The Missing User: Why Every Knowledge System Needs an Audience

What if the greatest failure in knowledge management is not losing information, but forgetting who the information is for?

A catalog can be technically accurate and still be intellectually useless. A research environment can contain thousands of carefully connected documents and still leave a reader stranded. The difference is often a single neglected question: Who is the audience?

That question sounds modest, even administrative. In practice, it determines what gets collected, how it is described, which relationships are made visible, and whether knowledge can travel from one mind to another. The future of documentation depends less on building ever larger stores of information than on designing systems that understand the people approaching them.

The deeper lesson is that knowledge is not complete when it has been recorded. It becomes useful only when it can be encountered, interpreted, and reused by someone with a particular purpose.

A library is not a pile of books

Imagine two libraries containing exactly the same collection. In the first, books are arranged according to the order in which they arrived. In the second, they are organized through a rich system of subjects, references, cross connections, and pathways for different kinds of readers. The contents are identical. The experience of knowledge is not.

This distinction separates storage from documentation. Storage preserves objects. Documentation gives objects a social and intellectual address. It helps someone answer questions such as: What is this? Why might it matter? What does it connect to? Who would find it useful? What should I read next?

The idea of a universal research instrument, associated with the work of Paul Otlet, was ambitious precisely because it treated documentation as more than the accumulation of texts. It imagined a world in which the relationships among recorded facts could be made navigable. A document would not be an isolated item, but a node in a larger field of inquiry.

That vision remains strikingly contemporary. Modern catalogs, knowledge graphs, search tools, and research platforms all attempt to solve a related problem: how can a person move from a question to the relevant parts of a vast informational environment?

Yet contemporary systems often make a crucial mistake. They describe the collection while barely describing the encounter. They tell us what exists, but not enough about how different people might use it.

A data engineer, a historian, a policy analyst, and a student may all search the same catalog. Their needs are not interchangeable. One may need lineage and technical ownership. Another may need historical context. A third may need evidence that supports a decision. The fourth may need a conceptual introduction. A single description can be factually correct for all four and practically adequate for none.

This is why audience is not a publishing detail. It is an organizing principle.

The hidden design choice inside every catalog

Every knowledge system makes choices about its audience, whether those choices are explicit or not. A highly technical catalog assumes that users already understand its vocabulary. A visually polished research interface assumes that users can infer the logic of its categories. A system that exposes every available field may be designed for administrators, even when it claims to serve researchers.

The absence of an identified audience does not create neutrality. It creates a default audience: usually the people who built the system.

This produces a common paradox. The more sophisticated a catalog becomes internally, the less accessible it may become externally. Its creators add fields, classifications, permissions, relationships, and workflows. Each addition makes sense from inside the institution. To a newcomer, the result can feel like entering a city without street signs. The infrastructure is impressive, but the path is unclear.

Consider a dataset titled Regional Health Outcomes 2018 to 2024. A technical description might state its owner, update frequency, schema, and storage location. Those details are valuable. But they do not answer several questions that determine whether the dataset can be responsibly used:

• What decision was this dataset created to support?
• Which population does it represent, and which does it exclude?
• What would a public health researcher misunderstand if they approached it without institutional context?
• Which related datasets should be consulted before drawing a conclusion?
• Is the intended user a data scientist, a policy maker, a journalist, or a member of the public?

These are not decorative questions. They affect interpretation. A field called status may mean legal status, workflow status, or publication status. A date may indicate collection, revision, approval, or observation. Without audience aware context, metadata can transmit ambiguity while appearing precise.

The practical consequence is that a catalog should be designed around use cases rather than objects alone. The central unit is not merely the document or dataset. It is the relationship between a resource, a question, and a reader.

A knowledge system does not truly organize information until it organizes the possible journeys through that information.

From universal classification to plural pathways

The dream of a comprehensive system for organizing knowledge is powerful, but the word “universal” can mislead. It suggests that one definitive structure could serve every inquiry. In reality, different audiences need different routes through the same intellectual terrain.

A novice may need a guided sequence: basic concepts first, then examples, then technical detail. An expert may want to bypass introductions and inspect provenance, contradictions, or raw records. A decision maker may need a concise synthesis with confidence levels and implications. A researcher may want to follow unexpected associations, including sources that challenge the dominant interpretation.

These are not competing versions of the truth. They are different interfaces to the same evidence.

A useful mental model is to think of a knowledge environment as a landscape rather than a warehouse. The landscape contains mountains, paths, landmarks, hazards, and viewpoints. A tourist, a geologist, and a rescue team can occupy the same terrain while requiring entirely different maps. The map does not change the landscape. It changes what becomes visible and what can be reached.

This suggests three layers of documentation.

The object layer answers: What is here? It includes titles, formats, dates, owners, subjects, and technical attributes.

The relationship layer answers: What is this connected to? It includes citations, dependencies, versions, related topics, institutional history, and conflicts or complementarities with other materials.

