Information scent

How clearly a link or label signals what you will find after clicking it. Strong scent makes the next move feel obvious.

does this link tell me where it goesI can't tell which menu item to clickmake link labels less vaguewhat is behind this buttonwill this click get me closerthe menu names don't match what I'm looking forlinks that smell like the answerinfromation scent

See it

Live demo coming soon

What it is

Information scent is how strongly a label, link, or preview predicts the useful information behind it. The term comes from Peter Pirolli and Stuart Card at Xerox PARC in the 1990s, who borrowed the maths of animal foraging: a bird works a patch until the cues say a better patch is elsewhere, and a person works a menu the same way. 'Shipping costs and delivery times' has strong scent; 'Learn more' asks for a blind click.

Reach for it when naming navigation, search results, cards, calls to action, and help links. Reuse the words people already have in mind, add a short preview where the destination is ambiguous, and make the landing page heading confirm the promise the link made.

The gotcha is optimizing labels for internal organization or cleverness. A team may know exactly what 'Solutions' or 'Explore' contains while a new visitor sees fog. Strong scent is specific, not merely short. Test the first click for a real task and watch where people go before explaining what the menu was supposed to mean.

Ask AI for it

Strengthen the information scent of this navigation. Replace every vague label such as 'Learn more', 'Explore', or 'Solutions' with a destination-specific noun phrase that uses the same words as the target page h1. Add a one-sentence preview under cards whose destinations are still ambiguous, keep link text unique within the page, and use breadcrumbs with aria-current='page' so people can predict their location before and after the click. Do not use icon-only navigation for unfamiliar destinations. Validate the rewrite with first-click testing: give five people a concrete find-or-do task, record their first choice and confidence, and revise any label that sends more than one person to the wrong destination.

You might have meant

discoverabilityaffordancerecognition over recallcontextual just in time helpmental model