01A guide for buyers

Ten minutes that tell you more about a web supplier than an hour in a meeting

You have three quotes on the table, all priced differently, and no way to compare them. Yet there is one piece of evidence you can check yourself, for free and without the supplier: their own website. This test works on us too, which is exactly why it is here.

Nothing is sent anywhere and we ask you for nothing. The scorecard is printable.

Price hides what it leaves out

Two quotes with the same number can mean entirely different work. The cheaper one often moves copy, photos and hosting onto your side, and that never shows up in a comparison table.

A portfolio shows someone else's compromises

On someone else's project you cannot tell what the client decided, what the budget decided and what the supplier decided. Portfolio shots also age better than the sites themselves. Open them live.

Their own site is their ceiling

They set the brief, the budget, the deadline and the taste themselves. Nobody forced four tracking scripts or a background video on them. What you see is the maximum, not a compromise.

And it shows their order of priorities

A site nobody has touched for two years does not prove incompetence. It proves that their own work goes last. After launch, your project joins the same queue.

You would not sit down in a dentist's chair if the dentist had bad teeth, and you need no medical training for that
You need no web training for this test either

Is a threshold of 90 too strict?

We measured 386 company websites in the Zlín region with one tool in a single week. The median mobile performance came out at 61 and 92 % of them sit below 90. The threshold is not strict, almost nobody just meets it.

See the survey results
  1. 01Open pagespeed.web.dev and paste the address of the supplier's home page.
  2. 02Switch to the Mobile tab. Almost everyone looks fine on desktop, but most people arrive on a phone.
  3. 03Run it three times and take the median. The test fluctuates by a few points and a single run proves nothing.
  4. 04Measure the address the page actually ends on. If the site redirects (from www, to a language, to a trailing slash), Google counts that hop as waiting and the score lands a few points below what a visitor experiences.
  5. 05Look at all four categories, not just performance.
CategoryWhat it actually measuresWhat you want to see
PerformanceHow long someone on a phone waits before they see content, and whether buttons jump under their thumb90 or above
AccessibilityWhether the site works for someone who cannot see detail, does not use a mouse or has enlarged their text90 or above
Best practicesSecurity, errors the browser reports and technology that should have retired years ago95 or above
SEOThe technical basics without which Google cannot present the site properly100

Thresholds for a supplier's own site, not for your future e-shop

Since 28 June 2025 accessibility is a duty for some companies, not a bonus. It does not apply to everyone, but whoever builds your website should be able to tell you whether it applies to you.

What those numbers mean and why a lab test differs from reality
  1. 01On a desktop press Ctrl and U (on a Mac Cmd, Option and U). The page source opens.
  2. 02Press Ctrl and F and search it for the words in the table below.
  3. 03On a phone you can do the same by typing view-source: in front of the address.

It is the combination that decides, not the name of the system. WordPress with green numbers and four plugins is solid work. WordPress scoring 38 with twenty plugins and a full screen banner describes what you are about to get.

If the source containsThe site probably runs onWhat to ask about
wp-content, wp-includes, and /wp-admin opens a login screenWordPressThe most widespread system in the world, and not a verdict in itself. Ask how many plugins there are, who keeps them updated for years, and what happens when two of them clash.
elementor, wpbakery, divi, avadaWordPress with a page builderThe design is assembled by mouse on top of a template. It appears quickly and holds up badly: every add on brings more scripts and the page gets heavier.
myshoptet.com, shoptetA rented e-shop platformFine up to a certain size, then it starts limiting you and the monthly fees grow with every add on.
wixstatic, parastorageWixThe site is tied to the platform and cannot be moved elsewhere. Ask what you would take with you if you left.
webnode, squarespace, webflow, framerA website builder or no-code toolA legitimate choice for a simple site. For a company that needs to connect to stock or an internal system it is a ceiling you hit sooner than you expect.
none of the above and the source looks leanA custom build or a modern frameworkAsk what it is and why that one. You do not care about the name, you care whether the supplier understands it and whether someone else could take it over.

