HRMIM
The Hotline Nobody Can Use: Two WCAG Failures at Level A

The Hotline Nobody Can Use: Two WCAG Failures at Level A

Timuçin Diş · Strategy Consultant· 11 November 2026· 10 min read

A reporting channel goes live. The address is published, posted on the noticeboard, added to onboarding. Six months later there are three reports in the box, and someone in the room says the obvious thing: clearly nothing serious is happening at the property.

There is one check to run before anyone says that, and it takes twenty minutes. Open the form, put your mouse in a drawer, and try to complete it end to end with the keyboard. If you get stuck anywhere, the channel is closed at that point. Not for everyone — but for everyone whose mouse is not an option.

Accessible form design gets treated here not as a design preference but as a measure of whether a channel functions at all. Below I take two WCAG 2.2 criteria, quote them as published, show how one very common interface shortcut fails both at once, and finish with three questions to put to a vendor. Two criteria — no more. A post naming eight is a checklist nobody completes. A post that names two, quotes them exactly and works one example through both is something you can act on this afternoon.

Accessibility is not a charity question, it is a does-the-channel-work question

In most companies accessibility sits in the wrong folder. It ends up under the sustainability report, in the goodwill section, on the list of things that would be nice if there is time. For a reporting channel that folder is wrong, because the question being asked is not one of goodwill. It is one of function: can the person who needs to file a report operate this form?

If they cannot, what you get is not a missing report. It is bad data. An empty box looks like a finding, and it is not one — it is a record of the instrument never reaching that person. This is a familiar measurement failure, the same one as zero reports as a measurement failure, and the data an unfilled form produces is always more misleading than the data from the ones filled in, because it never shows you who was left outside.

One thing needs saying up front, because this is the sentence most likely to be misread. I am not resting any of this on a Turkish legal requirement for accessible private-sector websites. I could not verify that such an obligation exists, so I am not asserting one. WCAG is the international reference standard — the criteria set buyers, corporate customers and public-sector tenders reach for. The argument in this piece is operational, not legal: a channel part of your workforce cannot operate is a channel that excludes them, and you do not need a statute to care about that.

Who reports from what device in a hotel

Most forms get built to the conditions on the desk of the team building them: a wide screen, a current browser, a fast connection, a mouse, one language and ten uninterrupted minutes. Every one of those conditions describes back-office staff.

The people most likely to file a report are in housekeeping, in the kitchen, in the laundry, in engineering and on the night shift. They open the form from a phone, in a corridor, between shifts. The screen is small, the signal drops in the service corridor, the handset may be several years old, and nobody has ten uninterrupted minutes.

Four questions to answer before anything is designed. None of them needs a statistic — the answers are all sitting in your own property:

  1. Which devices? Personal phones, a shared terminal, or both? Walk a shift and look, rather than assuming the device list.
  2. Which network? Guest wifi, staff wifi, or mobile data? What does the signal actually do in the service areas?
  3. Which language? What is the overlap between the language of the form and the languages spoken on the floor?
  4. How much time? Is the form built so it can be interrupted and resumed, or does it have to be finished in one sitting?

Answer those four and you will already know roughly where the two criteria below are going to break in your building.

Criterion 1 — WCAG 2.2 SC 2.1.1 Keyboard (Level A)

The first criterion, in the WCAG 2.2 text published by the W3C:

"All functionality of the content is operable through a keyboard interface without requiring specific timings for individual keystrokes…"

Two phrases carry the weight.

The first is "all functionality". Not most of it, not the main path, not the parts someone judged important. If something in the form can be done with a mouse, it has to be doable with a keyboard. One step outside that and the criterion is not met — and in practice one step is enough to close the whole form, because a form is a sequence and a sequence stops at its first broken link.

The second is "without requiring specific timings for individual keystrokes". Function cannot depend on how fast someone types. An older member of staff, someone with a tremor, someone typing one-handed while holding a phone in a corridor — all of them have to be able to finish.

What to do with that: for every component in your form that is reached with a mouse, ask separately how you get to it with a keyboard and how you get out of it again. Any component you cannot answer for is a component to test.

Criterion 2 — WCAG 2.2 SC 4.1.2 Name, Role, Value (Level A)

The second criterion, in the same document, at 4.1.2 Name, Role, Value:

"For all user interface components … the name and role can be programmatically determined; states, properties, and values that can be set by the user can be programmatically set…"

The key phrase here is "programmatically determined". Being visible on screen is not enough. What a component is, what it does, and what state it is currently in all have to be readable by software that cannot see the screen.

