investigate

Structure the guide in three parts: a 5-10 minute opening with consent and rapport building, 30-40 minutes of open-ended behavior-focused core questions with probes, and a closing summary. Avoid leading questions, hypotheticals, and asking participants to design solutions.

Five participants per round typically uncover about 85% of usability issues. Run multiple small rounds rather than one large study, and pilot test every task with a colleague before the real sessions.

Use 100+ responses for descriptive statistics and 300+ for segment-level analysis. For group comparisons, aim for at least 30 responses per group, and always report your sample size alongside any percentage.

Use interviews for generative questions when you do not know what to ask yet, since they reveal motivations and mental models. Use surveys for evaluative questions when you have a hypothesis and need to measure prevalence at scale.

Use affinity mapping to cluster observations bottom-up without preset categories, or follow Braun and Clarke's six-step thematic analysis. Structure each insight as observation plus inference plus implication, and tag it with an evidence strength rating.

Small samples reveal patterns but not prevalence, so findings from 5-8 interviews should be presented as signals, not statistics. Triangulate across multiple methods and flag uncertainty rather than generalizing to all users.