL.

Every product page needs a canonical pointing to the clean version. UTM parameters, sort orders, and session IDs should never be the indexed URL.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

I check server logs all the time. What I find always surprises people. Google spends most of its crawl budget on pages that don't matter.

Parameter URLs. Old redirects. Dead-end paths.

Log file analysis shows you the truth. Not what Google should be crawling – what it actually is crawling.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

I recently worked with a multilingual site where only the English version ranked. German and Spanish? Invisible. Zero traffic.

The cause was broken hreflang. Page A pointed to page B. Page B never pointed back. Google silently ignores the whole setup when that happens. And a missing sitemap killed it completely.

No error in Search Console. No warning. You're losing entire markets, and Google won't tell you.

Run Screaming Frog or Ahrefs Site Audit to catch missing return tags. Without them, you're flying blind.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 #TechSEO Myth Debunked

"Multiple H1 tags will hurt your rankings."

Not really. Not alone. Google has confirmed that multiple H1s on a page are "completely fine."

One H1 is still a best practice for clarity. What actually matters is a clean heading structure that reflects your content hierarchy. Not the count of any single tag.

Still, I'd avoid it. Google is only one system. There are plenty of others, and other reasons to stick with one H1.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

AI crawlers make way fewer requests than Googlebot. If your content isn't reachable in a shallow crawl, AI models will never find it.

That page buried five clicks deep from your homepage? GPTBot won't get there.

Keep your important content within 2–3 clicks from the homepage. Flat architecture wins for AI visibility.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Everyone talks about crawl budget. Nobody talks about "render budget".

Google might fetch all your JS-heavy pages but only render a fraction of them. Rendering delays can stretch to days on large sites.

Cut JavaScript dependency on your most important pages. Less rendering work for Google means faster indexing for you.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 #TechSEO Myth Debunked

"Longer content always ranks better."

Length doesn't determine rankings. Relevance and depth do.

A 500-word page that nails the answer will outrank a 3,000-word article that buries it in paragraph 15.

Write what the topic needs. If 800 words covers it, adding 2,000 more just makes it worse.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your site probably serves both www and non-www versions. And HTTP and HTTPS versions.

That's four versions of every URL. Before you even count trailing slashes or uppercase letters.

Redirect all versions to one canonical URL. One URL per page. No exceptions.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your sitemap has lastmod dates. Most sites set every page to today. Google sees that and stops trusting your dates entirely.

Only update lastmod when content actually changes.

Accurate dates tell Google what to recrawl first. Fake dates tell Google to ignore you.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Soft 404s are pages returning a 200 status while showing "Page not found" or empty content.

Google flags them in Search Console. Most people ignore the report.

These pages waste crawl budget and signal low quality. Fix the status codes, redirect them, or improve them overall.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your Lighthouse score is not what Google uses to rank you. Google uses real user data from CrUX – actual visitors on actual devices.

Sites scoring 95 in lab tools can completely fail Core Web Vitals in the field. Their real users are on slow connections and cheap phones.

Stop obsessing over Lighthouse. Open Search Console → Core Web Vitals. That's the number that matters.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

Do you know the difference between being found by AI and being known by it? You can find the explanation in my recent newsletter ⤵️

https://t.co/o9dn6BvzCE

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your XML sitemap file size matters. If it exceeds 50MB uncompressed or 50,000 URLs, Google won't process it. Full stop.

Split large sitemaps into smaller files and reference them from a sitemap index.

Here's the part people miss. The 50MB limit is on the uncompressed size. You can gzip a 60MB sitemap down to 5MB all you want. It still violates the limit.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your image alt text isn't just for accessibility. Google uses it to understand image content. AI crawlers rely on it even more because they process text, not pixels.

A product image with `alt="IMG_3847.jpg"` tells AI models nothing about your product.

Write descriptive, specific alt text. It helps Google Image Search and AI content understanding. Think of it as writing for machines, not just screen readers.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 #TechSEO Myth Debunked

I heard, "Submitting your sitemap makes Google index your pages faster."

A sitemap helps Google discover URLs. That's it. It doesn't guarantee indexing or speed it up. If Google doesn't think your content is worth indexing, submitting the sitemap ten times changes nothing.

Sitemaps are a discovery tool, not an indexing request. Fix your content quality and internal linking first.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 #TechSEO Myth Debunked

I heard - "Google only crawls your site once a day."

Not even close. A high-authority news site gets crawled thousands of times a day. A small business blog? Maybe few times a week.

It depends entirely on your site. Search Console → Settings → Crawl stats. That's the real number.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

GPTBot, ClaudeBot, and PerplexityBot all respect robots.txt. Block them and your content never appears in AI answers.

Worth flagging that Perplexity's compliance has been disputed – Cloudflare found evidence they used undeclared crawlers to bypass robots.txt.

I see site owners block these crawlers without realizing they killed an entire traffic channel.

Check your robots.txt right now. Look for blanket `Disallow: /` rules targeting AI user agents.

You might be invisible in AI search and not even know it.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Structured data isn't just for Google anymore. AI crawlers eat it up too. https://t.co/kp2cSMVqsA markup gives machines exactly the context they need without guessing.

Product details, FAQ answers, how-to steps, organization info. All machine-readable.

Make your data are explicit and both Google and AI will understand and cite you.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your CDN edge locations matter more than you think. If most of your users are in Europe but your CDN mainly has US edges, you're barely getting any performance benefit.

Compare your CDN provider's network map against your GA geography report. If they don't match, your CDN is basically doing nothing for you.

[Martin Stepanek 🏳️‍🌈 @techseovitals]

🟣 Underrated #TechSEO Tip

Your Google ranking and your AI citation are two completely different games. I see sites ranking top 3 on Google for competitive keywords that never get mentioned in ChatGPT or Perplexity.

An Ahrefs study found only 12% of URLs cited by ChatGPT, Gemini, and Copilot also rank in Google's top 10 for the same query. Perplexity overlaps more at around 33% – still low.

Google's AI Overviews pull heavily from top-ranking pages. Standalone LLMs run on entirely different source selection logic.