UX Research

Workflow

1. Define the research questionUnfocused research generates mountains of data with no actionable insight. A specific question keeps the study manageable and the findings usable.
2. Conduct user interviewsInterviews uncover motivations, mental models, and context that you'd never discover from usage data alone. They're the highest-signal research method.
3. Run moderated usability testsWatching real users attempt your tasks reveals usability issues invisible to designers who know the product too well.
4. Design and analyze surveysSurveys reach hundreds of users at once - validating whether insights from small qualitative studies are widespread.
5. Synthesize findings into insightsRaw interview transcripts and usability notes aren't actionable. Synthesis extracts patterns, root causes, and design implications.

Output Format

UX research deliverable: research question → method → findings → insights → design implications.

Research Report:
├── Research question and objectives
├── Method and participant profile
├── Key findings (themes with supporting quotes)
├── Insights (why users behave this way)
├── Design implications (what to change)
└── Appendix: interview guide, survey questions, raw data

Examples

Edge Cases & Gotchas

• Recruiting the wrong users - Testing with power users who know your product well masks issues that new users face. Recruit participants at the right stage (new, mid-tenure, power) for your research question.
• Confirmation bias - Researchers unconsciously look for evidence that confirms their existing hypothesis. Use a note-taker, share raw recordings with stakeholders, and actively look for disconfirming evidence.
• Sample size overconfidence - 5 interviews reveal usability issues but can't represent all user segments. Never extrapolate interview findings to 'all users' - use surveys for that.
• Research without follow-through - Research that doesn't influence product decisions creates research debt and cynicism. Connect findings to the backlog before starting new research.

When to Break These Rules

• Use analytics (funnel analysis, session recordings) instead of interviews when the behavior is already measurable - 'where do users drop off?' is a quantitative question.
• Skip research when you're in a time crunch and the decision is low-stakes and reversible - ship, measure, iterate. Research effort should match decision stakes.

Reference Files

• https://www.nngroup.com/articles/
• https://www.usability.gov/how-to-and-tools/methods/