If users can't tell where a link goes, they won't take the risk of clicking it.

Information scent comes from Information Foraging Theory, developed by Peter Pirolli and Stuart Card at Xerox PARC in the 1990s. They borrowed from evolutionary biology — specifically the way animals decide whether a food source is worth pursuing based on environmental cues.

The theory applied to digital interfaces: users behave like foragers. They scan an interface for cues about whether the information they need is "nearby." When those cues are strong — when the label, context, and surrounding text make the destination obvious — users follow the path. When the cues are weak or ambiguous, users back up, try a different path, or leave entirely.

Weak information scent is one of the most common causes of poor navigation — and one of the least diagnosed, because it's invisible to anyone who already knows the product well.

Signs of weak scent:

• Generic labels: "Resources", "More", "Information", "Details" — labels that could lead anywhere
• Icon-only navigation: Icons without text labels force users to guess, and most icons don't carry universal meaning across contexts
• Jargon-heavy category names: Internal language that maps perfectly to how the team thinks, but not how users search
• Mismatched landing pages: The link says one thing; the page delivers something slightly different. Users feel misled, even if the content is technically relevant.

Strong scent usually comes from three things working together:

1. Specificity in the label: "View invoices from the last 90 days" has stronger scent than "Billing". More words, more signal — especially for high-stakes navigation.
2. Context around the link: A link surrounded by related content gains meaning from its neighbours. Isolated links lose context and force users to make a blind decision.
3. Consistent follow-through: The page you land on matches the expectation set by the label. This is where many products actually break down — the navigation is fine, but the landing page doesn't deliver on the promise.

Information Architecture decisions live upstream of all three. A well-structured IA creates the conditions for strong scent; a poor one makes it nearly impossible to achieve at the label level alone.

The most common failure point isn't labeling — it's the gap between how products are organized and how users think about their own goals.

Users don't navigate by product features. They navigate by tasks. A product organized around features — "Analytics", "Reporting", "Insights" — creates weak scent for a user who just wants to know how many people signed up last month. Their mental model doesn't map to the structure.

This is an Information Architecture problem at its root, but it surfaces as poor scent. Card Sorting and Tree Testing are the two most reliable methods for testing whether your navigation structure matches user mental models before you build it.

Improving information scent is mostly language work and testing — not visual redesign:

• Run a click test: Show users a task and ask where they'd expect to find the answer. Tools like Optimal Workshop or Maze surface exactly where scent is breaking down, faster than session recordings alone.
• Label audit: Walk through every navigation item and category label. Would a first-time user know what's behind this? If not, sharpen the language — lean toward task-oriented labels over feature-oriented ones.
• Destination review: For every important navigation link, verify the landing page delivers what the label promised. This is often where scent actually breaks — not at the link, but after the click.
• Talk to users about their vocabulary: Contextual Inquiry or short interviews about how users describe what they're trying to find will surface the language gap between your labels and their mental models.