There is no single "correct" UX research method. The right method depends on what question you are trying to answer, how much time you have, and where you are in the product development cycle. This guide breaks down the most important methods and tells you when to use each.
The Two Dimensions: Qualitative vs Quantitative
Qualitative research gives you depth. It helps you understand the why behind user behaviour, the motivations, mental models, frustrations, and workarounds users develop when products fail them. The sample sizes are small (typically 5–15 participants), but the richness of insight is high.
Quantitative research gives you scale. It measures behaviour across large populations and tells you how many users do something, how often, and how successfully. It validates or disproves patterns found in qualitative work. Survey data and analytics fall here.
The most effective product teams use both in combination, qualitative research to understand behaviour, quantitative to measure its scope.
Generative Methods (Understanding Problems)
User interviews are the most versatile qualitative method. One-on-one conversations, typically 45–60 minutes, that explore user goals, mental models, workflow, and pain points. Best used when you need to understand your users deeply before designing anything. Even five well-structured interviews can surface the most critical insights.
Contextual inquiry involves observing users in their actual environment, at their desk, in their workflow, using their real tools. You see workarounds and adaptations that users never mention in interviews because they have stopped noticing them. Invaluable for enterprise and workplace tools.
Diary studies ask users to log their experiences over days or weeks as they naturally encounter the problem your product addresses. They surface longitudinal behaviour and emotional patterns invisible in single-session research.
Evaluative Methods (Testing Solutions)
Usability testing is the most important evaluative method. A participant tries to complete specific tasks using your product or prototype while you observe. You discover where users hesitate, make errors, or give up, insights that drive targeted improvements before launch. Moderated sessions (with a researcher present) give the richest data; unmoderated sessions (remote, asynchronous) give broader coverage faster.
Cognitive walkthroughs are expert evaluations where a researcher steps through the interface imagining the questions a first-time user would ask at each step. Faster than usability testing, they are best for early-stage designs or when participant access is limited.
A/B testing compares two design variants with real users at scale to see which performs better against a specific metric, conversion rate, task completion, click-through. It answers "which works better" but not "why." Use it to validate changes, not to discover what to change.
Surveys collect self-reported data from many users simultaneously. The Net Promoter Score (NPS) question ("How likely are you to recommend?") is one of the most common survey instruments. Surveys are limited by what users are aware of and willing to report, but at scale they provide statistically reliable signals.
Choosing the Right Method
Ask yourself three questions. First: am I trying to understand or measure? Understanding → qualitative; measuring → quantitative. Second: do I have a design to test, or am I still discovering the problem? Discovery → generative methods; testing → evaluative methods. Third: how much time and access do I have? A two-day sprint with five participants favours interviews plus a rapid usability test. A six-week programme allows contextual inquiry, diary studies, and multiple rounds of testing.
The 80/20 rule of research: Five usability test participants will reveal roughly 80% of the major usability issues in a design. You rarely need more than eight for a single round of evaluative testing. Recruit the right people, not the most people.
Key Takeaways
- Qualitative research reveals why users behave as they do; quantitative research measures how many and how often.
- Generative methods (interviews, contextual inquiry) are for problem discovery; evaluative methods (usability testing, A/B) are for solution testing.
- Match the method to the question, not the other way around.
- Five usability testing participants surface the majority of critical usability issues.
- Combining methods, qualitative insight with quantitative validation, produces the most reliable evidence.
