UX Research Methods: 15 Methods, When To Use Each And How To Choose

Short answer: UX research methods are the structured ways teams learn what users need, say and do, from user interviews and diary studies to usability testing, card sorting and A/B tests. Pick a method by the question you need answered: talk to people when you need to understand problems, watch people use a design when you need to find what breaks, and measure at scale when you need to know how much or how often. For most product teams, interviews, usability testing, surveys and analytics cover the bulk of everyday questions.

Choosing among UX research methods is less about knowing every technique and more about matching a method to a decision. The wrong method wastes weeks and produces confident but irrelevant answers.

Below are 15 user research methods grouped by product stage, a master table, a question-to-method lookup, research ops basics, time budgets and when to hire outside help.

• What Are UX Research Methods And How Are They Organized?
• All 15 Methods At A Glance
• Discover: Methods For Understanding Users
• Define And Explore: Methods For Structure And Concepts
• Test: Methods For Evaluating Designs
• Listen And Measure: Methods For Live Products
• How To Choose A Method By Question
• Research Ops Basics And Time Budgets
• When To Hire A UX Research Agency
• Frequently Asked Questions
• The Short Version

What Are UX Research Methods And How Are They Organized?

UX research methods are repeatable techniques for collecting evidence about users: their goals, their context, their behavior with a product and their reactions to it. The most widely used way to organize them comes from Nielsen Norman Group. Christian Rohrer’s article When to Use Which User-Experience Research Methods maps 20 methods on a “landscape” defined by three dimensions. Two of them do most of the work in practice.

Attitudinal Vs Behavioral

Attitudinal methods capture what people say: their opinions, stated preferences, beliefs and self-reported habits. Surveys, interviews and focus groups sit here. Behavioral methods capture what people do: where they click, where they get stuck, what they actually complete. Usability testing, analytics and A/B tests sit here.

The two often disagree: people say they would use features they never touch. When a decision is expensive, back attitudinal findings with behavioral evidence.

Qualitative Vs Quantitative

Qualitative methods observe people directly and answer “why” and “how”. They use small samples and produce insights, quotes and problem lists. Quantitative methods gather data indirectly, through an instrument such as a survey or analytics, and answer “how many” and “how much”. They need larger samples and produce numbers you can compare over time.

Qualitative vs quantitative UX research is about the kind of answer you need: a qualitative test shows that people misread a pricing toggle; a quantitative one shows what share fail and whether a redesign moved that number.

Generative Vs Evaluative Research

A separate, simpler framing splits research by purpose. Generative research (also called discovery or exploratory research) uncovers problems and opportunities before you have a solution. Evaluative research tests a solution you already have, from a sketch to a shipped product. NN/g’s article ties methods to three product phases along the same lines: strategize (generative), design and execute (formative), and launch and assess (summative).

Generative vs evaluative research is the fastest filter for picking a method. If you do not yet know what to build, you need generative methods. If you are deciding whether something works, you need evaluative ones.

All 15 Methods At A Glance

The table below summarizes each method. Sample sizes are planning rules of thumb, not statistical guarantees; quantitative studies should be sized with a power or margin of error calculation for your specific metric. Effort reflects a typical in-house study including recruiting, running sessions and analysis.

Method	Type	Best Stage	Sample Size Guidance	Typical Effort
User interviews	Attitudinal, qualitative	Discover	5 to 12 per user segment	1 to 3 weeks
Contextual inquiry	Behavioral and attitudinal, qualitative	Discover	4 to 8 sessions	2 to 4 weeks
Diary studies	Attitudinal and behavioral, qualitative	Discover	8 to 15 participants (allow for dropouts)	3 to 6 weeks
Focus groups	Attitudinal, qualitative	Discover	2 to 4 groups of 5 to 8	1 to 3 weeks
Card sorting	Attitudinal, qualitative or quantitative	Define and explore	About 15 for open sorts; more for closed	1 to 2 weeks
Tree testing	Behavioral, mostly quantitative	Define and explore	A few for pilots; dozens for metrics	1 to 2 weeks
First-click testing	Behavioral, quantitative	Define and explore	Dozens per design variant	Days to 1 week
Concept testing	Attitudinal, qualitative	Define and explore	5 to 10 per concept round	1 to 2 weeks
Usability testing	Behavioral, qualitative or quantitative	Test	About 5 per round (qualitative); 20 or more for metrics	1 to 2 weeks per round
Heuristic evaluation	Expert review, no users	Test	3 to 5 evaluators	Days to 1 week
Eye tracking	Behavioral, qualitative or quantitative	Test	5 to 8 (qualitative); far more for heatmaps	2 to 4 weeks
Accessibility testing with users	Behavioral, qualitative	Test	A few per assistive technology or disability group	2 to 4 weeks
Surveys	Attitudinal, quantitative	Listen and measure	Set by margin of error; often hundreds	1 to 3 weeks
Analytics review	Behavioral, quantitative	Listen and measure	All tracked traffic	Days, if tracking exists
A/B testing	Behavioral, quantitative	Listen and measure	Set by power calculation; usually thousands of visitors	2 to 6 weeks of traffic

