01Survey

How company websites are doing in the Zlín Region

We measured hundreds of company websites in the region with one tool in one week. No names, no league table, and with a methodology you can recalculate. The numbers below come straight from the measurement, not from the text.

8 August 2026

95sites measured
60median mobile performance
78 %sites below 70
94 %sites below 90
85median accessibility
37 %sites on WordPress
54 %sites with an outdated plugin
63 %sites on unsupported PHP
2,731 kBmedian page weight
23 %sites have enough traffic for real user data

How the sites are distributed

A median is one number. This chart shows the point: the sites do not cluster around the average, they pile up at the bottom.

0
0
0
10
0
20
3
30
10
40
28
50
28
60
10
70
4
80
5
90
Mobile performance score in tens along the bottom, number of sites vertically. The green band starts at ninety.
78 %sites below 70

What the sites run on

Detected from what each page loads by itself.

  • WordPress37 %
  • Joomla8 %
  • Webnode3 %
  • ostatní52 %

The bar is low and that is your opening

With 94 % of sites below ninety points and a regional median of 60, you do not need the best website in the country. A site that loads fast on a phone, reads well on a small screen and meets the legal minimum already puts you in the top few percent of the region. Most of those numbers do not move with a bigger ad budget, they move with how the site is built.

What we do about it
CategoryMeasuredMedian performanceBelow 90WordPress
Výroba95608335
Ostatní74616428
Obchod68606319
Služby firmám66636023
Stavebnictví32632912
Veřejný sektor1862184
Informační služby1869148
Doprava95982
Ubytování a gastro45931
Vzdělávání a zdravotnictví26321

What failed most often

The checks that failed most often across the measured sites. Not a matter of taste; each has a yes or no answer.

  • Insufficient text contrast against the background73 % · 69
  • Links that do not say where they lead67 % · 64
  • The page reports errors in the browser47 % · 45
  • Headings are out of order39 % · 37
  • Images without a text description23 % · 22
  • Zoom is blocked on mobile20 % · 19
  • Buttons too close together for touch18 % · 17
  • Buttons with no name for a screen reader15 % · 14
SystemSitesShareMedian performanceBelow 90
No system detected4547 %6038
WordPress3537 %6131
Joomla88 %628
Webnode33 %782
Drupal11 %471
Webflow11 %551
Nuxt11 %331
Next.js11 %471
Wix11 %561

Sites with no detected system are not necessarily hand built. The group holds static pages made years ago, sites built by a tool that leaves no trace, and pages with nothing but a heading and a phone number.

Add-ons on top of WordPress

Page builders and extensions that are not a system of their own. The share is counted from WordPress sites, not from all of them: a site with Elementor runs on WordPress, not on Elementor.

  • WooCommerce43 % · 15
  • Elementor37 % · 13
  • Divi17 % · 6
  • WPBakery14 % · 5

WordPress up close

The most common system in the sample. It is not a problem in itself, the problem is a version nobody updates.

35sites on WordPress
37 %of WordPress sites run an outdated core

Core version

  • WordPress 7.029 % · 10
  • WordPress 6.814 % · 5
  • WordPress 6.66 % · 2
  • WordPress 6.03 % · 1
  • older versions together11 % · 4

7.0.3 latest release at the time of measuring · 13 sites do not state their core version

Plugins

  • At least one plugin behind54 % · 19
  • Plugin untouched for over two years11 % · 4

PHP on the server

The server reports the version itself in the response header. Not reporting it is not a fault: it is a setting that costs nothing.

  • PHP 8.438 % · 3
  • PHP 7.438 % · 3 · no longer receives security fixes
  • PHP 7.325 % · 2 · no longer receives security fixes

87 sites do not state their PHP version, which is fine

How to read it: the large number is always the share of sites that are fine on that value. The bar below shows the whole split and the list under the bar says what each colour means. Amber "with a caveat" is not a fault: the site has the thing, just not the way it should. Red "needs work" is something to fix. The count of sites is in brackets.

