User intent is what information science has called the information need since 1962.

An information need is a person's conscious or unconscious desire for knowledge, to close a gap in understanding or to get a task done. It exists before any search and independent of it. User intent is this need, seen from the side of search.

A gap in knowledge is the absence of information that moves a person to look for facts, data or ideas. It is felt before it can be named. The search starts from the feeling of the gap.

Relevance links an information object to the task or question of a particular person. It is a relation between the person and the object. A document is relevant to someone, for something.

The visceral need is the first of Robert Taylor's four levels: an early, unexpressed and vague sense that something is missing. The person may not yet be able to put it into words. This is the kitchen table, before any query.

The conscious need is Taylor's second level: a mental description of a question that is still ill-defined. The person can think about it and may talk about it. It is still far from a search term.

The formalised need is Taylor's third level: a concrete statement of the question before it meets a system or a human helper. It is what a person would say if asked carefully. It is already narrower than what they meant.

The compromised need is Taylor's fourth level: the question reshaped to fit what a system can take. A query in a search box is a compromised need. Search intent is read from this last and smallest form.

The anomalous state of knowledge, a hypothesis by Nicholas Belkin, holds that searching begins when a person recognises an anomaly in what they know. The person cannot specify what would resolve it. That is why a query can describe the need only partly.

Sense-making, developed by Brenda Dervin, studies how people bridge a gap in a particular situation by making sense of information. It places the person, the situation and the use of information at the centre. A search is one step of building sense.

The Information Search Process, developed by Carol Kuhlthau, describes searching as stages over time, with thoughts, actions and feelings at each stage. Uncertainty and anxiety are part of the model. Searching has a course, and people feel it.

If the need were fixed at the first query, thoughts and confidence would stay where they started. Across one search they move, and half of the users still had no focus at the end.

A verificative information need seeks to check a known item or attribute, such as an author, a citation or a date. Peter Ingwersen distinguished it from topical needs. The person already knows what they are looking for.

A conscious topical information need aims to clarify or pursue a subject within a domain the person knows. The person can describe the topic. Search works best for this kind of need.

A muddled topical information need concerns new territory the person does not yet know how to describe. It comes with uncertainty about concepts and words. A query for it is the least reliable description of all.

Informational intent is the search class for queries aimed at knowledge or content on a topic. It is the class most often assigned. It covers everything from a quick fact to a long investigation.

Navigational intent is the search class for queries meant to reach a particular website or destination. The need is already specific. The query is a shortcut to a known place.

Transactional intent is the search class for queries that aim to perform an action, such as buying or downloading. It marks the moment of doing. The need that led there is usually older and larger.

A simulated work task is an evaluation method in interactive information retrieval, developed by Pia Borlund, that gives test persons a realistic situation to search from. It brings the need back into the experiment. It shows how different the same query becomes in different situations.

A query is the verbal or typed input a person submits to a search system. It is the visible trace of a need. It is never the need itself.

The information user is the person who experiences the gap, searches and uses what is found. Every model of need starts with that person. Intent taxonomies start with the query.

Robert S. Taylor described the information need in 1962 and set out its four levels in 1968, from visceral to compromised. His model shows how a need shrinks on its way into a system. Every discussion of user intent stands on it, whether it names him or not.

Time, the task and the feelings. The usual accounts treat intent as a static category attached to one query. They leave out that needs change while people search, that tasks have stages and span sessions, that uncertainty and anxiety shape every query, and that people sometimes avoid the very information they seem to look for. The user intent is larger than the query, it moves, and it is felt.

An evolving information need changes as the person reads what they find. Marcia Bates described searching as berrypicking, where each find reshapes the question. An intent label fixed at the first query describes a need that no longer exists.

A latent information need is a need the person has without having expressed it, sometimes without being aware of it. It lies before the first query. It is where recommendations and good pages can help before anyone asks.

An affective information need is the wish for reassurance, confidence or the experience of others, alongside facts. People searching about an illness often want to know that they will be all right. A definition answers the query and leaves the person alone.

Task uncertainty is how unsure a person is about what the task actually requires. It changes how they judge relevance. A system that treats an uncertain searcher like a certain one serves neither.

The pre-focus phase is the early stage of Kuhlthau's search process, when the person explores without a clear focus and feels confusion. Broad orientation helps most here. A precise answer snippet arrives too early.

A task stage is the point a person has reached in the larger work their search serves. What counts as useful changes from the beginning of a task to its end, as Pertti Vakkari showed. The same query at two stages needs two answers.

Goal ambiguity means the person does not yet know exactly what would count as a good result. They cannot judge usefulness reliably. A quick exit then says more about the vague goal than about the page.

Information drift is the gradual shift of topic across a series of related searches. Each search moves the question a little. Treating the searches as separate events misses the path.