The five user guidance for qualitative usability testing and the roughly 15 participant guidance for card sorting both come from Jakob Nielsen’s article Why You Only Need to Test with 5 Users, which also recommends about 20 users for quantitative usability studies.

Discover: Methods For Understanding Users

Discovery methods are generative. Use them before committing to a solution or when metrics show a problem but not why.

User Interviews

User interviews are one-on-one conversations, usually 30 to 60 minutes, about a person’s goals, workflows, frustrations and past decisions. They are the default starting point because they are cheap, flexible and work remotely.

Ask about specific past events rather than hypotheticals, because people describe what happened far better than they predict what they would do. Stop recruiting when new interviews in a segment stop surfacing new themes.

Contextual Inquiry

Contextual inquiry combines observation and interview in the user’s real environment. The researcher watches someone do their actual work and asks questions as it happens, in a master and apprentice style. It surfaces workarounds, sticky notes, spreadsheets on the side and interruptions that never come up in an interview. It is especially valuable for B2B and operational software, at the cost of travel and workplace access.

Diary Studies

A diary study asks participants to log experiences over days or weeks, often with short prompts, photos or screen recordings sent through a mobile tool. Use it for behavior that unfolds over time: habit formation, onboarding over the first month, travel, health routines or recurring work tasks. Keep entries short, send reminders, and over-recruit because some participants drop out.

Focus Groups (With Caveats)

Focus groups bring 5 to 8 people together for a moderated discussion. They come from market research and are good at surfacing language, attitudes and reactions to messaging or brand ideas. They are poor at usability questions: group dynamics push people toward agreement, a vocal participant can steer the room, and nobody is actually using the product. Never use one to judge usability, and treat “would you buy this” answers with suspicion.

Define And Explore: Methods For Structure And Concepts

Once you understand the problem, these methods help shape the information architecture and pick between directions before detailed design starts.

Card Sorting

In card sorting, participants group labeled cards (content items or features) into categories. In an open sort they create and name their own groups, which reveals their mental model. In a closed sort they place cards into categories you define, which checks whether your proposed structure makes sense. It tells you how people think content should be grouped, not whether they can find things in the final navigation. That is tree testing’s job.

Tree Testing

Tree testing, sometimes called reverse card sorting, gives participants a text-only version of your navigation hierarchy and asks them to find where specific items live. Because there is no visual design, it isolates the labels and structure. NN/g’s guide to tree testing notes that a small qualitative round can yield useful insights, while quantitative tree testing needs a larger sample, and recommends piloting with a small qualitative study first. Pair it with card sorting: sort to generate the structure, tree test to validate it.

First-Click Testing

First-click testing shows participants a screen (a wireframe, mockup or live page) and a task, then records where they click first. It answers a narrow but important question: does the layout point people in the right direction? It is fast, unmoderated and works on static Figma frames, so it is a cheap way to compare two layouts before building either.

Concept Testing

Concept testing puts an early idea in front of target users, usually as a sketch, storyboard, landing page or rough prototype, and explores whether it addresses a real need and how people understand it. Show two or three concepts so participants compare instead of politely approving; stated purchase intent is weak evidence.

Test: Methods For Evaluating Designs

Evaluative methods find what breaks in a design. Run them in small rounds between iterations, not one big study before launch.

Usability Testing

Usability testing asks representative users to complete realistic tasks while thinking aloud, and the researcher watches where they hesitate, misread or fail. It can be moderated or unmoderated, and it works on anything from paper prototypes to production software. For planning, scripting and analysis in depth, see our usability testing guide.

