Accessibility: what we measured
This page does not claim that this site is accessible. It tells you what we measured, what came out of it, what we do not know yet and what still has to be fixed. With the figures, so you can check it.
- DOC.TG1zWhy this is not a translationframe
- DOC.TG1aWhat we are working towardsaim
- DOC.TG1bHow we measured thismeasurement
- DOC.TG1cWhat we foundresult
- DOC.TG1dWhat has to happenopen
- DOC.TG1eCannot use something?reporting
- DOC.TG1fWhat this statement coversscope
# The order of this text, with the code that stands above each part.
# The pointer walks the parts; it does not show where you are reading.
Why this is not a translation of the Dutch statement
So this page does what is honest in every one of those countries: it says what we measured, on what date, with what, what came out, and what we have not looked at. No claim of conformity, no badge, no statement that we meet a legal duty. If you need a formal statement for a procurement process, tell us which form you need and we will say plainly whether we can produce it.
Exactly which form of accessibility statement fits a private provider in the United Kingdom, and what the European Accessibility Act means for us in Ireland, is a question for a lawyer and has not been settled. It is named here rather than answered with a Dutch form wearing an English label.
What we are working towards
We aim at WCAG 2.1 level AA. That is the international standard for accessible websites and we hold to it as a guideline, including where no specific legal obligation applies to us. In practice that means we watch for keyboard only operation, visible focus, sufficient contrast, text you can enlarge, clear headings and labels, alternative text on images, understandable error messages, respect for the reduced motion setting, and accessible sign-up and payment steps.
What deliberately does not appear here: that this site conforms to WCAG 2.1 AA. That statement requires a full test of all success criteria, by hand, with a screen reader and preferably by somebody from outside. That test has not been done. While that is the case, we call no part of this site conformant, and we do not use the word accessible as a badge.
What we could do, we did. Below is exactly what that was.
How we measured this
The measurement was carried out on 12 August 2026 on the new version of this site, on a desktop browser based on Chromium. It consisted of three things.
- A calculation of the contrast ratios. We read the actually rendered text colours out of the browser and set them against the background using the calculation rule from WCAG 2.1. The background of this site is a gradient, so we calculated both its darkest and its lightest point and wrote down the whole range.
- A scan of every page on this site for declared language, alt attributes on images, blocked zoom, heading structure and the link between form fields and their label.
- A manual keyboard check on one page, tabbing through the page and establishing which element took focus, whether that was visible and whether the menu behaves as it should.
Just as important is what we did not do. We did not test with a screen reader, not on a phone or with touch operation, not with voice control, not for real readability at 200 per cent and 400 per cent magnification, and not with the reduced motion system setting actually switched on. We also did not judge whether the alternative texts that are there are any good in substance, only whether they exist. And there has been no audit by an outside party.
Note: this is a sample, not an audit. A sample can show that something is wrong, but not that everything is right. Read the results in that light.
The measurement of 12 August 2026 was run over the site as it stood on that date. The English pages have grown since then and have not been separately measured. So do not read the figures below as a measurement of the page you are looking at now: read them as the state of the build these pages come out of. A new measurement, covering all three languages, is what the next version of this page will be based on.
What we found
Contrast of the text
The WCAG 2.1 AA requirement is 4.5 to 1 for ordinary text and 3 to 1 for large text. Below is what we measured per kind of text. The range comes from the background being a gradient: the top left is slightly lighter than the rest.
| Element | Size | Measured contrast | Reaches 4.5 to 1 |
|---|---|---|---|
| Body text and lists | 16.6 px | 7.1 to 1 up to 7.9 to 1 | yes |
| Headings | 27 px and larger | 18.2 to 1 up to 20.1 to 1 | yes |
| Links in body text | 15.5 px | 6.8 to 1 up to 7.5 to 1 | yes |
| Bold text and table cells | 15.5 px | 15.0 to 1 up to 16.6 to 1 | yes |
| Reference row at the foot of a section | 11.5 px | 7.5 to 1 up to 8.3 to 1 | yes |
| Small code labels, such as DOC.TG1 above and the line in the footer | 11 px | 3.8 to 1 up to 4.2 to 1 | no |
The body text sits well above the line. The small code labels in the monospace typeface sit below it. In most places those labels carry no information you need in order to understand the page, but in the footer there are company details in that colour. That is a real point and it is in the list of what has to happen below.
Keyboard and visible focus
Focus is visible. Our own stylesheet nowhere sets outline: none and so does not suppress the focus ring. Nor is there a focus style of our own, which means you see your browser's default ring. At the check that was a clearly visible ring around the element with focus. The drawback of having no style of our own is that the appearance differs per browser and that we cannot guarantee it stands out equally well everywhere against this dark background.