The audience layer answers: Who might use this, for what purpose, and with what prior knowledge? It includes intended users, likely questions, recommended entry points, cautions, and possible next steps.

Many catalogs are strong at the first layer and uneven at the second. The third is often left to chance. But without it, the other layers remain underinterpreted. A relationship is not meaningful in the same way for every user. A dependency that matters to an engineer may be irrelevant to a social historian. A citation that looks peripheral to a manager may be essential to a researcher investigating bias.

Audience aware documentation does not mean simplifying everything. It means making complexity navigable. It allows the same resource to support multiple depths of engagement without forcing every user through the same doorway.

The ethics of making knowledge findable

There is also an ethical dimension. To make information discoverable is to shape what people can know, question, and act upon. Classification systems are never merely mirrors. They foreground some connections and hide others. They define which terms count as legitimate, which sources appear central, and which users are expected to do the work of interpretation.

An inaccessible system often shifts its costs onto the least powerful audience. Experts can ask colleagues for context. Insiders know the local abbreviations and institutional history. Newcomers, independent researchers, and members of the public encounter only the visible surface. If that surface is poorly designed, exclusion can look like user incompetence rather than system failure.

This is particularly important when documentation concerns public resources, scientific evidence, or decisions that affect communities. A catalog that is technically open but practically opaque may satisfy a formal requirement while failing a democratic one.

The remedy is not to erase specialized language. Specialized language can be precise and necessary. The remedy is to provide translation layers. A resource can retain its formal schema while also offering a plain language description, examples of appropriate use, known limitations, and links to background concepts.

Think of this as interpretive infrastructure. Roads make physical movement possible. Interpretive infrastructure makes intellectual movement possible. It includes definitions, context, provenance, audience labels, examples, and routes to adjacent knowledge.

A mature system should therefore ask not only whether a resource is available, but whether a reasonable person can understand its significance without already belonging to the institution that produced it.

A practical framework for audience centered knowledge

Audience can be turned from an abstract value into a design method. Before adding another field or building another search feature, map the system around four questions.

1. Who is arriving?

List the actual communities that approach the collection. Do not stop at job titles. Identify their levels of expertise, languages, constraints, authority, and reasons for searching. A researcher looking for historical evidence and an executive looking for a decision signal may use identical keywords while needing very different forms of support.

2. What are they trying to accomplish?

Replace vague goals such as “find information” with observable tasks. Are users comparing sources, verifying a claim, locating a reusable dataset, learning a field, tracing a decision, or generating a new hypothesis? The task determines which metadata matters most.

3. What could they misunderstand?

This question is unusually powerful. It moves documentation beyond description toward risk prevention. Identify ambiguous terms, missing populations, outdated assumptions, misleading visual cues, and relationships that are invisible to outsiders.

4. What should happen next?

A successful encounter rarely ends with a single item. The user may need a related source, a methodological note, a newer version, a dissenting interpretation, or a person who can answer a question. Every important resource should offer meaningful next steps.

These questions can be condensed into a simple design test:

For a specific audience, can the system support discovery, interpretation, and responsible reuse without requiring invisible local knowledge?

If the answer is no, adding more content may worsen the problem. The system needs better pathways, not merely a larger inventory.

A small team can apply this framework immediately. Select one frequently used resource and document three audiences. For each audience, write a one sentence purpose statement, one likely misunderstanding, and three recommended next steps. Then ask real users to complete a task using the revised description. Measure not just whether they found the resource, but whether they interpreted it correctly and knew where to go afterward.

That last measure matters. Search success is not the same as knowledge success. A person can click the correct result and still draw the wrong conclusion.

Key Takeaways

• Name the audience before designing the description. A resource should not have one generic explanation when its users have different purposes and levels of expertise.
• Document relationships, not just attributes. Explain dependencies, context, provenance, related materials, and disagreements so that users can understand significance rather than merely locate objects.
• Treat misunderstanding as a design problem. Record what an outsider could reasonably infer incorrectly, then address those risks directly.
• Create multiple entry points. Offer plain language orientation, technical detail, historical context, and expert pathways without forcing every user into the same route.
• Measure interpretation and reuse. Ask whether people can apply what they find responsibly, not only whether they can find it.

The ambition behind a universal research environment should not be abandoned. It should be revised. The goal is not one map that claims to be equally intuitive to everyone. The goal is a connected world of knowledge with enough maps, legends, and guided routes that different people can enter without losing their way.

A catalog is often described as a record of what an institution knows. More accurately, it is a proposal about how that knowledge may be encountered. Its categories, labels, links, and omissions quietly tell readers what is important, what is related, and who is expected to understand.

That is why the smallest visible field can carry the largest consequence. A simple audience label may determine whether a document becomes useful evidence or remains institutional furniture. The future of knowledge systems will belong not to those that collect the most, but to those that make the relationship between information and human purpose impossible to overlook.

The central question is therefore not, “How much knowledge have we organized?” It is this: “Whose next question can our organization of knowledge make possible?”