UX research methods: which one to use
UX research methods compared: interviews, usability testing, surveys and analytics, and when to use each.

UX research answers two kinds of question: what are people doing, and why are they doing it? Analytics and surveys are good at the first. Interviews and usability testing are good at the second. Most projects need a bit of both.
This guide compares the common methods and shows when to use each.
The main methods
- User interviews: one-on-one conversations to understand needs, motivations and context
- Usability testing: watching people try to complete tasks with your product
- Surveys: structured questions to many people
- Analytics: data on what people actually do on your site or app
- Card sorting and tree testing: how people group and find information
- A/B testing: comparing two versions with real traffic
Which method answers what
- Explains why
- Measures how many
- Speed and cost
Our view, scored out of 5.
Choose by your question
- “What do customers need?” → interviews
- “Can people use this?” → usability testing
- “How many people have this problem?” → survey or analytics
- “Where do people drop off?” → analytics, then usability testing
- “How should we organise the menu?” → card sorting and tree testing
- “Which version performs better?” → A/B test
How many participants
For qualitative methods like interviews and usability tests, five to eight people per user group usually reveals the main patterns. For surveys and A/B tests, you need far more — often hundreds — for reliable numbers.
Plan a research round
- The decision this research will inform.
- User groups and how we’ll recruit them.
- Which method and why it suits the question.
- What we’ll ask people to do or talk about.
- Dates, incentives, tools.
- Who needs to see the findings and in what format.
Spam check by Cloudflare
How this looks in your industry
Illustrative examples
Closest to you
B2B
A logistics software company
The problem: Debated new features based on opinions.
What changed: Ran eight customer interviews before deciding.
Result: Dropped one planned feature and prioritised a different problem.
Closest to you
E-commerce
An online furniture store
The problem: Analytics showed drop-off on product pages.
What changed: Ran five usability tests to find out why.
Result: Found shoppers couldn’t find dimensions and delivery info.
Closest to you
Government
A local council website
The problem: Residents couldn’t find services.
What changed: Used card sorting and tree testing to rebuild the menu.
Result: Navigation matched how residents think.
Common mistakes
- Asking people what they want instead of observing what they do
- Using surveys to answer “why” questions
- Testing with colleagues instead of real users
- Doing research after decisions are made
- No plan for sharing findings
Key takeaways
- Interviews for why, analytics for what
- Match the method to the question
- Five to eight people per group for qualitative work
- Research before you design
Frequently asked questions
Can we do research on a small budget?
Yes. Five remote usability tests and a quick analytics review can be done in a week or two.
Should we pay participants?
Usually yes. A gift card or payment respects their time and improves show rates.
Who should run the research?
Someone neutral. Designers can run it, but avoid having people test their own work without guidance.
Your next step
Get answers before you build
We plan and run UX research that gives your team clear direction, fast.
Explore UX research