The menu behaves as it should. While the index is closed it sits at visibility: hidden, which keeps the links inside it out of the tab order. So you do not tab blindly through them. The button that opens it reports through aria-expanded whether it is open or closed, and Escape closes it again. On the page where we counted, eleven elements were visible and reachable with the tab key.
What is missing is a link that takes you straight to the main content. Anyone working with the keyboard currently has to pass the logo and the menu button on every page.
Alternative text on images
Across all pages together there are 965 images. All 965 have an alt attribute. On 19 of them that attribute is deliberately empty, which is the correct way to let a screen reader skip a purely decorative image.
Note what this does and does not say. It says the attribute is there. It does not say the text in it is useful to somebody who cannot see the image. We have not judged that one by one.
Language and enlarging
Every page declares its language on the html element. The Dutch pages carry lang="nl", these English pages carry lang="en-GB". A screen reader therefore knows which language to pronounce. Individual foreign words or quotations inside a page are not marked separately; we did not check that.
No page blocks zooming. Nowhere is there user-scalable=no or a maximum zoom level in the viewport setting. Whether the text really reflows neatly at 200 per cent magnification is a separate test that has not been done.
Headings
Except on two pages, every page has exactly one h1. Those two have none: the demo environments of the MyParcel Dashboard and of Jarvis. On five pages the heading structure skips a level, for instance from an h1 straight to an h3. For somebody navigating by headings with a screen reader, the structure there is wrong.
Forms
There are 36 input fields on seven pages. Of those, 22 have a label or an aria-label that is genuinely linked to them. The other 14 lean only on the grey example text in the field. That disappears as soon as you start typing and is not always passed on by assistive software as the name of the field. That has to be put right.
Motion
The stylesheet takes the reduced motion setting into account in eight places, including the transition between pages, which is then switched off entirely. We read that in the code and did not measure it with the setting actually on. Every figure on these English pages is built the same way: the continuous motion in it stops when that setting is on.
What has to happen
These are the points that came out of the measurement. They are here because we think an accessibility statement containing only good news is worth nothing.
# No finding is not the same as conformant. There has been no manual test with keyboard and screen reader, so nowhere on this site does it say that it meets WCAG 2.1 AA.
# The dashes count only the points in the table below. They are not a score and not a percentage.
# The bar that keeps travelling across the four columns is the measurement itself: it is never finished.
| Point | Where | What it means for you |
|---|---|---|
| Small code labels do not reach 4.5 to 1 | The annotations at every section and the company line in the footer | That small grey text is hard to read in low light or with less sharp sight |
| No focus style of our own | Site wide | You do see where you are with the keyboard, but how clear that is depends on your browser |
| No link to the main content | Site wide | With the keyboard you have to pass the header on every page first |
| 14 input fields without a linked label | The free tools and the Jarvis sign-up screen | Assistive software does not always say what belongs in the field |
| Two pages without an h1 and five with a heading jump | The demo environments and a few knowledge pages | Navigating by headings gives a messy picture there |
These points deliberately carry no target date. We only put a date next to one once it has been agreed and is achievable. Target date per point: not yet established.
What we do not do here: call a function accessible without a manual check with keyboard and screen reader having gone over it. Known problems per product go into our internal release and accessibility register, with an owner named.
Cannot use something?
Email sales@theseo.nl with the word accessibility in the subject line. Help us with three things:
- the page where it goes wrong, preferably as a full web address;
- which device and which aid you use, for instance the name of your screen reader or the fact that you work with the keyboard only;
- a short description of what you were trying to do and where it stalled.
We confirm your report within two working days and aim to come back within ten working days with a usable solution or a working alternative. If that does not work out, you hear from us what is possible and when. If the point is that you cannot reach a service you are paying for, that goes ahead of everything else.
If you cannot resolve it with us, you are free to take your complaint elsewhere. If it is about your personal data rather than about accessibility, the route for that is in our privacy statement.
What this statement covers
The measurements above cover this website. Our products with a sign-in environment of their own, such as the MyParcel Dashboard and Jarvis, are included only in so far as their public pages are part of this site. The demo environments were scanned and produced two of the points above.
For those products the same starting point applies: no function is described as accessible without a manual check with keyboard and screen reader alongside an automated test. What comes out of that per product goes into the product documentation.
We are not certified, against any standard, and accessibility is not a certified process at TheSEO. We do not name somebody else's standard as backing here either, because naming a standard is not evidence. What does exist is on this page itself: a measurement with a date, the points that came out of it and a version number that moves as soon as something changes.
Does this page change?
Yes. As soon as one of the points above is solved or there is a new measurement, we amend this page and raise the version number. What was here before stays available on request.
See also our privacy statement, our terms and conditions and our AI transparency statement.
Version 1.0 · measurement carried out on 12 August 2026 · English edition published on 25 August 2026