Conduct neutral discovery and synthesize evidence

After mapping the room-booking journey, the specialist sees that many calendar viewers did not submit a request. Asking “Wasn't the calendar confusing?” may produce polite agreement rather than an account of what happened. A discovery interview is not a vote on a solution already chosen by the team. Its purpose is to reconstruct a recent real episode in the participant's words while keeping observation separate from interpretation. Begin with a specific occasion: “Tell me about the last time you wanted to book a room. Where did you start? What happened after you saw the available times?” Follow with actions: “What did you select, and what did you expect next?” Do not present a redesigned page before hearing the old path; that would test stated preference and politeness instead of past behavior.

Prepare a short interview guide around one decision. Start with context, ask three questions about the last episode, then ask about the result and any workaround. A follow-up may request a description of a screen or message, but it should not supply the cause you favor. “Did you abandon the request because 'provisionally available' was unclear?” embeds a cause in the question. “When you saw that phrase, what did it mean to you?” tests understanding without supplying an answer. If someone says “the calendar is bad,” ask what they tried to do and what blocked them. The adjective alone is not a design decision.

Keep three columns in the note: quoted words or observed action, a possible interpretation, and a later check. In a fictional case, Reem says she saw a green time slot and then “provisionally available,” assumed that booking was uncertain, but submitted anyway. Sami stopped at a membership requirement and never opened the calendar. Huda opened the calendar twice to compare days, then submitted. Do not code all three as “confused by calendar.” Reem points to uncertainty about confirmation, but it did not stop her submitting; Sami encountered eligibility earlier; Huda's repeat opening may be ordinary comparison, not lost demand. This reading also prevents treating calendar openings as distinct people.

To synthesize interviews, create a matrix with participants as rows and journey stages as columns. Put a specific act or phrase in a cell, not a generic tick mark. Add a column for contradictory or exceptional evidence. A fourth participant might find the wording clear but receive confirmation late. Do not erase disagreement by majority vote. Check whether participants saw the same version under comparable conditions. If one used a phone and another a desktop computer, interface differences form a testable explanation. Repetition in a small qualitative sample guides the next inquiry; it is not a population prevalence estimate. Four interviews cannot justify saying “75% of users are confused,” even if three mention similar trouble.

Before interviewing, clarify operational and ethical boundaries. What purpose has been explained to participants? Who can see the notes? Is a recording necessary at all, and who authorizes it? Follow the organization's consent and retention policy rather than collecting sensitive material simply because it might help analysis. If accessibility or personal information arises, do not seek medical details; describe the service obstacle and refer it to the appropriate team. The specialist can organize evidence under a mandate, but cannot invent a privacy policy or promise a remedy the organization has not approved.

Consider who is missing from the interviews. If only successful bookers were recruited, people who could not find the link or stopped at membership are absent. This selection bias is not fixed merely by asking the same group more questions. Seek an authorized way to include people who stopped at different stages, or observe sessions beginning before the calendar. When a complaint repeats, check whether it comes from independent experiences or one story recirculated by several people. Repetition of words does not imply many separate incidents. Mark any important unrepresented group as a coverage gap. In the synthesis matrix, keep a column for when the problem did not occur; an exception may point to device, booking type or staff practice. Neither suppress contrary evidence to simplify the story nor promote one exception into a universal rule. Report that the obstacle appeared in defined cases and name a check that separates rival explanations.