Why must color never be the only way an interface conveys meaning, and what is a second indicator?

## The rule, and the four things it covers Success criterion **1.4.1 Use of Color**, at level A, states that color must not be used as the only visual means of conveying information, indicating an action, prompting a response, or distinguishing one visual element from another. Those four verbs are worth reading separately, because each one names a different everyday mistake: - **Conveying information** - a status dot whose color is the entire message. - **Indicating an action** - a required-field marker that is only a red tint on the label. - **Prompting a response** - a validation error signalled by turning a field's border red and nothing else. - **Distinguishing elements** - a link inside a paragraph that differs from the surrounding words only in color, or two chart series told apart only by line color. A closely related criterion, **1.3.3 Sensory Characteristics**, also level A, blocks the verbal form of the same failure: instructions that identify something purely by its color, shape, size, visual location, orientation or sound. Note that this is a *separate* requirement from contrast. A cue can be beautifully contrasted and still carry only one channel. Passing one says nothing about the other. ## Who this protects It is tempting to file the rule under color-vision deficiency and stop there. The real list is longer, and naming it is what makes an answer sound like experience rather than a checklist. | Group | What a color-only cue does to them | | --- | --- | | Color-vision deficiency | Two hues collapse into one and the distinction disappears | | Low vision, cataract, ageing eye | Hue and saturation discrimination degrade before acuity does | | Bright ambient light or glare | The screen washes out and hues flatten toward each other | | A dimmed, aged or low-quality display | The same effect, permanently, for that user | | An uncolored printout | Every hue becomes the same grey | | Anyone using speech output | Hue is never spoken at all, so the cue is absent rather than weak | | Anyone working at speed | Decoding a hue is slower than reading a word, for everybody | The last two rows are the ones candidates forget, and they are the ones that make the rule universal rather than a minority accommodation. ## What actually counts as a second channel The second channel has to be **perceivable in the same place, at rest, without an interaction**. That rules out several popular non-fixes: - A **legend elsewhere on the screen** mapping colors to meanings is not a channel on the item. It is a lookup table the user has to hold in their head while scanning. - A **hint that appears only while a pointer rests on the item** does not exist for a keyboard, touch or speech user. - A **second glyph in a different color but the same shape** is still one channel. Two identical dots, one green and one red, carry color and nothing else. What does count: 1. **A word.** Overdue beside the red dot. Cheapest, most robust, survives speech and print. 2. **A distinct shape.** A tick, a cross and an exclamation are three shapes; three circles are not. 3. **A pattern or texture.** Useful in dense data displays where words will not fit. 4. **Position or grouping.** A dedicated column, a section heading, an ordered position. 5. **Weight or size**, as a supporting cue rather than as the only one. The field test needs no tooling and works on any platform: **cover the color**. Render the screen without it, or put a thumb over the colored element, and ask whether every distinction survives. If it does not, the screen fails. ## Applying it when the schedule is already committed A hospital patient check-in app went through a redesign that was already in build when the accessibility review landed, with a 12-person delivery team and 43 screens in scope. The queue view showed each appointment's status as a single colored dot: green for arrived, amber for waiting, red for overdue. Every dot measured comfortably above 3:1 against its surface, and the screen still failed, because contrast and channel count are unrelated properties. The remedy that survived the schedule was not a redesign. It was one decision, taken once: every status in the product has a **name** as well as a color, and the name renders beside the dot. Seventeen of the 43 screens changed; none of them changed layout, and no new component was needed. The lasting benefit was not the 17 screens. It was that the status vocabulary became fixed, so the next screen the team built could not reintroduce the defect. The rule is cheap when it is a property of the product's vocabulary and expensive when it is a review finding raised against each screen in turn.