Reading the three words separately makes it usable:

  • Name. What is this component called? "Choose a category" — or nothing at all?
  • Role. What kind of control is it? A button, a listbox, a checkbox? Looking like a button is not a role.
  • Value, states and properties. What is selected right now? Is the field required? Is it in an error state? Is it checked?

A quick test you can run in your head: if you had to describe this component down a phone line to someone who cannot see the screen, could you say what it is called, what type it is and what state it is in? If you cannot, neither can the software.

The clickable div: one component failing both criteria

Now put both criteria on one very common interface shortcut. This is a pattern to look for in a form in front of you — a thing to test for, not a thing to attribute to anyone.

The pattern: the mandatory selector that asks what the report is about is built not from real buttons or a real list, but from a stack of clickable div elements styled to look like buttons. It looks immaculate. It works perfectly with a mouse. It passes design review without a comment.

That single component fails both criteria:

  • It fails SC 2.1.1, because an ordinary div does not take keyboard focus. You cannot tab to it, and because you cannot reach it, you cannot make the selection. "All functionality … operable through a keyboard interface" is not satisfied.
  • It fails SC 4.1.2, because that same div has no programmatically determinable name, role or value. A screen reader either announces nothing at that point or announces something meaningless. The user has no way of knowing a choice is required there.

The worst part of the pattern is where it sits. The category selector is a required field — the form cannot be submitted without it. So when the component fails, the experience does not degrade. The channel closes. The person can read the form, knows exactly what they want to report, and cannot send it. From the outside, that person simply did not file a report.

The test is one keystroke long: drop the mouse and try to tab to the selector. If you cannot reach it, nothing else about the form matters yet.

Why Level A sits inside AA conformance

Here is the part that earns its place in a vendor conversation. Both criteria above are Level A — WCAG's most basic tier. And Level A sits inside AA conformance: saying an interface meets AA is saying it also meets the Level A criteria.

The practical consequence is direct. A vendor claiming AA has to answer for SC 2.1.1 and SC 4.1.2 individually. These are not advanced criteria to be revisited in a later release; they are the floor of the level being claimed. A form whose required category selector cannot be reached with a keyboard does not have AA conformance, wherever that sentence is printed.

So ask for the claim as criteria, not as a letter. "AA" is a letter. "SC 2.1.1 met, SC 4.1.2 met, tested by this person on this date" is a result. A claim that cannot be broken down to that level is a claim and nothing more.

Mobile, low bandwidth, and shared devices

A form that meets both criteria can still break in three places in a hotel. These are fit-for-the-floor problems rather than criteria problems, and it is worth not confusing the two.

Small screens. Keyboard access that works on a desktop can overflow on a handset, or move focus to somewhere outside the visible area. Test on a real phone in a real hand, not in an emulator.

Weak connections. The signal in a service corridor, a basement or a laundry is not the signal on the office floor. The decision to make is what happens when the form is interrupted. If the text someone has written disappears, they will not sit down and write it a second time. People tell a difficult story once.

Shared devices. The terminal in the staff room or behind the desk is one of the likeliest entry points and also the riskiest. What does the next person see when they open the browser? History, autofill and a half-finished session can identify a reporter as effectively as a name field.

Those three are the whole difference between a channel built around the team on the floor and one built around the assumptions on a desk. The criteria tell you the form is technically operable. These three questions tell you it is operable in your hotel.

Language: the form is in Turkish, the worker is not

This gets its own heading precisely so it is not confused with the two criteria above. Language is not what either of them is about. But it cuts off access in exactly the same way: the person opens the form, does not understand what is being asked, and closes it.

In a seasonal hotel operation this is not an edge case. More than one language is spoken in the same kitchen, and the person filing a report is often not fluent in legal register even in their first language.

Three rules that are cheap to apply:

  1. Put the language choice above the first field. A flag icon buried in the footer does not exist for the person who needs it.
  2. Do not write the form in the language of the statute. Words like "disclosure", "declaration" and "consent" belong on the operator's screen. The reporter's screen should read like speech.
  3. Accept free text in whatever language it arrives in. Someone writing in their own language beats them not writing at all, every time.

How mandatory-field design breaks accessibility

Every required field is a gate. If the gate does not open, the quality of the form behind it is irrelevant — which is why required fields are where accessibility faults cost the most.

Two concrete break points.

