UX Research Methods: 15 Approaches & When to Use Each

UX Research Methods: 15 Approaches & When to Use Each

UX research methods are the structured techniques teams use to learn about users — what they need, how they behave, and where current designs fall short. This guide covers 15 methods organized by what each one is best at, with sample sizes, time-to-result, and the specific question each method answers most cleanly.

The trick to using research well isn't running every method — it's matching the method to the question. The fastest signal usually comes from the method that exactly fits your specific question, not from the most thorough method.

The Two Axes of UX Research

Most strong research mixes axes. A typical sequence: generative qual → generative quant → evaluative qual → evaluative quant.

1. User Interviews

Question answered: Why do users behave the way they do? What's their mental model?

Type: Qualitative + generative

Sample size: 5-8 per segment for saturation

Time-to-result: 1-2 weeks (recruit + interview + analyze)

Best for: Discovery, understanding workflows, surfacing motivations the team doesn't know exist yet. Use behavior-grounded questions ("Walk me through the last time you…") instead of hypotheticals. See user interview questions for copy-ready prompts.

2. Contextual Inquiry

Question answered: What do users actually do in their environment, not what they say they do?

Type: Qualitative + generative

Sample size: 4-8 sessions

Time-to-result: 2-3 weeks (field logistics add time)

Best for: Surfacing workarounds and unconscious workflow steps. Observation in the real environment captures friction that interview self-report misses. Pair with "Can you show me?" prompts that force demonstration over description.

3. Jobs-to-be-Done (JTBD) Interviews

Question answered: What outcome is the user trying to achieve, beyond features they describe wanting?

Type: Qualitative + generative

Sample size: 8-12

Time-to-result: 2-3 weeks

Best for: Reframing feature decisions around outcomes. JTBD shifts focus from "users want X feature" to "users are trying to accomplish Y outcome." See JTBD framework.

4. Diary Studies

Question answered: What does the user experience look like across days or weeks, in their own moments?

Type: Qualitative + generative

Sample size: 10-20 over 5-7 days

Time-to-result: 2-3 weeks (fieldwork dominates)

Best for: Capturing in-the-moment context that interviews lose to recall bias. Keep prompts short (1-2 sentences). Light + long beats heavy + short for diary participation.

5. Card Sorting

Question answered: How do users group and categorize your content? What does their mental model of your information look like?

Type: Qual + generative (open sort) or quant + evaluative (closed sort)

Sample size: 15-30 per segment

Time-to-result: 1 week

Best for: Information architecture decisions before design exists. Open sorts surface user-natural categories; closed sorts validate an existing IA. See card sorting.

6. Tree Testing

Question answered: Does the navigation structure work when stripped of visual design?

Type: Quantitative + evaluative

Sample size: 30-50

Time-to-result: 3-5 days

Best for: Validating an IA before it ships into visual design. Tree tests strip away color and hierarchy so labels and structure get tested in isolation. Run tree tests free with unlimited responses.

7. First-Click Testing

Question answered: Does the visual layout direct users to the right element?

Type: Quantitative + evaluative

Sample size: 30-50

Time-to-result: 3-5 days

Best for: Catching layout problems on screenshots before development. Users who click the right area first succeed at the overall task ~87% of the time. Run first-click tests on uploaded screenshots — no Figma prototype required.

8. Usability Testing (Moderated)

Question answered: Can users complete real tasks? Where do they get stuck and why?

Type: Qualitative + evaluative

Sample size: 5-8

Time-to-result: 1-2 weeks

Best for: Surfacing specific friction with participant quotes. Use think-aloud probes. Measure SEQ after each task; SUS at the end. See usability testing questions.

9. Usability Testing (Unmoderated)

Question answered: Same as moderated — but at scale and lower cost.

Type: Quantitative + evaluative (when N ≥ 30) or qual (when smaller)

Sample size: 30-100

Time-to-result: 1-2 weeks

Best for: Validation at scale once you've nailed the test design. Cheaper and faster than moderated; the tradeoff is no real-time probing.

10. Surveys

Question answered: What do users think, prefer, or do? — at scale.

Type: Quantitative + either generative or evaluative

Sample size: 100-500 for stable rates

Time-to-result: 1-2 weeks

Best for: Measuring magnitude at scale; benchmarking with standardized instruments like SUS or NPS. See UX survey questions.

