The average Slovak small-business website takes almost 10 seconds to load its main content on mobile. We measured 51 real sites: median LCP 9.9 seconds, and 94% of them sit in the band Google calls poor. A visitor on a phone does not wait that long. They leave before they see anything. Below are the exact numbers, why the sites are slow, and how to check and fix yours.
I spent an evening opening 51 Slovak small-business websites on a phone and counting the seconds until they loaded. The average stopped me: almost ten. I am Tair, I run DevNova in Bratislava, and that one number says more about Slovak small business than it first appears.
Straight up: at DevNova we build and speed up websites, so I have a stake in this. That is exactly why there is no scare-mongering here, only the raw numbers from the measurement and concrete steps you can take yourself, for free.
The hype: the site looks fine, why worry about speed
Most owners look at their site on a laptop, see nice photos, and call it done: it works, it looks good, leave it. But the customer does not see what you see in the office on fast wifi. They see what loads on a four-year-old phone in a shop on mobile data. And that is often a different site: the same design, but two or three ad-break-length pauses before anything appears.
The owner does not judge the site on a laptop. The customer judges it on a phone, and the phone counts seconds.
The reality: what we measured across 51 sites
So this would not be a feeling, we took 51 live Slovak small-business sites and ran each through Google PageSpeed Insights on mobile. Here is what came out:
| What we measured (mobile, n=51) | Result | Verdict |
|---|---|---|
| LCP, main content load | median 9.9 s · 94% poor · 0% good | very slow |
| Performance (Lighthouse) | median 60 / 100 | below par |
| Missing structured data (schema) | 55% | Google/AI cannot read it |
| Missing main heading (h1) | 33% | a third of sites |
| CLS, visual stability | 80% good | this they handle |
Method: 51 publicly available live sites, Google PageSpeed Insights (mobile, Lighthouse lab) + HTML audit, June 2026. Convenience sample, NOT representative of the whole Slovak web; it shows a pattern, not the country's exact average. Data free to cite (CC-BY): State of SK Web Speed.
What does 9.9 seconds mean? Google treats an LCP up to 2.5 s as good. Almost ten is four times the line. And a simple thing holds, which we see constantly: the longer a site takes, the more people leave before they see anything at all. On a phone, patience is measured in seconds, not minutes, and the first three decide.
Why are they so slow? Plainly: not on purpose. From the same thing we find almost every time, an old WordPress or builder with dozens of plugins, full-size images with no compression, no caching, heavy widgets (maps, chats, sliders). A site is not built slow deliberately. It grows in layers nobody turned off for the owner.
A slow site is not about pretty design. It is about layers under the hood that nobody turned off for you.
What do those extra seven seconds cost? A brutally simple rule holds: after the first two or three seconds, some people just close the tab. Not because the site is ugly, they never saw it. When buying something or looking for a dinner spot, the decision happens in seconds, and a slow site loses before it gets a chance. And that lost customer never even shows up in your analytics, because they did not wait for the page to load.
And a second, newer tax has appeared. Over half of those sites had no structured data. That is the invisible description that lets Google, and increasingly assistants like ChatGPT and Perplexity, understand that you are a business, where you are, and what you offer. Without it, both the algorithm and the language model read you worse and recommend you less. In an age when people ask AI for a recommendation directly, being invisible to machines is as expensive as being slow for humans.
Picture a restaurant in Petrzalka. A hungry person at six in the evening looks on their phone for somewhere to eat and finds three sites. Two load in under two seconds and show the menu and a reservation. The third, the pretty one that takes ten seconds, they never see: they already chose from the first two. The difference between first and third place in that moment is not design. It is speed, and unlike taste, speed can be measured and fixed.
A real example: from 11 seconds to 1.5
Concretely. One site we rebuilt scored Performance 56/100 on mobile and loaded its main content in 11.4 seconds, right in the band most of those 51 sit in. We rebuilt it from scratch in custom code (web development): 1.5 seconds and 99/100, and the SEO score rose from 91 to 100. The full case is on the project page; the wider context on AI builders is in the article on AI websites and SEO.
