UX Research Methods: Which One to Use and When
Basics4 minInnovialabs editorial team
In short
Pick the method by the question: "why" and "what people think" call for interviews and diary studies, "how many" and "how often" for surveys and analytics, "can people use it" for usability tests, and "how to structure navigation" for card sorting.
The short answer
There is no best UX research method, only the method that answers your question. So start not with a tool but with a statement: what exactly do we want to learn, and what decision will we make based on the result?
- To understand reasons and motives, use qualitative methods: interviews, diary studies.
- To learn scale and frequency, use quantitative ones: surveys, product analytics.
- To check whether people can handle the interface, run usability tests.
- To shape structure and navigation, use card sorting.
Two axes behind every method
Qualitative vs quantitative. Qualitative methods answer "why" and "how": few participants, deep observations. Quantitative methods answer "how many" and "how often": lots of data, little explanation.
Attitudinal vs behavioral. Some methods capture what people say and think, others what they do. There is often a gap between the two: someone is sure they always read the product description, while analytics show they go straight to "Buy".
| Method | Qualitative / quantitative | Attitudinal / behavioral |
|---|---|---|
| Interviews | Qualitative | Attitudinal |
| Surveys | Quantitative | Attitudinal |
| Usability tests | Mostly qualitative | Behavioral |
| Card sorting | Either | Attitudinal |
| Product analytics | Quantitative | Behavioral |
| Diary studies | Qualitative | Both |
Good research often combines methods from different corners: analytics shows where the problem is, interviews and tests show why.
Each method in brief
- Interviews. One-on-one conversations about tasks, context and experience. Great at the start, when you need to understand the audience. Ask about past experience, not a hypothetical future.
- Surveys. Confirm or reject hypotheses on a large sample. Poor at uncovering new ideas: you only get answers to the questions you asked.
- Usability tests. A user performs tasks in a prototype or product while you observe. They can be moderated or unmoderated. Even a few sessions usually reveal the most visible problems.
- Card sorting. Participants group cards with menu or content items. In an open sort they name the groups themselves; in a closed sort they use predefined ones. Useful when designing navigation.
- Product analytics. Funnels, clicks, time on step, drop-offs. Shows real behavior at scale but does not explain causes.
- Diary studies. Participants record their experience over several days or weeks. A fit for scenarios stretched over time: habits, repeat purchases, using a service throughout a week.
Choosing by project stage
| Stage | Key question | Suitable methods |
|---|---|---|
| Problem discovery | Who are our users and what gets in their way? | Interviews, diary studies, review of existing data |
| Structure and concept | How should content and features be organized? | Card sorting, interviews |
| Prototype | Is the interface clear? | Usability tests on the prototype |
| After launch | What actually happens, and where do we lose people? | Analytics, surveys, usability tests |
Choosing by budget
Cost depends less on the method itself than on the number of participants, recruiting difficulty and depth of analysis.
- Minimal budget. A few interviews or usability tests with reachable audience members, a review of the analytics you already have, a short survey of your own user base.
- Medium budget. Regular prototype tests, online card sorting, recruiting through specialized services.
- Large budget. Diary studies, research across several segments or regions, a mix of qualitative and quantitative methods.
Even with modest resources, a small study beats no study.
Common mistakes
- Picking a method before the question. "Let's run a survey" without knowing what to do with the answers.
- Leading questions. "This is convenient for you, right?" gives a predictable and useless result.
- Words instead of behavior. Decisions are based only on opinions, without checking actions.
- Research for the sake of a report. Findings never turn into concrete product changes.