BlogWebsites for tradespeople
Heating-firm website speed: median of 53 sites, eight before and after
The typical website of a heating firm in Bratislava scored 66 points out of 100 on a phone in October 2026 and showed its main content after 6.3 s. Eight other sites we rebuilt as previews went from a median of 2.9 s to 0.5 s.
AuthorMatúš KališFounder
"A fast website" can be said without a number. So here are two measurements we made ourselves, and for each we say what was measured and what was not.
Two measurements, two methods
| Item | Median of 53 sites | Eight sites before and after |
|---|---|---|
| What was measured | The websites of 53 heating firms in Bratislava | The home page of the current site and the home page of our preview |
| Tool | Lighthouse 13.5, mobile profile, default throttling | Chromium through Playwright, emulated iPhone at 393 px |
| Network and CPU | Lighthouse default throttling | 4G at 1.6 Mb/s, 150 ms, CPU 4 times slower, cache off |
| When | October 2026 | 20 October to 5 November 2026 |
| Repeats | Not recorded per site, we report the median of 53 | Current site once, new site the median of three runs |
| What was recorded | Performance score and main-content time | Times, data transferred, requests, third-party servers, tap targets |
The figures in the first column and the second are not comparable point for point. They use a different tool, different throttling and different firms.
What the median of 53 says
The median is 66 points out of 100 and 6.3 s until the main content shows. Our preview for a heating contractor, Teplo Váh, had a median of 99 points and 2.1 s in three runs on its live address. Google recommends that the main content shows within 2.5 s for most real visitors. Our figure is a lab figure, not from real visitors, so we compare it with that threshold only as a rough guide.
We do not name individual firms or list them. So this measurement says nothing about a specific website, only about the typical one.
Eight sites before and after
Eight firms for which we built a preview of a new website. They are not all heating firms. The table shows the median of the eight values for each indicator, calculated from the figures on the Measurements page, where you can also see the individual firms.
| Indicator | Current sites (median) | New sites (median) |
|---|---|---|
| Main-content time (LCP) | 2.9 s | 0.5 s |
| First-content time (FCP) | 1.8 s | 0.4 s |
| Data transferred | 1,704 kB | 216 kB |
| Requests | 46.5 | 14 |
| Third-party servers | 12.5 | 0 |
The spread is wide: the fastest current site took 0.78 s, the slowest 51.64 s. So the median describes the middle, not the range.
In total, the eight current sites transferred 21,135 kB in 363 requests, the new ones 1,966 kB in 100 requests. Of 216 tap targets on the current sites, 92 were smaller than 44 px. On the new ones, 5 of 315.
What this does not show
- It does not show that a faster website brings more customers. We did not measure customers.
- It does not show which thing slowed which site. One site had 50 third-party servers and showed its main content after 0.78 s, another had none and needed 2.46 s. The number of third-party servers does not explain the time on its own.
- These are not visitor data. We have no Chrome UX Report data, so we do not measure INP.
- The eight sites are not a random sample. They are firms we built a preview for.
What to take from it for your own site
Three checks take five minutes: the size of your images, the number of third-party scripts, and whether you can hit a button with a finger on a phone without zooming. The PageSpeed Insights steps are in the article How fast should a website be on a phone. What we do and what it costs is on the Prices page, and you can ask us through the contact page.
Sources
We prepared the text with the help of AI; the figures are from our own measurements.
Related pagesMeasurementsPricesContact