Every mistaken state needs a clear exit
Aliases: exit state · return path · mistake recovery
What it is
User control and freedom requires every mode, filter, flow, dialog, or wizard that users may enter by mistake to have a visible, predictable, nonpunitive exit. The exit should say what will happen: close the panel, discard a draft, return a step, or keep a background task.
Why it happens
Wrong entry arises through exploration, mistargeted input, defaults, or permission jumps. Without an exit, users become trapped in a mode and resort to refresh, killing the app, signing out, or submitting unwanted data; when the exit’s meaning is unclear, they fear losing work and avoid it. A clear exit turns an unintended action from a risky commitment back into reversible exploration.
Studying it
Create mistap and exploration tasks, then observe how participants leave a nontarget state. Record time to find the exit, use of refresh/back/app kill, repeated actions, anxiety about data loss, and abandonment. Compare exit placement, labels, close semantics, and confirmation; test separately on mobile, keyboard, screen readers, and stacked dialogs.
Where it stops holding
A clear exit is not unconditional exit. Payment, deletion, examination, and legal signing need confirmation, delay, or administrative intervention; unsaved work must not disappear silently. Immersive mode, security locks, and managed devices may restrict return, but they must explain why and how to get help.
Applying it
- Define close, cancel, back, and save-draft semantics for every mode, wizard, filter, and dialog.
- Put exits where the current context predicts them and label outcomes explicitly, such as “Cancel and discard draft.”
- Support hardware/keyboard back, Escape, gestures, and assistive shortcuts, and keep close controls unobstructed.
- Walk through every reachable state: enter it by plausible random clicks and check return to the main task within three steps without loss.