A comparison of the most useful research methods for product teams, organised by the question each one can actually answer, with participant numbers and a simple way to start.
In short
UX research methods differ in whether they capture what people say or what they do, and whether they explain why or measure how many. Use interviews to understand needs, usability testing to see where a design fails, surveys and analytics to measure scale, and card sorting or tree testing to structure navigation. Pick the method from the question, not the other way round.
Teams often choose a research method because it is familiar: a survey because it is easy to send, or user testing because someone ran one last year. This comparison from Devies, a Gothenburg-based product and engineering consultancy, is for product owners, designers and managers who want their UX research to answer the question they actually have. It explains how methods differ, compares the five most useful ones and shows which to pick when.
What is UX research, and how do methods differ?
UX research is the systematic study of users, their goals and their behaviour, to inform what a product should do and how it should work. It covers everything from early interviews about needs to measuring how a released feature performs.
Christian Rohrer's framework at Nielsen Norman Group, when to use which user-experience research methods, sorts methods along three dimensions:
- Attitudinal or behavioural: what people say versus what people do. Surveys capture stated views; analytics and A/B tests capture behaviour.
- Qualitative or quantitative: qualitative methods are better suited to answering why, and how to fix a problem; quantitative methods answer how many and how much.
- Context of use: whether participants use the product naturally, follow a script, or do not use it at all during the study.
The same article maps methods to product phases. In the strategy phase, generative methods such as interviews and field studies find new directions. During design, formative methods such as usability testing, card sorting and tree testing improve a design. After launch, summative methods such as benchmarking, A/B testing and analytics measure performance.
How do the main UX research methods compare?
The five methods most product teams need are interviews, usability testing, surveys, analytics with A/B testing, and card sorting with tree testing. They differ mainly in what they reveal and what they cannot.
Criterion
Interviews
Usability testing
Survey
Analytics and A/B testing
Card sorting and tree testing
Answers
Interviews
What do people need, and why?
Usability testing
Where and why does the design fail?
Survey
How common is an attitude or need?
Analytics and A/B testing
What do people do, at scale?
Card sorting and tree testing
How do people group and find content?
Says or does
Interviews
Says
Usability testing
Does
Survey
Says
Analytics and A/B testing
Does
Card sorting and tree testing
Mostly does
Qualitative or quantitative
Interviews
Qualitative
Usability testing
Mostly qualitative
Survey
Quantitative
Analytics and A/B testing
Quantitative
Card sorting and tree testing
Both
Typical participants
Interviews
5 to 10 per group
Usability testing
About 5 per round
Survey
Hundreds
Analytics and A/B testing
All users
Card sorting and tree testing
15 or more for card sorting
Needs a design?
Interviews
No
Usability testing
Yes, a prototype or product
Survey
No
Analytics and A/B testing
Yes, a live product
Card sorting and tree testing
Only a content list or tree
Main limitation
Interviews
People misreport behaviour
Usability testing
Small samples, no statistics
Survey
Cannot explain why
Analytics and A/B testing
Cannot explain why
Card sorting and tree testing
Says little about visual design
No method covers everything. The strongest research combines one method that shows what happens with one that explains why.
Which method answers which question?
Choose the method by writing down the decision you need to make and the question behind it, then matching it to the method that answers that kind of question.
Choose interviews when you are early, the problem is unclear, and you need to understand goals, context and current workarounds. Interviews are the starting point for journey maps; our guide to customer journey mapping shows how to turn them into backlog items.
Choose usability testing when you have a prototype or a product and need to know whether people can complete key tasks. It is the most reliable way to find design problems before they reach production. User testing (användartester in Swedish) in most teams means exactly this. The GOV.UK Service Manual's guidance on moderated usability testing describes the core approach, and our guide to usability testing on a budget gives a one-week method.
Choose a survey when you already understand the issues and need to know how widespread they are, or want to track satisfaction over time. Write it after qualitative research, not instead of it.
Choose analytics or A/B testing when the product is live, you have enough traffic, and you need to know how many people do something or which of two versions performs better. Pair it with testing to learn why.
Choose card sorting or tree testing when the problem is structure: navigation labels, menus or how content is grouped. Card sorting generates ideas; tree testing checks a proposed structure. Our guide to information architecture covers both in context.
How many participants do you need?
The number of participants depends on whether you are looking for problems or measuring them: qualitative methods need few people, quantitative methods need many. Getting this wrong is the most common way to waste a research budget.
- Qualitative usability testing: about five users per round, repeated after changes, is the widely used guideline from Nielsen Norman Group for finding most problems in a focused test.
- Interviews: five to ten per distinct user group usually reveal the main patterns; stop when new interviews stop adding new themes.
- Card sorting: Nielsen Norman Group's card sorting guidance recommends at least 15 participants for a qualitative card sort and 30 to 50 for a quantitative one.
- Surveys and A/B tests: the number depends on the size of the difference you want to detect and the margin of error you accept. Calculate it before you start, and expect hundreds of responses or conversions rather than dozens.
A small qualitative study is not a lesser version of a large quantitative one. It answers a different question.
How do you start UX research with a small budget?
Start with the question that blocks your next decision, use the cheapest method that answers it, and make research a regular habit rather than a one-off project. These steps work for teams without a dedicated researcher.
- List open questions that the team argues about or guesses at, and rank them by the cost of being wrong.
- Reuse existing data first: support tickets, sales notes, analytics and earlier studies.
- Pick one method per question using the comparison above.
- Recruit from your own customers where possible, with consent and a small thank-you.
- Involve the team in sessions, so developers and stakeholders see users first-hand.
- Write a one-page summary with findings, evidence and decisions, and store it where the next team member will find it.
- Repeat monthly or per release, adding new questions as old ones are answered.
Keep reading
From our work
Client Case: Lea BankLea Bank rebuilt its digital platform with Strapi, React, GrowthBook, GA4, and BigQuery to improve SEO, A/B testing, and marketing autonomy.Read the caseDevies
Want a second opinion?
Devies, a Gothenburg-based product and engineering consultancy, works on questions like this with Nordic product and IT teams. Tell us where you are and we will suggest a sensible first step.
Frequently asked questions
What is the difference between UX research and market research?
Market research studies markets, segments and buying behaviour to inform business decisions such as pricing or positioning. UX research studies how people use, or would use, a product and where it helps or fails them. The two overlap in interviews and surveys, but UX research focuses on tasks, behaviour and design decisions rather than market size or brand perception.
Can we do UX research without real users?
Some methods work without users. Expert reviews against usability heuristics, analytics and support ticket analysis can reveal likely problems quickly. They cannot replace observing real users, because experts and data both miss problems that only appear when people bring their own goals, knowledge and context. Use expert methods as a first filter, then test the most important journeys with users.
Should UX research be moderated or unmoderated?
Moderated research, where a researcher guides the session, is better for exploring why people act as they do and for complex products. Unmoderated research, where participants complete tasks alone with a tool, is faster and cheaper at scale and suits simple, well-defined tasks. Many teams use moderated sessions to find problems and unmoderated tests to check fixes.
How often should a product team do UX research?
Small, regular research beats large, rare studies. A useful rhythm for many teams is a short usability test of key flows every few sprints, a handful of interviews whenever a new problem area opens, and continuous review of analytics and support data. The aim is that no significant design decision is made purely on internal opinion.