It was not magic or a prettier design. We measured what exactly pulled those 11 seconds, threw out the bloat, compressed the images, turned on caching, and deferred the scripts that do not need to run immediately. That is the whole difference between a site that looks finished and a site that works.
This is not a one-off. Most of those 51 sites would need a similar rebuild, not because they were badly designed, but because after launch nobody watched what piled up under the hood. A website is like a car: you buy it fast, but after years without a service it slows down. The difference is that with a website you do not hear it, until the customers leave.
And we see the same pattern across industries. A bakery, an auto service, a physiotherapist, a small e-shop: a 2019 site, bundled plugins, photos straight off the camera at four megabytes, and a phone that chokes. That is exactly why the fix is predictable: we roughly know what we will find, and we know what to do about it. It is not guessing, it is a procedure.
What to do: how to check your own site
Here is what I would do in your place, free, today:
- Open PageSpeed Insights on mobile and enter your site. LCP over 2.5 s = weak Core Web Vitals, and it is holding you back.
- Look at what loads. Lots of plugins, unoptimised images, heavy widgets: that is your tax for a quick build long ago.
- Check the basics via the Rich Results Test: do you have structured data and a main heading? Half of sites do not.
If the site is slow, small patches will not change it, real numbers move only with a rebuild on clean code. That is why we made a fast entry format: Speed Rescue, a measurement, speed fixes, structured data and a contact funnel, with a before/after report at a fixed price. See the pricing or just send your site for a check. For an e-shop, speed decides revenue twice over, see custom e-commerce.
What usually does NOT help: switching hosting (the problem is what loads, not where it sits), adding one cache plugin to a site full of other plugins, or manually shrinking a few images. Those steps shave off tenths of a second, not seconds. When slowness is built into the architecture of the site, cosmetics will not beat it, which is why genuinely slow sites are worth rebuilding, not patching.
The good news is that speed is one of the few website problems you can measure with a number and fix to a number. It is not a matter of taste or opinion: LCP is either under 2.5 s or it is not. And when it is, the visitor, Google, and your inquiries all feel it at once.
Who pays most for a slow site
Anyone who expects customers from the site, not just a business card. It hits hardest: e-shops (every extra second = fewer orders), restaurants and cafes with reservations, clinics, dentists and salons with online bookings, real estate agencies, and trades and local services that depend on Google showing them on the map. If your site is meant to bring inquiries and calls, 9.9 seconds is a bill you pay in lost customers, you just do not see it on an invoice.
If a site only needs to exist, slowness does not matter. If it must bring inquiries, it is a bill you pay quietly.
A simple test: if your site vanished tomorrow, would you lose customers? If yes, its speed is not cosmetics, it is a business tool that either sells or gets in the way.
Frequently asked questions
Is 9.9 s really a problem if the site works? Yes. Google treats over 2.5 s as poor, and on a phone most people do not wait that long. Works and loads fast are two different things.
Can it be sped up without a rebuild? Partly, compressing images and disabling unused plugins helps by a few tenths of a second. The big numbers (from 10 s to 1.5) only change with a rebuild on clean code.
How much does fixing speed cost? It depends on the state of the site; the entry-level Speed Rescue is a fixed price, with an exact quote after the check in the brief.
Will I lose content or Google rankings? No, we move the content over and usually improve speed and structure.
What is LCP and why 2.5 s? The time until the main content shows for a visitor. Google treats up to 2.5 s as good. Our median was 9.9 s.
Are these 51 sites representative of Slovakia? No, it is a convenience sample, not a study of the whole market. It shows a pattern, not the country's exact average.
Where is the data? Free (CC-BY) on the dataset page, including a CSV download.
About the author
Tair Khamitov, founder of DevNova, a web studio in Bratislava that builds fast websites and automations for small and mid-sized companies. This piece grew out of measuring 51 real Slovak business websites.
Suspect your site is slow? Send it for a check and we will tell you straight what is holding it back and whether a fix is enough or a rebuild is worth it. No strings.
External sources
Related: why AI websites score high on SEO yet Google cannot see them.