Four Google scores

The green band starts at ninety.

Mobile performance

6 %are fine

  • 90 a víc6 %(5)
  • 50 až 8980 %(70)
  • pod 5015 %(13)
  • median 60

Accessibility

29 %are fine

  • 90 a víc29 %(26)
  • 50 až 8971 %(64)
  • median 85

Best practices

74 %are fine

  • 90 a víc74 %(67)
  • 50 až 8924 %(22)
  • pod 501 %(1)
  • median 96

SEO

77 %are fine

  • 90 a víc77 %(69)
  • 50 až 8923 %(21)
  • median 92

What speed is made of

The thresholds are Google's, not ours.

Time to main content (LCP)

3 %are fine

  • do 2,5 s3 %(3)
  • 2,5 až 4 s9 %(8)
  • přes 4 s88 %(77)

Layout shift (CLS)

87 %are fine

  • do 0,187 %(78)
  • 0,1 až 0,253 %(3)
  • přes 0,2510 %(9)

Time the page cannot respond (TBT)

76 %are fine

  • do 200 ms76 %(67)
  • 200 až 600 ms17 %(15)
  • přes 600 ms7 %(6)

Server response (TTFB)

100 %are fine

  • do 0,8 s100 %(90)

Home page weight

13 %are fine

  • do 1 MB13 %(12)
  • 1 až 2,5 MB34 %(31)
  • přes 2,5 MB52 %(47)

Files downloaded

36 %are fine

  • do 5036 %(32)
  • 50 až 10047 %(42)
  • přes 10018 %(16)

Data from real visitors

Only sites with enough traffic.

Google only publishes this data for sites with enough traffic. 22 of the 95 measured sites have it, that is 23 %. For the rest only lab measurement on a simulated phone is available.

LCP for real visitors

68 %are fine

  • do 2,5 s68 %(15)
  • 2,5 až 4 s27 %(6)
  • přes 4 s5 %(1)

Touch response for real visitors

95 %are fine

  • do 200 ms95 %(18)
  • 200 až 500 ms5 %(1)

Hosting and security

Things set once that cost nothing.

Secure connection

86 %are fine

  • fine86 %(82)
  • with a caveat13 %(12)
  • needs work1 %(1)

Enforced https

25 %are fine

  • fine25 %(24)
  • with a caveat75 %(71)

Security headers

14 %are fine

  • fine14 %(13)
  • with a caveat20 %(19)
  • needs work66 %(63)

PHP version

38 %are fine

  • fine38 %(3)
  • needs work63 %(5)

Tracking and privacy

Third party code and cookie consent.

Tracking and advertising code

36 %are fine

  • fine36 %(34)
  • with a caveat62 %(59)
  • needs work2 %(2)

Cookie consent

26 %are fine

  • fine26 %(25)
  • with a caveat55 %(52)
  • needs work19 %(18)

Scripts from other domains

23 %are fine

  • fine23 %(22)
  • with a caveat65 %(62)
  • needs work12 %(11)

Accessibility and controls

Whether the site works for weaker eyesight too.

Declared page language

91 %are fine

  • fine91 %(86)
  • needs work9 %(9)

Zoom on mobile

66 %are fine

  • fine66 %(63)
  • needs work34 %(32)

Images with descriptions

63 %are fine

  • fine63 %(57)
  • with a caveat21 %(19)
  • needs work16 %(15)

Dark mode

1 %are fine

  • fine1 %(1)
  • with a caveat99 %(85)

Legal minimum

What every trader should have on their site.

Privacy policy

71 %are fine

  • fine71 %(65)
  • with a caveat29 %(27)

Business identification

43 %are fine

  • fine43 %(37)
  • with a caveat57 %(49)

Findability and upkeep

How the site shows in search and whether it lives.