First, being required is itself a state. The text of SC 4.1.2 covers states, properties and values explicitly: they have to be programmatically determinable. A red asterisk is a visual cue, not a programmatic state. If the fact that a field is required lives only on the screen, the person who cannot see the screen does not know what they are being asked to complete.

Second, the error message. If an incomplete submission turns a field red and does nothing else, the user never learns why it failed. I am not naming a third criterion here — the scope of this piece is two, and it stays two. But the operational outcome is not in dispute: the person presses the button, nothing appears to happen, presses it again, and gives up.

The cheapest fix on the design side: cut the number of required fields to the minimum. Every required field is one more place the channel can close. Ask which fields are genuinely needed before a report can be actioned. In most forms that number is smaller than the current one.

Four tests you can run in twenty minutes

No tooling, no budget, no specialist. Open the form and work through them in order. For each one, do not record pass or fail — record which step it broke on. That note is the thing that works in a vendor conversation.

  1. Put the mouse in a drawer and complete the form end to end. Tab, Shift+Tab, arrow keys, Enter and space only. The first place you get stuck is the place the channel closes. If you reach the end and cannot submit, the form does not exist for that user.
  2. Tab through and watch the focus indicator. Is the outline showing you where you are visible at every step? If it vanishes at one step, write that step down. Someone who cannot see where they are cannot finish.
  3. Zoom the browser to 200%. Do fields overlap, is text cut off, does the submit button leave the screen? This takes ten seconds and breaks more often than people expect.
  4. Submit the form empty. What happens? Is there an error message, or do fields just quietly turn red? If there is a message, can you reach it with the keyboard?

Run all four once, write down the result, and date it. That single page will be the most concrete document you bring to your next vendor meeting.

Three questions for your vendor

The tests above tell you the state of your own form. These three tell you the state of someone else's before you buy it.

1. Which version, which level, which criteria? "We're accessible" is not an answer. Which version of WCAG, which conformance level, and separate results for at least SC 2.1.1 and SC 4.1.2. Ask who tested it and when. A conformance claim that cannot be broken down by criterion is a marketing line.

2. Can I see the required fields completed with a keyboard? Not a screenshot — a recording. And not in the admin panel: on the screen the person filing the report will see. Those two surfaces are rarely built with equal care, and the one we care about is the one that gets less.

3. Can I try it on my own property's setup? The simplest question and the most revealing: I want to see a reporting form you can complete with a keyboard, not in a demo account but in the configuration my staff will actually be given. Every time that request is deflected or postponed, the deflection is the answer.

The value of a reporting channel is not that it exists on paper. It is that the person in the hardest position in the building can use it. If you want to work out what that position looks like in your hotel and whether the form survives it, get in touch — we will run the four tests above on your own screen together.

Get New Posts by Email

We will email you when a new blog post is published.

Loading the form…

Related Posts

Off the Payroll: Who Can Actually Use Your Ethics Hotline

Off the Payroll: Who Can Actually Use Your Ethics Hotline

Laundry, security, landscaping and spa are rarely on a hotel's payroll — but they work the same corridors. Limiting a reporting channel to payrolled staff is usually a habit, not a decision. This post frames the scope question with two reference points: joint principal–subcontractor liability under Labour Act art. 2, and the reporting persons listed in Directive art. 4. Then three costs of narrow scope, two of wide scope, and a four-question frame for writing the decision down.

30 September 2026 · 11 min read

No Company Email, No Desk: How a Seasonal Worker Reports Harassment

No Company Email, No Desk: How a Seasonal Worker Reports Harassment

Workplace harassment reporting almost never happens in a seasonal hotel. The reason is structural rather than cultural: no company email, no desk, a contract that ends in October. TBK 417 puts an affirmative duty on the employer, and Labour Act 24/II(d) attaches a direct consequence to sitting on a report. This piece names the six places the channel breaks in a seasonal team and closes with a checklist for building the internal channel around that constraint.

16 September 2026 · 11 min read

Before You Build an Ethics Hotline: Why Zero Reports Is Bad News

Before You Build an Ethics Hotline: Why Zero Reports Is Bad News

In most Turkish hotel groups the ethics hotline is a shared mailbox, and for years nothing arrives in it. This piece reads zero reports as a measurement fault rather than good news: it sets out the four conditions a reporting channel has to meet, shows that a shared inbox fails all four by construction rather than by neglect, looks at the two Turkish instruments that have written those conditions down, and closes with five questions to ask before you build anything.

2 September 2026 · 12 min read

Let's work with HRMIM

Explore our platform free of charge.