Search goal modification is the refinement or redirection of a goal in response to what was found. The target moves during the search. Results tuned to the first goal keep answering a question the person has already left.

A multi-session search task extends across several sittings, days or weeks. Planning, research and purchases often work this way. Each new session starts from where the person was, and the system usually starts from nothing.

If every query stood for a fresh need, repeats would be rare and shared across people. A third of all queries repeated the same user's earlier query, while only 7 % were typed by more than one user.

Commercial intent describes comparing and evaluating before a purchase, a stage between informational and transactional. The classic three-part taxonomy has no place for it, and SEO practice added it. Comparison content filed as a transaction page misses the person who is still weighing.

Investigational intent is the aim to understand a subject in depth, as distinct from looking up a fact. Gary Marchionini separated lookup, learning and investigation in exploratory search. A short answer widget cuts an investigation off at the first step.

A subjective knowledge gap is the gap a person believes they have, which can differ from the gap they actually have. People who see themselves as beginners read differently from experts. Content pitched at the wrong self-image loses both.

A post-purchase information need arises after a transaction: setting up, using, fixing or returning. The need does not end with the purchase. A returning buyer sent back to the sales page has been misread.

A user knowledge state is what a person already knows when they search. It decides how they read a snippet and what they ask next. The same content depth cannot fit a novice and an expert.

Task determinacy is how far a task has fixed rules for completion. Open tasks need discovery and structure, closed tasks need direct answers, as Katriina Byström and Kalervo Järvelin showed for task complexity. A system built for closed tasks fails open ones.

Information avoidance is the tendency to steer clear of information that could cause discomfort, such as a feared diagnosis. People search around a subject as much as into it. A page that confronts too early loses the person it was written for.

The need for cognitive closure, studied by Arie Kruglanski, is the desire for a firm answer and an end to uncertainty. It differs between people and situations. It shapes how long someone searches and which result they accept.

Information seeking anxiety is the unease people feel when searching is complex or unfamiliar, first described as library anxiety by Constance Mellon. It changes how people phrase queries and how fast they move. An overloaded interface adds to it.

Serendipitous knowledge discovery is the unplanned finding of valuable information while looking for something else. Sanda Erdelez called it information encountering. Intent models that assume a fixed target filter it out.

A task-based user model describes a searcher by role, task and situation, beyond the words of a query. It lets a system adapt to where the person is in their work. Keyword categories cannot carry that.

Information satisfaction is a person's own judgement of whether the need was met. Topical match and satisfaction often diverge. A result can fit the query perfectly and leave the need open.

Task completion time measures how long the whole task takes, beyond the time on one page. A short session can mean success or surrender. Only the task tells which.

The query reformulation rate is how often people change their query within a search. Each reformulation shows something the previous query did not carry. A high rate is the visible form of an unmet need.

If one query expressed the whole need, the session would end there. Nearly 40 % of all queries reworked the previous one, and these are the ways they did it.

Cognitive authority, a concept from Patrick Wilson in 1983, is the influence a source has on what a person accepts as true. People adopt answers from sources they trust. Topical relevance without authority is ignored.

Epistemic trust is a person's willingness to believe information from a source they cannot check. It decides whether a found answer is used. A need for knowledge is also a need for someone to rely on.

Cognitive dissonance, described by Leon Festinger, is the discomfort of holding beliefs that conflict with new information. In search, it shapes which results people open and which they reject. Relevant opposing views are often passed over for that reason.

Task workload is the mental effort a search demands, measurable with instruments such as the NASA Task Load Index. Dense result pages raise it. Effort spent sorting results is effort taken from the need.

Perceived task difficulty is how hard a person believes a task to be. It predicts how they search and when they give up. The same interface for easy and hard tasks helps neither well.

Cognitive friction is the resistance people meet when a system does not work the way they think, a term Alan Cooper made known. In search, it lies between the thought and the words the box will accept. Failed sessions are often friction, and they are counted as missing content.

Information exhaustiveness is the requirement of some tasks to see everything relevant, beyond the best few results. Legal research and systematic reviews work this way. A short summary answer fails a need for completeness.

The hedonic value of a search task is the enjoyment people get from searching and browsing itself. Not every session serves a goal. Recreational browsing misread as failed searching distorts every measure of success.

Search intent is read from the compromised form of the need, and content is built for that form. The person arrives with a larger need that moves, carries feelings and belongs to a task. Pages that serve only the query lose that person at the point where the need goes beyond the box.

Answer the user intent and carry the search intent inside it. A page can name the situations people arrive from, address the stage they are in and the worry behind the question, and ask back what brought them. That is the idea of Supervised Search: the publisher supervises the last stretch of the search.