Page title

73 %are fine

  • fine73 %(69)
  • with a caveat27 %(26)

Search description

74 %are fine

  • fine74 %(70)
  • needs work26 %(25)

Link preview

34 %are fine

  • fine34 %(32)
  • with a caveat66 %(63)

Sitemap

75 %are fine

  • fine75 %(71)
  • with a caveat25 %(24)

Last change on the site

79 %are fine

  • fine79 %(31)
  • needs work21 %(8)

Year in the footer

60 %are fine

  • fine60 %(33)
  • with a caveat11 %(6)
  • needs work29 %(16)

A high score on its own does not make a good website

The fastest sites in the sample are not well built modern sites. They are fast because there is almost nothing on them: static pages with no system, untouched for years, sometimes just a heading and a phone number. Speed is a condition, not a goal.

185 kBmedian weight of the fast sites
15median number of files on the fast sites
80 %of the fast sites run on no recognised system
3,033 kBmedian weight of the other sites

So read the score together with the other numbers. A site that is fast and also has something to say, works on a phone and meets the legal minimum is a different thing entirely from an empty page scoring ninety.

ExampleWe do not publish the company name or the address and never will.
68Performance
92Accessibility
100Best practices
92SEO
72 %faster than this share of the measured sites in the region
17.9 snež se objeví hlavní obsah
4509 kBváha úvodní stránky

What the site runs on

WordPress, Elementor, WooCommerce · WordPress 7.0.3 (poslední vydaná 7.0.3) · 5 doplňků, z toho 5 pozadu a 0 bez aktualizace přes dva roky

WordPress plugins

DoplněkNa webuPosledníAktualizován
Blocksy Companion2.1.252.1.522026-08-06
Elementor Website Builder – more than just a page builder3.34.34.2.22026-08-06
Essential Addons for Elementor – Popular Elementor Templates & Widgets6.5.96.7.22026-07-29
Ultimate Addons for Elementor2.8.22.9.22026-07-20
WooCommerce10.4.411.0.02026-08-04

What failed in the measurement

  • Insufficient text contrast against the background
This is what a company gets when it asks for its own report. It is a real measurement of a real site from this sample, only without the name, address, town or sector. It is chosen deterministically, so it does not change on every load.

We only send the report to an address on that site's own domain. It is the only way to tell, without registration, that the site is yours, and to stop someone else from requesting your company's result.

The sample came from the regional company directory and from a manual list. For every company we verified in the business register that it exists, and live that the website really belongs to it. Companies listed in the regional directory but registered elsewhere stayed in: they operate here.

Measurement used Google PageSpeed Insights, on mobile, on the final address after redirects. A site that redirects from its domain to a language version would otherwise score worse than what a visitor experiences.

Our own reading takes only what the server serves to an ordinary browser: the home page, the response headers, the sitemap and the stylesheets the page links to. No admin addresses were probed. Before each site we read robots.txt and honoured it.

Parked domains and pages behind bot protection do not count as a company website. An empty domain would otherwise score beautifully on speed and skew the whole summary.

What these numbers do not say

The sample is not a random selection of all companies in the region. These are companies that once joined the regional directory, plus manual additions. Smaller firms and sole traders are under represented.

Lab measurement fluctuates by a few points between runs. That matters for a single site; across hundreds it evens out.

Some checks are uncertain for an individual site. A page assembled in the browser does not show us its footer, so we report that a privacy policy was not found, never that it is missing.

A finding is not an accusation. A low score does not mean the site is broken or that a bad supplier built it. It means there is something to work on.

We own up to our own errors. The first published version of this report carried two detection bugs: the WordPress core version was also read from jQuery, which lives in the same directory, and a plugin version from a library bundled inside the plugin. Sites therefore came out as years out of date while running the latest release. Both are fixed and all 386 sites were re-examined; the Google scores were left untouched so the report stays comparable with itself.

Manual check: we went through twelve randomly picked sites by hand against what their servers actually serve. The detected system, core version, PHP version, tracking code, cookie consent and legal details matched on all twelve.

What to do about it

If the numbers interest you for your own site, read the guide to choosing a supplier: it has a ten minute test you can run yourself and a tool that takes a site apart on the spot.