Distracting elements must be hideable

What it is

A support bubble beside the form, a promo overlay, an unread badge, a stack of toasts that insert themselves: for someone with an attention disability these are a second task running in parallel. Hide distracting content requires that regions unrelated to the current step, and that steal attention, can be turned off or hidden—and that they do not let themselves back in before the task is done.

Motion of the primary content is a different leaf: that one turns off movement. This one turns off things that should not be here at all. A static banner still counts if it keeps inserting, covering, or stealing focus.

Why it happens

Selective attention is capacity-limited. Every extra independent region is a capture source: a new message, a badge count, an avatar that twitches. People with attention disabilities are not “not trying”; the cost of inhibiting those captures is high enough that the primary task cannot be held. Hiding them returns the resource to the current step.

The difference between a distractor and the main content is whether the user chose to open it. A chat that expands itself inserts a new task by force. A toast that can be dismissed but returns with the same payload tens of seconds later was never hidden. Legally required copy may appear once, but it still needs to collapse; it cannot be a permanent film over the form.

Studying it

Dual-task: the same form with live chat and toasts versus all non-essential layers off. Compare completion time, off-task glances, errors. A cheaper cognitive walkthrough lists every region that updates or inserts on its own and asks whether it appears without a user click.

Independent variables: number of non-essential layers, auto-reappear interval, whether a persistent hide/focus control exists. Dependent variables: main-path completion, interruption count, reinsertion after dismiss.

Where it stops holding

Safety alerts (account under login, payment failed) may interrupt, but they must be rare, understandable, and not loop after dismiss. A still empty-state illustration that does not block action is not a distractor. Chat the user opened should stay while they use it; the task is “turn off what I did not ask for.” A primary button that is a timed promo countdown writes the distractor into the main path; “it can be closed” does not excuse it.

Applying it

• Offer a findable “hide extras / focus” control that suppresses chat entry points, promo bars, and non-safety notifications, and keep that choice for the session.
• Do not auto-expand chat or surveys. Toasts must not steal focus; they must be dismissible; the same payload must not reinsert.
• Verify: dismiss every overlay the user did not open, then complete the main task. Fail if an unsolicited region returns within thirty seconds. If the product advertises a focus mode, turn it on and count competing regions again—only the current step should remain.

Related

• Same group: J4.08.1 Autoplay and animation must be stoppable · J4.08.3 Long tasks must be breakable and savable
• Nearby: A5.07 Attentional capture · A5.08 Interruption cost and task resumption · J5.12 Live regions and dynamic announcements
• Search terms: hide distracting content · focus mode · attention disability