Information Scent
Information scent is the strength of the cues that help users predict whether a link, label, or result will move them toward their goal.
Download Markdown (.md)Soroush’s Product Wiki · https://wiki.soroush.uk/topic/information-scent/
01What it means
Information scent describes the perceived relevance of available cues—such as link text, headings, snippets, categories, and visual context—to a user’s current information need. The idea comes from information foraging theory: people estimate the likely value and cost of following a path, then continue, switch direction, or leave based on those signals.[1][2][3]
- 01Information needGoal and language used to judge relevance
- 02Available cuesLink, heading, snippet, category, metadata
- 03ExpectationPredicted destination, scope, and value
- 04Route choiceContinue, switch direction, or leave
- 05Evidence after clickConfirm scent or trace the mismatch
Weak, vague, or deceptive cues cause wrong turns even when the destination itself is well designed.
02Why it matters
Strong scent lets people choose efficiently without opening every option. It improves navigation, search-result selection, and confidence because labels communicate both destination and expected value. Weak or misleading scent causes pogo-sticking, abandoned journeys, and reliance on search even when the desired content is nearby.
03When to use it
Use information scent as a diagnostic lens when users hesitate among links, choose plausible but wrong categories, bounce between results, or overlook useful content. Apply it while naming navigation, writing links and search snippets, designing cards, and deciding what context to expose before a click.
04How it works
- State the user’s likely goal and the words, distinctions, and evidence they use to judge relevance.
- Inventory the cues available before each click, including labels, surrounding text, metadata, imagery, and placement.
- Rewrite vague or promotional cues so they describe the destination, scope, and distinguishing value in user language.
- Compare neighboring options to ensure each predicts a different destination and does not overpromise.
- Observe route choices and wrong turns in realistic tasks, then trace failures back to a missing, weak, or deceptive cue.
05A practical example
In a software help center, “Manage access and roles” replaces “Administration,” while article cards include task-specific titles and product-version labels. A user looking to invite a contractor can distinguish role settings from account security before clicking, reducing visits to irrelevant pages.
06Where it breaks
Scent is relative to a person’s vocabulary, experience, and goal; a strong cue for an expert may be meaningless to a novice. Metrics such as click-through can reward sensational or deceptive labels, so successful scent must be judged by downstream task success. Visual prominence cannot compensate for an inaccurate promise.
Common mistakes
- Using generic labels such as “Learn more” when several links appear together.
- Optimizing clicks with exaggerated cues that do not match the destination.
- Evaluating a label alone instead of comparing it with neighboring choices and context.
07Sources
- Information Scent: How Users Decide Where to Go NextNielsen Norman Group · Research-based guidanceInformation Scent: How Users Decide Where to Go Next
- 3 Common IA MistakesNielsen Norman Group · Usability guidance3 Common IA Mistakes
- Better Link Labels: 4Ss for Encouraging ClicksNielsen Norman Group · Research-based guidanceBetter Link Labels: 4Ss for Encouraging Clicks