What to search for in the source and what follows from it

Why we do not build company websites on WordPress
One failed deal breaker is enough on its own. For the rest: three or more failures from a supplier who is otherwise well priced means you will pay that price back in work on your side.

Only the list of checks prints, without backgrounds or buttons.

Once you have a shortlist, the second half begins: the questions a quote never answers.

Twenty questions for a web supplier, each with a good answer and a warning sign

The test is not a verdict, it is a list of questions you want answered without evasion. A supplier who knows their own numbers and can explain them is a better bet than one who has never looked.

Measure anyone, including us

Paste the address of the supplier you are considering. Once you have the result, one button measures us too, so you have them side by side.

And this you can verify right here

The values below were measured by your own browser on this page, just now. They are not numbers typed into the text.

Server response
measuring
Weight of the page
measuring
Files downloaded
measuring
Third party scripts
measuring
What these numbers mean

Measured directly in your browser, nothing is sent anywhere. The average page on the web weighs around 2.5 MB today, roughly eight times more. Zero third party scripts means no tracking, no cookies, and therefore no consent banner.

Google marks anything from 90 points up as green, and no visitor can tell 96 from 100. A site below 90 deserves attention, below 80 the speed puts people off, and below 75 the company measurably loses enquiries.

Verify with Google

The same threshold is in our contract

A score above 90 on mobile and desktop is part of our terms and conditions. If a delivered site does not reach it, we keep tuning at our own cost until it does. It is not a promise made in a meeting, it is a clause in the contract we sign.

Read that clause in the terms and conditions

And where we come off worse

  • There are two of us. We handle a limited number of projects at once, so sometimes we will offer a later date than an agency of twenty.
  • We have no night shift. We have monitoring that wakes us up, but we do not run shifts.
  • Our site is a presentation, not an e-shop with fourteen integrations. That makes our job easier than that of someone who builds e-shops, and it is fair to say so. Which is precisely why the whole test runs on the one thing every supplier has in common: their own presentation site.
How do I recognise a good web supplier?
By their own website, not by their portfolio. Measure their home page in PageSpeed Insights on mobile, walk through it on a phone, check the cookie banner and the legal details in the footer, and send them an enquiry through their own form. Their own site is the one project where they set the brief, the budget and the deadline themselves, so it is their ceiling rather than a compromise with a client.
What PageSpeed score should a web supplier have?
On mobile, 90 or above for performance and accessibility, 95 or above for best practices and 100 for SEO. Measure the home page three times and take the median, because the test fluctuates. Below 70 is a reason to walk away: on their own presentation site none of the excuses that hold for a client project apply.
Is a cheaper quote always the worse choice?
No. The difference in price is usually not in quality but in what the price leaves out: copy, photos, integrations, hosting and the hours someone in your company will spend on it. Compare quotes over three years of running rather than the launch day price.
What if a supplier comes recommended but their site scores badly?
A recommendation is a strong signal and should not be thrown away. Ask about what you found and listen for whether they already know. „We know, here is the reason and here is the date“ is fine. „Customers do not look at that“ tells you how your site will look in two years.
How do I tell what a website is built on?
Open the page source with Ctrl and U and search it for wp-content (WordPress), myshoptet (a rented e-shop platform), wixstatic (Wix) or elementor (a page builder on top of WordPress). With WordPress you can also add /wp-admin to the address and a login screen appears. The system alone is not a verdict; what matters is how it combines with the measured numbers and the number of plugins.
How many quotes should I ask for?
Three is enough. Beyond that the decision collapses into comparing prices, because nobody remembers the rest. Ask for them against the same brief, otherwise there is nothing to compare, and run this test on all three first. It saves you meetings that would lead nowhere.

If someone else comes out better, go with them

This guide is not here so that you pick us. It is here so that you do not pick blind. If our numbers and the way we work suit you, get in touch and we will go through your project. If not, at least you know what to ask elsewhere.