11. A/B Testing

Question answered: Which of two designs performs better on a specific metric?

Type: Quantitative + evaluative

Sample size: 30-50 per variant for prototype-stage; 1,000+ per arm for production A/B with subtle metrics

Time-to-result: Hours to weeks (depends on traffic + effect size)

Best for: Resolving design debates with quantitative answers. Pair with a small qualitative study to understand the why behind the win.

12. Preference Testing

Question answered: Which of two or more designs do users prefer and why?

Type: Mix qual + quant

Sample size: 20-30

Time-to-result: 3-5 days

Best for: Early-stage design exploration when you're choosing between visual directions. Faster than A/B testing because no task completion is required. Note: preference ≠ performance — pair with usability or first-click tests for high-stakes decisions.

13. Concept Testing

Question answered: Does the underlying idea resonate before you build anything visual?

Type: Mix qual + quant

Sample size: 50-100

Time-to-result: 1-2 weeks

Best for: Validating demand for a concept before design or development investment. See concept testing questions.

14. Heuristic Evaluation

Question answered: Does the design follow established usability principles?

Type: Expert-driven qual + evaluative

Sample size: 3-5 evaluators (not users)

Time-to-result: 1 week

Best for: Quick expert review before user testing. Catches obvious issues that don't need user data to spot. Use Nielsen's 10 Usability Heuristics. See heuristic evaluation.

15. Analytics + Behavioral Data

Question answered: What are users actually doing at scale?

Type: Quantitative + either generative or evaluative

Sample size: All users (or representative sample)

Time-to-result: Real-time or days

Best for: Identifying what's happening (drop-off rates, funnel conversion, feature adoption). Pair with qualitative methods for the why — analytics never tells you why a 30% drop-off happens, only that it does.

How to Pick the Right Method

Start with the question, not the method. A common decision tree:

When your question doesn't fit one method cleanly, combine two. Generative qual + evaluative quant is the most common pairing: interview 8 users to discover the problem, then survey 200 to measure its prevalence.

Plan a Research Sequence

A practical product cycle uses 3-5 methods in sequence:

1. Discovery — user interviews + JTBD → understand the problem
2. IA validation — card sort → confirm content groupings
3. Layout validation — first-click test on wireframes → confirm visual hierarchy works
4. Usability — moderated usability testing on hi-fi prototype → surface friction
5. Launch validation — survey + analytics post-launch → measure actual outcomes

Each method takes days, not weeks. Skipping early methods to "save time" usually costs more time later when fundamental issues surface in usability testing — when they're 10x more expensive to fix.

Run Multiple Methods in One Workspace

ValidateThat supports user interviews, card sorts, tree tests, first-click tests, and surveys in a single workspace. Free plan covers 3 studies with unlimited responses across all methods — so the follow-up test after a research finding lives in the same project.

Start research free →

Further Reading

• 25 Qualitative Research Questions by Method
• 30 Concept Testing Questions: Ready-to-Use Survey Bank
• 24 Usability Testing Questions: Before, During & After
• Prototype Testing: 6 Methods, When to Use Each
• User Research (UX Glossary)
• UX Research Methods (UX Glossary)

Frequently Asked Questions

What are the main types of UX research methods? UX research methods divide along two axes: qualitative vs quantitative (stories + themes vs measured metrics) and generative vs evaluative (discover what to build vs test what exists). User interviews are qual + generative. Surveys can be either gen or eval depending on use. Usability testing and tree testing are evaluative; qual/quant depends on sample size.

What's the difference between qualitative and quantitative UX research? Qualitative UX research surfaces the why through methods like interviews, contextual inquiry, and card sort debriefs. Quantitative UX research measures how much, how many, or how often through methods like surveys, A/B tests, and unmoderated usability tests at scale. Most strong research mixes both.

What's the difference between generative and evaluative research? Generative research discovers what to build. Evaluative research tests something that already exists. Generative comes first; evaluative comes later when there's something to test.

How many participants do I need for UX research? Qualitative methods: 5-12 (Nielsen's 5 catches ~85% of usability issues). Quantitative methods: 30-200 depending on effect size. Tree tests and first-click tests typically use 30-50.

How do I pick the right UX research method? Match the method to the question. The fastest signal usually comes from the method that exactly matches your specific question — not from the most thorough method.