Feedforward tells users what will happen before action occurs
Aliases: feedforward · pre-action information · outcome preview
What it is
Feedforward is information provided before a person acts that communicates whether an action is available, what it will affect, and what result may follow. Button labels, reasons for unavailability, highlighted drop targets, a pre-delete impact summary, and input-format examples can all be feedforward. It complements feedback: feedback explains a result that has occurred, while feedforward helps form a reasonable expectation before action.
Why it happens
Among possible actions, people need to predict which is useful and whether its consequence is acceptable. Feedforward externalizes system rules, current state, and expected outcome, reducing reliance on probing and memory. It may persist in a static interface or appear contextually on hover, focus, drag start, or during entry. Clear expectation lets people change course before a consequence happens rather than recover afterward.
Studying it
Before acting, ask first-time users to state their predicted result, conditions, and risks, then compare prediction with reality. Observe error attempts, hesitation, undo, help seeking, and completion quality with and without feedforward. For consequential action, test understanding of scope and irreversibility, not merely whether the cue was seen. Also assess whether feedforward appears when needed rather than being submerged by irrelevant information.
Where it stops holding
Feedforward cannot precisely promise outcomes that are genuinely uncertain; it should state uncertainty, influencing factors, and controllable range instead of pretending to guarantee. Showing every rule in advance creates overload, especially in complex systems. For highly familiar, low-consequence actions, extra explanation may be noise. Feedforward should vary with task novelty, risk, and context.
Applying it
- Before people commit, present the most important availability conditions, scope of impact, outcome, and irreversible consequences.
- During dynamic operation, show accepted targets and anticipated change through drag previews, input validation, and pre-action summaries.
- Test whether people can predict correctly, not merely whether explanation exists; revise information or simplify action where misunderstanding appears.