Trade Website Speed Study: We Tested 556 UK Sites

• On a mobile phone, the typical trade website in our study takes 6.7 seconds to show its main content. Google counts 2.5 seconds or less as good, and over 4 seconds as poor.
• 72% of the sites we speed-tested (350 of 489) were over 4 seconds. Only 9% (46 of 489) passed all three of the speed and stability checks we used.
• Pages mostly stay put as they load (88% passed the layout stability check). Loading speed is what lets them down.
• 41% of homepages (193 of 473) had no tap-to-call link, which matters when the phone call is the sale.
• This is lab data from one test per site, on a sample that is not random. The full method, and what the numbers cannot tell you, is below.

By Sam and Harriet at Brambla. Data collected by RnkRocket, our sister SEO tool, in September 2026.

On a mobile phone, the typical trade website in our study takes 6.7 seconds to show its main content. If you run a trade business and your website feels a bit slow, you are in very good company. You may also be losing some of the people who search for you.

We looked at 556 websites belonging to electricians, plumbers, roofers, builders, joiners, painters and similar trades, spread across Devon, Cornwall and 27 towns and cities. We tested the homepage of each one, on a simulated mobile phone. Below is what we found, what it means, and what you can do about it. The method is near the bottom, including what these numbers can and cannot tell you.

The speed figures cover the 489 sites where Google's PageSpeed Insights returned a result. The page-content figures cover the 473 homepages we could read, and the HTTPS figure covers the 495 homepages that answered at all.

1. The typical site takes 6.7 seconds to show its main content

The median Largest Contentful Paint (LCP) was 6.7 seconds across 489 sites. LCP is the moment the biggest thing on the screen, usually your main photo or headline, has appeared. Google treats up to 2.5 seconds as good and over 4 seconds as poor. 72% of the sites we tested (350 of 489) were over 4 seconds. Only 11% (55 of 489) were within 2.5 seconds.

What it means for you: someone searching for an emergency plumber on their phone, probably in a hurry, is waiting on a page that has not finished showing up. Some will go back and tap the next result.

2. The median PageSpeed Insights score was 67 out of 100

Google's PageSpeed Insights gives each page a mobile performance score out of 100. The median in our study was 67 (489 sites). 15% scored 90 or more (73 of 489), and 10% scored under 50 (51 of 489).

What it means for you: a 67 sits in the "needs improvement" band. It is not a disaster, but it is a long way from fast, and the layout stability figures below suggest the gap is mostly about how long the page takes to show up, not about it jumping around.

3. Pages stay put, but they load slowly

This is the good news. 88% of the sites (432 of 489) passed Google's layout stability check (CLS of 0.1 or under), so buttons and text mostly do not jump about as the page loads. 74% (360 of 489) also passed the blocking-time check (TBT of 200 ms or under), which is a lab measure of how quickly a page responds once it has loaded. So the weak spot is loading speed. We did not test the cause, but heavy images and scripts are common reasons for it.

What it means for you: a slow site is not always a broken site. Often it can be put right without starting again.

4. Four in ten homepages have no tap-to-call link

41% of the homepages we could read (193 of 473) had no tap-to-call link. If a visitor on a phone can see your number but has to remember it, switch apps and type it in, some will not bother. Our check looks at the page's HTML, so a call button added by JavaScript might be missed. We think this is a small effect, but we cannot rule it out.

What it means for you: for a trade business, the phone call is the sale. A tap-to-call button is one of the cheapest improvements you can make.

5. One in five has no meta description, and more than a third have no structured data

21% of homepages (100 of 473) had no meta description, the short summary that often appears under your name in Google. 36% (170 of 473) had no structured data (JSON-LD), the behind-the-scenes labelling that tells Google what kind of business you are and where you work. 23% (111 of 473) had no H1 main heading.

What it means for you: none of these will fix a slow site, but they are the basics of being understood by Google, and they are quick to put right.

A few things that were fine

Almost all sites used HTTPS: only 5% (25 of 495) did not. And only 7% of the homepages we read (35 of 473) were missing the viewport tag that tells a phone how to display the page.

Devon and Cornwall are the largest single part of our sample: 298 of the 556 sites (54%), of which 270 were speed-tested and 259 had a readable homepage. We compared them with the sites outside the South West. Bristol and Bath (28 sites) are left out of both columns.

On speed, Devon and Cornwall sit close to the rest of the sample. The gap is small, the two parts of the sample were built in different ways (see the method below), and a single lab run per site is a noisy measure. So we are not claiming that Devon or Cornwall sites are slower.

The clearest difference was the phone button: 48% of Devon and Cornwall homepages (124 of 259) had no tap-to-call link, against 33% (64 of 193) elsewhere. 28% (73 of 259) had no H1 heading, against 16% (31 of 193). Because of how the two lists were put together, treat even these as a pointer rather than proof.

We cannot split Devon from Cornwall. The county was recorded for only a small part of the list, so we show the two together.

