Write recruitment criteria, a voluntary consent plan and data minimization tied to the research question without making universal legal claims.
Recruitment determines what a study can reveal and what it will miss. Testing an appointment portal only with staff who know its terminology can conceal a new visitor's difficulty with labels. Start with people who actually perform the task, then set inclusion criteria connected to the question: first-time phone use, returning to change a booking, or navigating by keyboard. Age or job title is not automatically a proxy for the behavior to be understood. Convenient participants can help pilot a discussion guide, but cannot replace the usage variety relevant to the decision. Record who was excluded and why, as well as the interpretive gap that exclusion creates. A small exploratory set is not a promise of population representation. Its purpose is to understand failure mechanisms; use a broader measure later if the decision requires prevalence.
Consent should be informed and voluntary, not a box signed after recording starts. Before a session, explain the purpose, tasks, data to be collected, who will see them, retention under the organization's policy, and whether audio capture is optional. Distinguish participation from permission to record audio or screen. Refusing a recording should not remove a participant's ability to understand the study or withdraw. Collect the minimum: if a fictional booking number is enough to test the journey, do not request a real number or identity document. Use a participant code rather than a name in analytic notes, separate contact data from research excerpts and restrict access. Where a service is subject to a particular legal regime, consult the organization's privacy specialist rather than treating one country's service guidance as a universal legal rule.
Lesson summary
Consent should be informed and voluntary, not a box signed after recording starts.
