Framing UX Problems and Context of Use · Lesson 2

Map actors, tasks and context

4 min read

Lesson objectives

Create a context-of-use matrix that separates direct users from stakeholders and identifies the task and testable constraints.

Context of use is not a generic fictional persona. It is the relationship among a particular user, goal, tool, environment and constraints while a task is being completed. A receptionist may operate a booking system directly, while the person waiting for a notice is affected by the result and a manager owns the policy decision. Putting all three into one user box hides different needs. Begin with a task to study: confirm a reservation, correct details or recover an application number. Specify its entry point, completion criterion and resources available to the person. Record conditions that could change performance, such as a shared device, intermittent connection, interface language, keyboard use or help from another person. Do not attribute such conditions to a whole group without observation; treat them as recruitment hypotheses or questions to verify.

A useful matrix has a row for each materially different task, rather than a row for every job title. Its columns can be goal, decision step, channel, environment, dependency on another actor, known evidence and evidence gap. Write 'the participant needed a booking number before leaving' if a session showed it, but write 'a printable copy may be needed' if that remains unobserved. Ask whether the task changes with a lost connection or a language switch. If it does, a single successful path cannot represent the whole context. Recruit for the usage patterns relevant to the decision, not merely the people easiest to contact. A manager interview can reveal operating rules but usually cannot replace observing a customer's behavior. End the map with a small research decision: which actor and step to observe next, and what finding would change the design?

Example

Hypothetical training scenario in Jeddah: someone submits an application to a community-activity portal from a phone while travelling, then a centre employee reviews it on a desktop. The applicant's immediate goal is a tracking number before connectivity drops; the employee's goal is to identify missing documents. The context matrix gives the applicant a phone, possibly interrupted connection and need for confirmation, while leaving notification preference unproven. The employee has an application queue and permission to update file status, with a potentially busy desk. A manager decides document rules but cannot stand in for either user's experience. Test interrupted submission and draft recovery with applicants, then the visibility of document status with employees. Jeddah is only the setting for this invented case and implies no real system or local statistic.

Exercise

Exercise 1 — In the booking case, a volunteer uses a shared phone to change a visit date, while a centre employee records attendance. Produce two context-matrix rows: goal, task entry, completion test, one constraint and evidence still needed. Output two separate rows and one research question for each role.

Check your answer

Solution 1 — Volunteer row: start from the booking number; goal is a changed date; completion requires the new date and a confirmation number. The shared phone creates a privacy constraint. Observe how the booking is found and the session closed. Ask whether the person can recover the booking and confirm the change without exposing details to the next device user. Employee row: start from today's visit list; goal is recording attendance; completion is the updated visit status. A busy desk may constrain attention. Observe list ordering and correction of a mistaken entry. Ask how the employee distinguishes a cancelled visit from an attended one after a date change. Do not combine the two success measures because they concern different goals.

Exercise

Exercise 2 — A manager says, 'Everyone uses a desktop; I saw two do so.' The log shows six phone sessions and four desktop sessions, with possible repeat participants. Correct the map and claim, then select one usage pattern for the next round. Output an evidenced context constraint, an inference limit and a recruitment decision with a reason.

Check your answer

Solution 2 — Seeing two desktop users cannot support 'everyone'. The log records six phone and four desktop sessions, not six and four distinct people or population shares. Add 'phone was used in recorded sessions' to the matrix with its conditions still open, and retain the desktop path. For the next round, recruit people who reschedule on a phone, including an interrupted-connection case if that is the question, because the phone path appears in logs yet has not been examined. Keep a desktop comparison if the decision affects both channels. If reliable records later show that phone sessions were merely page views rather than task attempts, redefine the usage unit before treating them as completed work.

Lesson summary

A useful matrix has a row for each materially different task, rather than a row for every job title.

Report an error in this lesson

Lesson objectives

Lesson summary

Share this page