You do not need to become a web developer. A few things in rough order of effort:

1. Check your own site on a phone, on mobile data, not on your office wifi. Open it, start a timer, and see how long until you can read your main heading. If it is more than three or four seconds, it is worth fixing.
2. Run your homepage through Google's free PageSpeed Insights tool and look at the mobile tab. The "Largest Contentful Paint" number is the one to watch. It is a single simulated test, so run it a couple of times.
3. Shrink your photos. Large, uncompressed images are a common reason a trade site is slow. Ask whoever looks after your site to resize and compress them.
4. Add a tap-to-call button to the top of your homepage on mobile.
5. Write a one-line meta description for your homepage: who you are, what you do and where.
6. Remove what you do not use. Old plugins, sliders and chat widgets all add weight.

If you would like a second pair of eyes, our free Website Audit looks at speed, mobile set-up and the basics above and tells you plainly what is worth fixing. And if you decide your current site is not worth patching, our 7 Day Website is a fast, clean build for a small business. No pressure either way.

Who ran it. The scan was run by RnkRocket, a UK SEO tool built by the same person who owns Brambla. We are not a neutral third party, and we are in the business of building websites, so we want to be clear about how this was done. Homepage checks were run on 25 September 2026 and speed tests on 26 September 2026.

How sites were sampled. The sample has two parts. The first is a list of 298 Devon and Cornwall trade websites from Brambla's own research list of local businesses. We checked that this list was not filtered for weak sites: it was built from Google Maps results for each trade and town, keeping every business that listed a website. The second part is 258 sites found through Google Maps local search in 27 UK towns and cities: electricians and plumbers in all of them, plus roofers, builders, painters and decorators, and joiners in the smaller places. Searches covered a 15 km radius of each city centre, and we kept at most ten sites per city per trade. Only businesses that list a website were included. This is not a random sample of every UK trade business, and it is weighted towards Devon and Cornwall (54% of the sites). We call it a study of "UK trade websites" because that is where the businesses are, not because it is a statistical picture of the whole country.

What was excluded. From 622 listings, we removed 17 sites on free website-builder addresses, 13 from a network of template landing pages, 2 booking-portal links and 32 near-identical "emergency plumber" sites that read as one network (the same copy with only the town and brand swapped). That left 558 listings, which are 556 distinct websites (two sites are listed under two trades each). The Devon and Cornwall list had already left out chains, merchants, directories and the websites of our own clients. Otherwise we did not choose sites by how good or bad they are.

What we measured. For each site, we fetched the homepage once, politely, obeying its robots.txt file, and read its HTML (no JavaScript was run). We then ran Google's PageSpeed Insights once on the homepage in mobile mode. Speed figures are based on the 489 sites where PageSpeed Insights returned a result. Sites that did not load, blocked our scanner through robots.txt or could not be scored are not in the speed figures. Page-content figures (meta descriptions, headings, structured data, phone links) are based on the 473 homepages we could read. HTTPS figures are based on the 495 homepages that answered at all.

What PageSpeed Insights lab data is and is not. It is a simulated test of a mobile phone on a slowish connection, run by Google's Lighthouse tool. It is lab data, not real-user data: it shows what a typical visitor on a typical mid-range phone might experience, not what your actual visitors did. We ran it once per site, and a single run can vary from one test to the next, so one site's result should not be treated as exact. Across hundreds of sites those differences largely even out, but they do not vanish.

The thresholds. Google treats a Largest Contentful Paint of 2.5 seconds or less as good and over 4 seconds as poor, and a Cumulative Layout Shift of 0.1 or less as good. A site "passed all three" if, in that one lab run, its Largest Contentful Paint was 2.5 seconds or less, its Cumulative Layout Shift was 0.1 or less, and its Total Blocking Time was 200 milliseconds or less. Total Blocking Time is the lab stand-in for Interaction to Next Paint (INP), which can only be measured with real visitors. This is our simplified pass mark, not Google's official Core Web Vitals assessment, which uses real-user data that most small trade sites do not have.

How many websites did you test, and which trades?

We looked at 556 websites. Speed tests came back for 489 of them, and we could read the page content on 473. They belong to electricians, plumbers, roofers, builders, joiners, painters and decorators and similar trades.

Is a 6.7 second load time really that bad?

Google's guidance is that the largest part of a page should appear within 2.5 seconds, and over 4 seconds is poor. At a median of 6.7 seconds, the typical site we tested is well into "poor". It is a lab test on a simulated phone, so real visitors on faster connections may see better times, but the direction is clear.

Why did you publish this when you build websites?

Because the numbers are useful to a small business owner whoever builds the site, and because we would rather show how we got them than ask you to trust us. The method is above, and we have not named or shamed any business.

Can I check my own website?

Yes. Google's PageSpeed Insights is free to use. If you would like us to look at it, our free Website Audit covers speed, mobile set-up and the basics, with no obligation.