Contextual Inquiry

Contextual inquiry is a research method in which the researcher observes and interviews people while they do real work in their own environment. Developed as part of the Contextual Design methodology by Hugh Beyer and Karen Holtzblatt, it combines observation with in-the-moment questioning, and rests on the premise that people cannot accurately describe their own work from memory, but can explain it while doing it.

The method is structured around a master-apprentice relationship. The researcher takes the position of an apprentice learning the work from the person doing it, which changes the dynamic from an interview, where the participant summarizes and rationalizes, to a demonstration, where they show and narrate. This surfaces the details that summaries omit: the spreadsheet maintained alongside the official system, the printed notes taped to a monitor, the colleague consulted before every exception, the steps performed in a different order than the documented process specifies.

What it reveals is typically invisible to every other method. Workarounds are the most valuable category, because each one marks a place where the system fails to support the actual work, and because people stop noticing their own workarounds once they become habitual. Interruptions, environmental constraints, parallel tools, and the social context of decisions are similarly absent from interviews and from analytics, and they frequently explain why a well-designed feature goes unused or why a metric behaves in ways nobody can account for.

It is most valuable when the domain is unfamiliar to the design team, which is the normal situation in business software, operational tools, and specialist services. A team building software for warehouse staff, clinicians, insurance assessors, or call centre agents cannot design competently from a requirements document, because the document reflects how the process is supposed to work rather than how it does. A few days of observation typically reorders the priorities of a project more than months of stakeholder workshops.

The costs are real. Sessions are lengthy, access requires negotiation with organizations and sometimes with regulators, travel and scheduling are expensive, and analysis of rich observational data is slow and skilled work. Participants may also alter their behavior when watched, and confidentiality constraints in regulated environments can restrict what may be observed or recorded. These constraints mean the method is used selectively, with a small number of sessions chosen for maximum diversity of context rather than for statistical representation.

A lighter version of the method is available to organizations that cannot arrange full field visits, and it is considerably better than nothing. Asking a customer to share their screen and walk through a task they were going to perform anyway, with the researcher observing and asking questions at natural pauses, captures a useful proportion of what a site visit would reveal, particularly the workarounds and the parallel tools. It misses the physical environment, the interruptions, and the social context, which matters more in some domains than others. For software used at a desk, remote observation is a reasonable substitute; for work performed in a warehouse, a clinic, or a vehicle, it is not.

Analysis usually involves building models of the work as it actually happens, covering sequence, communication flow, physical environment, and the artifacts people create for themselves, and these models then drive design decisions directly. In practice contextual inquiry sits at the front of engagements where the problem space is not already understood, feeding product research and shaping what gets built in product design. It is disproportionately valuable in health and other regulated or operationally complex sectors, where the gap between documented process and real practice is widest and the cost of designing for the documented version is highest.