Scandiary

A real report

This one is not a mockup

The report below came out of the actual scanner, on a real website, on 12 August 2026. Nothing in it was written by hand or tidied up afterwards.

That one is the free plan. Further down there is a second report, from the same website on the paid plan, so you can see what the money changes.

The free-plan report

Whose website is it, and why not a customer’s

It is the W3C’s Before and After Demonstration — a website the people who write the web accessibility standards built deliberately broken, so that tools and students can be tested against it.

We have scanned real shops, and we are not going to publish their names next to a list of their failures. That is somebody else’s house. A demo built to be scanned is the one site where showing the findings costs nobody anything.

What the scan found

  • 5Issue types
  • 44Elements affected
  • 5Pages scanned

The interesting part is not the total. It is where the problems were. All 44 elements sit on a single page — the demo’s deliberately broken before home page. The other four came back with nothing, and two of those are W3C’s own real pages.

That is the check we most wanted to pass. A scanner that reports problems everywhere is not sensitive, it is useless — you would learn to ignore it by the third week. This one found the broken page, and stayed quiet on the rest.

Two of the quiet pages are inside the demo as well. That does not make them accessible — it means an automated check could not prove anything about them. The same is true of any quiet page in your own report.

The part we actually sell

Read what it says before it says any number

Open the report and you will find a section called Since the last scan. In this one it holds a real comparison, against the scan of 10 August 2026. Before a single number, it admits something about us: we changed the browser window we scan with, from 800×600 to 1440×900 pixels, so some of the differences below come from our change and not from the site.

Nobody would have caught that from the outside. A wider window shows navigation and footers a narrow one hides, the counts move, and the honest thing is to say so where you cannot miss it — above the numbers, not in a footnote.

The comparison itself reports five issue types unchanged. A week with nothing in it is a real result, not a failure — and it is also not proof that a site is accessible. That is the thing we sell, and it only exists from the second scan onward.

Here is a second report: a second scan of the same site, on 29 August 2026, run with the paid page limit instead of the free one. It is raw scanner output as well — nothing added, nothing tidied.

Those numbers are bigger because the scan looked at fifteen more pages, not because the website got worse between the two. The report says that itself, above its own comparison: the set of pages scanned changed since last time. We would rather point at it than let you read the jump as a discovery.

The part you are actually paying for is further down, under each issue. There is a list headed Where they are, and it gives the CSS selector of every affected element together with the page it sits on. On the free plan that section does not exist — you get one example element per issue, and the report tells you plainly that the rest is what the paid plan lists.

Three lines out of the twenty-five listed under one issue in that report:

  • #home > a[onfocus="blur();"] https://www.w3.org/WAI/demos/bad/before/home.html
  • #news > a[onfocus="blur();"] https://www.w3.org/WAI/demos/bad/before/home.html
  • a[href$="#s-ATAG20"] https://www.w3.org/Translations/

The same issue, two different pages. That is the whole point of the list: the first two are one template, the third is somewhere else entirely.

It stops somewhere. The largest issue in this report affects 173 elements and the list is headed first 25 of 173. Twenty-five is enough to find the template that produced them; 173 lines of selectors is not a report, it is a log — and we would rather say where the limit is than let you find it.

The paid-plan report

One thing you will notice

Some of the scanned pages are not part of the demo at all — the scanner follows links within the same domain, and the demo sits inside the much larger w3.org. On a normal shop this is exactly what you want, because your pages link to your other pages. We are pointing it out because you would have spotted it yourself, and a tool that hopes you will not look closely is not worth having.

Your own would look like this

Same layout, your pages, your findings, arriving on a Monday. The first one is generated as soon as you sign up.

Watch it every Monday