Accessibility, measured rather than claimed

Your volunteers did not choose your software: you chose it for them. The one who magnifies the screen, the one who never touches a mouse and the one who listens to the page all need to sign up, see their schedule and state their availability. We measure where we stand, and we publish the result, including what it does not cover.

100% of the applicable RGAA criteria

The grid of the 106 criteria of the French RGAA 4.1 is kept screen by screen, and published: 87 applicable criteria out of 87 are met, the other 19 not applying to this product (no video, no CAPTCHA, no time limit). It is a self-assessment, not a third-party audit, and the accessibility statement says so plainly, along with the share of criteria judged by hand.

Three measurements, re-run whenever a screen changes

Accessibility degrades without warning: a button added on Tuesday, a colour changed on Thursday. It only holds if it measures itself, before every release.

• The rules a machine can check, WCAG 2.0 and 2.1 levels A and AA, over 54 screens: contrast, accessible names, ARIA, heading and form structure. No violation.
• A full keyboard walk, over 190 screens: forty tab presses per screen, tab order compared with visual order, visible focus, traps, the skip link, and controls only a mouse can reach. No finding.
• Script errors, over 77 screens: a broken script leaves a screen half usable, and the keyboard user is the first to get stuck in it.

These three cover every screen of the product, not a sample. The inventory is computed from the application's routes, crossed with the test data set: a screen no data can open is measured by nothing, and that is exactly where the defects we found were hiding.

What that means in practice

With a keyboard. The first tab stop on every page offers "Skip to content", which jumps the bar, the search box and the event tabs. Focus is always visible. No trap: you get out of every field, every menu, every dialog. Where dragging exists, in the availability calendar and the global schedule, it doubles a form that does the same thing, because drawing with a mouse has no keyboard equivalent.

Magnified and in high contrast. Text holds at least 4.5:1, including over the team colours the server computes on the fly: an assignment's label picks black or white according to its team's hue, and the worst case measured across the 360 possible hues stays above the threshold. Pages read at 375 px wide with no horizontal scrolling, and list tables turn into cards on a phone.

With a screen reader. Every page carries a main landmark, a heading hierarchy and labels linked to their fields. Row actions say what they act on: "Edit", in a twenty-row list, becomes the same word twenty times when you browse controls out of their row.

Without motion. When the visitor's system asks for less animation, transitions stop: that is an accessibility setting for people with vestibular disorders, not a matter of taste.

What we do not measure, and why we say so

Automated checking catches between a third and a half of real accessibility defects. The rest needs a human, and sometimes a human who is concerned. We have not yet tested with a screen reader, nor with disabled people, and no tool judges the quality of a label.

It is written plainly in the accessibility statement, with the full list of limits and the address for reporting a problem. Software that shows an "accessible" badge without saying what it checked leaves you carrying the risk; a statement that says where it stops gives you something to answer your own obligation with.

Why this concerns you

If you are a local authority, a public body or a publicly funded association, French law (article 47 of the law of 11 February 2005) applies to your online services, and a tool you put in your volunteers' hands is one of them. Your own accessibility statement has to rest on something. We give you the figures, the date and the method.

And if you are an association under no particular obligation, the question is the same: out of a hundred volunteers who sign up, some enlarge the page, some cannot handle a mouse, and some listen to the screen. They will not tell you, they will give up.

Frequently asked questions

Is Volunteo WCAG conformant?

Partially, and we write it down rather than display a badge. No audit against the 106 criteria of the French RGAA has been carried out. What is measured is measured exhaustively and published with its figures; what is not is listed by name in the accessibility statement.

Can a blind volunteer sign up unaided?

The public sign-up form can be reached and filled in with a keyboard, its fields have linked labels, and its error messages are announced. We have not yet tested it with an actual screen reader, and we say so: automated checking catches between a third and a half of real defects.

Is there an option to switch on?

No. There is no "accessible mode": it is the same application for everyone. A separate mode is a mode you forget to keep up to date.

Our authority has to document its tools. What can you provide?

The accessibility statement, which is public, and on request the detail of the measurements: the list of screens covered, the rules checked and the date of the last run. Write to us through the contact form.

Read next