Heuristic Evaluation

In a heuristic evaluation, a few experienced evaluators independently review an interface against a set of principles, most often Jakob Nielsen’s 10 usability heuristics, developed with Rolf Molich in 1990 and refined in 1994. It involves no users, so it is fast and cheap, but it catches rule violations rather than real-world confusion. Our heuristic evaluation guide walks through the process step by step.

Eye Tracking

Eye tracking records where participants look and for how long, using screen-based or wearable hardware. It shows whether people see key content at all, which is something clicks and verbal reports cannot tell you. It is also expensive and slow, and heatmaps need much larger samples than a standard usability test to be stable. For most product teams, a well-run usability test answers the same question at a fraction of the cost.

Accessibility Testing With Users

Automated checkers and WCAG conformance audits find many technical issues but not all of them. The W3C’s guidance on involving users in evaluating web accessibility notes that evaluating with people with disabilities and older users identifies issues conformance evaluation alone does not, and cautions against assuming one person’s input applies to all people with disabilities. Recruit people who use screen readers, magnification, voice control or keyboard-only navigation, let them use their own setups, and combine sessions with an expert audit.

Listen And Measure: Methods For Live Products

Once something ships, these methods show how it performs at scale. Pair them with a qualitative method that explains the numbers.

Surveys

Surveys collect structured answers from many people at once. They are good for measuring satisfaction, sizing how common a problem is, segmenting users and tracking attitudes over time. They are bad at discovering problems you did not already think to ask about. Write surveys after interviews so answer options reflect real user language, avoid leading or double-barreled questions, and work out the responses you need for your target precision before launch.

Analytics Review

An analytics review mines product and web analytics (funnels, retention, feature adoption, search logs, session recordings) to find where users drop off or behave unexpectedly. If tracking exists, it is the cheapest research you can do and the best way to aim qualitative work. Analytics tells you what happened, never why. A sharp drop at one checkout step is a reason to run a usability test on that step, not a diagnosis.

A/B Testing

A/B testing randomly splits live traffic between two or more versions and compares a predefined metric, such as sign-up rate. It is the strongest evidence of causal impact you can get in UX work. But it needs enough traffic, only compares options you thought of, and says nothing about why a variant won. Fix the metric and run length in advance, and do not stop early because one variant looks ahead.

How To Choose A Method By Question

Start with the decision you need to make, write it as a question, and pick the cheapest method that can answer it with enough confidence for the stakes. This lookup covers the questions product teams ask most often.

Your Question	Start With	Add If Stakes Are High
What problems do our users have?	User interviews	Contextual inquiry
How does this behavior change over weeks?	Diary study	Analytics cohort review
How common is this problem across our users?	Survey	Analytics review
How should we group and label our content?	Open card sort	Tree test
Can people find things in our navigation?	Tree test	Usability testing on the live menu
Does this layout send people the right way?	First-click test	Usability testing
Which direction should we pursue?	Concept testing	Prototype usability testing
Can people complete key tasks?	Usability testing	Quantitative benchmark study
What are the obvious issues before we test?	Heuristic evaluation	Usability testing
Do people notice this element at all?	Usability testing	Eye tracking
Can people with disabilities use it?	Expert accessibility audit	Accessibility testing with users
Where are users dropping off?	Analytics review	Usability testing on that step
Did this change improve the metric?	A/B test	Follow-up interviews on the result
How do people react to our messaging?	Interviews or concept testing	Focus groups or a survey

Three rules help when the table is not enough:

1. Match the method to the phase. Early on, generative and qualitative. Mid-design, evaluative and qualitative. After launch, quantitative with qualitative follow-up.
2. Triangulate on big decisions. Pair one attitudinal and one behavioral method, or one qualitative and one quantitative, so a single method’s blind spot does not drive the decision.
3. Prefer several small studies over one large one. Five users, fix, five more users finds more problems than fifteen users at once, which is the core argument of Nielsen’s five user article.

Our product design process guide shows where each of these methods plugs into discovery, design, delivery and iteration.

Research Ops Basics And Time Budgets

Most research delays come from recruiting, scheduling, legal review and lost findings, not the method. A little research operations work removes most of that friction.

