Glossary · 2 · Topics and information design

Information foraging

Also known as: Information foraging theory, IFT, Informavore model

In technical communication and human-computer interaction, information foraging is a theory that explains information seeking as behavior shaped by the value of the information sought and the effort needed to get it. Developed by Peter Pirolli and Stuart K. Card at Xerox PARC and published in Psychological Review in 1999, it borrows concepts from optimal foraging theory in ecology: readers move between information "patches", follow "information scent" given off by proximal cues such as link labels, headings, titles and metadata, and abandon a patch when the expected yield drops below what another patch promises. The theory is descriptive, not a standard or a method, and it predicts behavior rather than prescribing document structure. In documentation practice it is used as a rationale for making individual pages self-sufficient and for strengthening the cues by which readers judge, enter and leave a page.

Information foraging
Information scent
Findability
Every Page is Page One
Reading behavior

In one sentence

Information foraging explains how readers follow information scent between content patches — and why documentation pages must stand alone.

Example

After a support review showed that users reached the lubrication interval through search rather than the manual's table of contents, the writing team restated the machine model and the safety prerequisites at the top of that topic, because foragers landing there had no preceding page to rely on.

Information foraging theory grew out of research at Xerox PARC in the mid-1990s: a CHI 1995 paper on information foraging in information access environments was followed by the full theory in Psychological Review in 1999. Its vocabulary — informavore, patch, scent, diet, enrichment — is an analogy to animal foraging, and its core claim is an economic one: people spend attention where the expected gain per unit of effort is highest, and they leave as soon as a different patch looks better. In user experience work the theory became widely known through its link to the concept of information scent, which describes how readers judge a link or heading by its wording and its surrounding context before committing a click.

Mark Baker made the theory one of the explicit foundations of Every Page is Page One (XML Press, 2013). His argument is that readers of technical documentation arrive through search engines, help links and portals, so any page may be the first page a reader sees; a page that assumes the reader came from the previous page offers weak scent and little value, and the forager moves on. Foraging therefore supports pages that declare their subject and type, establish their context, stay within one purpose and link outward to adjacent subjects instead of relying on a linear reading sequence.

How it applies

Treat every topic as a landing page. State the product, variant, software release or machine configuration the topic applies to, and repeat prerequisites and warnings there; a reader who entered by search has not read the front matter. This does not make a document complete or compliant — required content for information for use and instructions for use follows from the applicable regulations and standards, not from a reading model.
Invest in the cues, not only the content. Titles, headings, link labels, summaries, breadcrumbs, facets and metadata are the scent a forager smells before deciding. Task-shaped wording ("Replace the filter cartridge") carries more scent than a feature label ("Filter module").
Use a controlled vocabulary for scent at scale. A consistent taxonomy and product metadata let a portal show readers which patch they are in and which neighboring patches exist; inconsistent labels break the trail.
Design for abandonment. Assume readers leave mid-page. Put the answer early, keep procedures short, and make exits explicit — related topics, parent assembly, error-code references — rather than hoping readers scroll.
Software and machinery documentation differ in patch size. Context-sensitive help for a user interface delivers small patches reached from the application, while maintenance documentation is often reached from an error code, a part number or a QR code on the machine; both benefit from topics that are addressable and self-contained, for example as a DITA topic or an iiRDS topic.
Measure the trail. Search logs, zero-result queries, entry pages and exit pages are the observable side of foraging and show where scent is missing. Treat such data as indicative, not as proof that content is sufficient.
Expect machine foragers too. Retrieval systems and AI assistants also consume documentation in fragments; topics with explicit context and clean metadata survive that fragmentation better than chapters that depend on their position in a book.

Information foraging vs. information scent

Information foraging is the overall theory of how people seek information under cost and value constraints. Information scent is one construct inside it: the reader's estimate, formed from proximal cues, of how likely a path leads to the desired information. Foraging explains why readers switch patches; scent explains what they read to decide. A page can offer strong scent in its link labels and still be a poor patch if the content behind the link is thin.

Information foraging vs. Every Page is Page One

Foraging is a descriptive theory from cognitive psychology and applies to any information environment. Every Page is Page One is a documentation model that uses foraging as part of its justification and adds prescriptive characteristics for topics, such as self-containedness, a declared type and rich linking. Citing foraging therefore does not commit a team to that model, and adopting the model does not follow automatically from the theory.