A critical grouping relationship needs a non-visual, perceivable encoding as a fallback

What it is

Color, spacing, and container background all share one assumption as Gestalt grouping cues: that the user can see the screen and resolve fine differences on it. Users with color vision deficiency, low vision, or a reliance on a screen reader each fail a different part of that assumption — someone with a color vision deficiency can see spacing fine but not a specific color contrast; a screen reader user does not depend on the picture at all. That means no single purely visual cue can cover everyone at once. The only reliable approach is to pair a genuinely important grouping relationship with an encoding perceivable through a non-visual channel, so the grouping information does not vanish entirely just because one particular visual ability is missing.

Why it happens

If a grouping relationship is conveyed through only one visual channel, the moment the perceptual ability tied to that channel is missing, the whole piece of information disappears with it, because no other channel holds a copy of it. Giving a critical relationship a second encoding that runs through a different perceptual channel — writing the visual grouping into the markup structure as well, or simply stating in text that these items belong together — stores the same information twice, through two mutually independent means, so the relationship survives as long as at least one copy remains perceivable. This is the same underlying principle as adding a shape alongside color in the color-vision-deficiency case, just applied across a broader range of failure modes: don't bet critical information on any single cue holding up everywhere — keep a backup for it.

Where it stops holding

This principle applies only to grouping relationships that are genuinely "critical" — it is not a license to pile up multiple redundant encodings on every single grouping regardless of cost. A purely decorative grouping, one whose loss would not affect a user's understanding or ability to complete a task, does not need a dedicated fallback, and over-applying this principle just makes the interface itself more complex. Whether a grouping relationship counts as critical is decided by what happens if it is lost: if losing it would cause the user to misread which content belongs where, or to miss information the task requires, it is a critical relationship that needs a fallback.

Applying it

• First sort every grouping relationship in the interface into two categories: purely decorative reinforcement, and relationships users genuinely need to understand content or complete a task. Only invest the cost of a non-visual fallback encoding in the latter.
• For a critical relationship, identify which single visual channel it currently rides on (color, spacing, container), then add an encoding that does not draw on that same channel at all — write the grouping into the markup structure, or simply spell out the relationship in text — rather than "backing up" one visual cue with another cue that still depends on vision.
• To verify: assume, one at a time, that a given visual ability (color perception, sharp visual acuity, seeing the screen at all) is completely unavailable, and check each critical grouping relationship to see whether it can still be perceived. If any critical relationship becomes entirely imperceptible under one of these assumptions, the fallback is not adequate and a genuinely independent encoding needs to be added.

Related

• Same group: A2.15.1 Grouping by color alone fails for users with color vision deficiency · A2.15.2 Spacing-only visual grouping does not survive a screen reader's linear read-out · A2.15.3 Low-contrast container cues can be entirely invisible to users with low vision · A2.15.5 Purely visual grouping can fail after zooming or resizing text, once spacing ratios shift
• Nearby: U1.10 Dual and redundant encoding · J5.01 Screen readers
• Search terms: redundant encoding · fallback encoding · non-visual channel