Recruiting

Write a short screener that filters on behavior (“used a project management tool weekly in the last three months”) rather than demographics alone. Recruit from your own customer base for existing products, and use a recruiting panel or agency for niche B2B roles or people who are not yet customers. Over-recruit by one or two people per round to cover no-shows, and pay higher incentives for specialist professionals.

Consent And Data Handling

Every study needs informed consent that explains what you are recording, how it will be used, who will see it, how long it is kept and how participants can withdraw. Have a template reviewed once by legal, reuse it, restrict access to recordings, and follow the privacy rules for your participants’ locations.

A Research Repository

A repository is a searchable home for findings, so the next team does not re-run the same study. Whether it is a dedicated tool or a shared drive, tag each insight by product area, user segment, method and date, and link it back to the evidence (clips, quotes, raw notes). A one-page summary per study, written for people who will never watch a recording, does more for adoption than any tool.

Budgeting Time

A typical qualitative study breaks down roughly like this:

• Planning and research questions: 1 to 3 days.
• Recruiting and scheduling: 3 to 10 days, longer for specialist B2B roles.
• Running sessions: 2 to 5 days for 5 to 10 sessions.
• Analysis and synthesis: roughly as long as the sessions took, often longer.
• Reporting and readout: 1 to 2 days.

Recruiting can overlap with planning, so a focused usability round often fits in one to two weeks. Diary studies, contextual inquiry and A/B tests take longer because the calendar, not the team, sets the pace. With no time at all, a heuristic evaluation plus an analytics review takes a few days; treat the output as hypotheses.

When To Hire A UX Research Agency

Many teams can run interviews and usability tests themselves with a bit of practice. Outside help is worth considering when:

• You have no researcher in-house and a decision is too expensive to get wrong, such as a new product line or a major redesign.
• You need a hard-to-reach audience, like clinicians, procurement leads or users in several countries and languages.
• The method needs specialist skills or equipment, such as eye tracking, large quantitative benchmarks, or accessibility studies with assistive technology users.
• You need independence. An outside team can test a product without internal politics shaping the questions or the readout.

Ask partners for a sample research plan, an anonymized report and their recruiting approach. Our list of top UX research and usability testing agencies is a good starting shortlist. If you need an expert review rather than user studies, see the top UX audit companies. If research is part of a larger build, the top product design agencies typically include discovery research in their engagements.

Frequently Asked Questions

What Are The Most Common UX Research Methods?

User interviews, usability testing, surveys and analytics review are the most common, because they are flexible and relatively cheap. Card sorting and tree testing are standard for navigation work.

What Is The Difference Between UX Research Methods And User Research Methods?

In practice the terms are used interchangeably. “User research methods” is sometimes used more broadly to include market research techniques, while “UX research techniques” usually focuses on how people interact with a product. The methods themselves overlap almost entirely.

What Is Qualitative Vs Quantitative UX Research?

Qualitative UX research observes a small number of people directly to understand why and how things happen. Quantitative UX research measures many people through instruments like surveys or analytics to answer how many and how much. Strong research programs use both, with qualitative work explaining the numbers.

What Is Generative Vs Evaluative Research?

Generative research explores problems and opportunities before a solution exists, using methods like interviews, contextual inquiry and diary studies. Evaluative research tests an existing design or product, using methods like usability testing, heuristic evaluation and A/B testing. The split tells you which group of methods to look at first.

How Many User Interviews Do I Need?

A common rule of thumb is 5 to 12 per distinct user segment, stopping when new interviews stop producing new themes. If you have three very different segments, plan interviews for each rather than one combined pool. Depth and good recruiting matter more than raw count.

Are Focus Groups Good For UX Research?

Focus groups are useful for exploring attitudes, language and reactions to messaging, but they are a poor fit for usability questions. Group dynamics bias answers, and participants are discussing a product rather than using it. For most UX questions, individual interviews or usability tests give more reliable results.

The Short Version

UX research methods fall along two useful axes, attitudinal vs behavioral and qualitative vs quantitative, and serve one of two purposes: generating ideas or evaluating designs. Discover problems with interviews, contextual inquiry and diary studies; shape structure with card sorting, tree testing, first-click and concept tests; evaluate designs with usability testing and heuristic evaluation; and measure live products with surveys, analytics and A/B tests.

Pick the cheapest method that answers your actual question, run small studies often, pair what people say with what they do, and keep findings in a repository so they compound. Bring in a specialist agency when the audience, method or stakes are beyond what your team can cover.