In progress Course · 13 sections · about 7 min · Published

UX Research 101: Methods, Interviews and Usability Testing

A practical introduction to UX research: when to research, how to choose a method, run interviews and usability tests, and turn notes into insights.

The notation in one minute

The diagrams in these courses use The Shape of Clarity, a small visual notation. Five shapes are enough.

Small labels or icons say what an interface holds: a question (?), an insight (!), ideas (a light bulb), a decision (a check mark). Speech bubbles and the chart symbol, for evidence, are added on top of the notation when needed; they are not part of it.

Read the notation
Contents
  1. About this course
  2. 1. The scene: people, interface, environment
  3. 2. Research in the loop: Understand and Test
  4. 3. What you can observe: say vs. do, why vs. how many
  5. 4. When to research (and when to push back)
  6. 5. Where research sits in the timeline
  7. 6. Two directions: top-down and bottom-up
  8. 7. Choosing a method
  9. 8. Triangulation
  10. 9. The pragmatic process: cast wide, then zero in
  11. 10. Three methods in practice
  12. 11. Synthesis: from observations to insights
  13. 12. Research as an operation: the loop inside a team

About this course

Research is how the design loop touches reality. Design Thinking 101 teaches that loop as five steps: Understand, Ideate, Decide, Prototype, Test. Research concentrates on two of them, Understand and Test, and the two are closely linked: testing a prototype is understanding again, and it closes the loop. In an evolve case, understanding often starts by testing the real product.

Every research method is a different way of looking at the same scene: people, an interface, an environment.

  • Who it is for: anyone who needs to learn what people actually do and need before (and after) designing something for them.
  • How it works: self-paced. Each section has a short explanation, a diagram, an exercise (Do), and readings (Read), data-oriented sources first.
  • Your case: use the same digital product as in Design Thinking 101, either a build case (it doesn’t exist yet) or an evolve case (it exists and must change). Your findings will feed the design work.
  • Out of scope: Figma, accessibility and performance are covered in UX Design 101.

The vocabulary used in this course:

  • Observation: what you saw or heard.
  • Finding: a pattern across observations.
  • Insight: a finding plus its implication.
  • Evidence: anything measurable or verifiable in reality. Tests produce it, and careful observation captures it while you are still understanding the problem.
Inside an environment, a user triangle acts on an interface and gets feedback; its insight reaches a dark triangle, whose hypothesis returns to the interface.
Empathy is the discipline of standing where the user stands.

1. The scene: people, interface, environment

Every UX question is about three things: someone (a user) acting on something (an interface) somewhere (an environment). If you can’t name all three, you don’t have a research question yet.

Empathy starts here. It is the discipline of seeing the scene from the user’s position, not from yours.

Do. Pick an app you used today and draw the scene. Who else is in the environment?

Read.

The five-step loop with Understand and Test highlighted; a grey triangle steps outside to observe, returns with evidence.
Research is what happens where the loop touches reality.

2. Research in the loop: Understand and Test

There is one loop: the algorithm of Design Thinking 101. The designer steps outside the scene to understand it. The loop continues through Ideate, Decide and Prototype. Then Test brings the designer back to understanding the scene, which has now changed. Research is what happens at the two moments where the loop touches reality.

Do. Place your last project on the loop. Where did it stop?

Read.

A two-by-two map: say vs. do on one axis, why vs. how many on the other; six methods placed as labelled dots.
Every research method sits somewhere on these two axes.

3. What you can observe: say vs. do, why vs. how many

Two axes classify every method:

  • Attitudinal vs. behavioural: what people say vs. what they do.
  • Qualitative vs. quantitative: why and how vs. how many and how much.

When saying and doing disagree, trust doing.

Do. Place six methods on the map: interview, survey, analytics, usability test, card sort, A/B test. Which quadrant is empty in your practice?

Read.

A question passes four yes-or-no checkpoints; failing any leads to "push back", passing all to a research plan.
Research is worth its cost only if all four checkpoints pass.

4. When to research (and when to push back)

Research costs time. Only start it if all four are true:

  1. you are uncertain;
  2. you have a specific question;
  3. you haven’t already answered it;
  4. you can’t find the answer elsewhere.

Otherwise, push back or go straight to design.

In an evolve case, the cheapest first diagnosis is an expert review of the existing product against Nielsen’s heuristics (see UX Design 101, “Good vs. bad UX”). It tells you where to point the research.

Do. Rewrite three vague requests (such as “users hate the dashboard”) as specific research questions, and run each one through the four checks. Evolve case: first run a 20-minute heuristic pass on the existing product and list its three worst problems.

Read.

Two diverge-converge shapes side by side: generative, formative, summative in order, evaluative dropped in anywhere.
The Double Diamond shows where each research type sits in time.

5. Where research sits in the timeline

  • Early research generates: it explores the problem.
  • Mid-project research forms: it shapes the solution.
  • Late research sums up: it measures the result.
  • Evaluative research can happen at any time, on an existing product or on a prototype.

The Double Diamond draws the same idea twice: diverge, then converge. It is a composite of two moves: divergence (arrows leaving one point in different directions) and convergence (arrows meeting at one point).

Do. For your case, list which type of research you need this week and why.

Read.

An environment with one arrow descending into it (a hypothesis tested) and one rising out of it (a finding emerging).
Good research alternates between testing and discovering.

6. Two directions: top-down and bottom-up

  • Top-down: start from a hypothesis and check it against reality (deductive).
  • Bottom-up: start from observations and let findings emerge (inductive).

Good research alternates between the two.

Do. Take one finding from your interviews. Was it a confirmed hypothesis or an emerging pattern?

Read.

Two columns joined by arrows: typical requests on the left, research methods on the right.
Each kind of request points to one or two matching methods.

7. Choosing a method

Start from the request, not from the method you like. Each type of question points to one method or a few.

Do. Take five requests from people you work with. For each one, pick a method and defend the choice in one sentence.

Read.

Three views of the same scene — data, interviews, observation — each sending an arrow that meets at one insight.
Three kinds of evidence converging give a position you can trust.

8. Triangulation

One method gives you one angle. Combining quantitative data, qualitative data and an expert review gives you a position you can trust.

Do. For one finding, name what the other two views would need to show to confirm it.

Read.

Many small environments at the top, arrows converging downward to a single user acting on a single interface.
Start broad across many environments, then narrow to one.

9. The pragmatic process: cast wide, then zero in

Start broad so you don’t miss the real problem. Then narrow down to the specific risk or opportunity worth a decision.

Do. Write a one-page research plan: question, method, participants, timeline, and the decision it should inform.

Read.

A grey and an orange triangle face to face; the orange triangle's speech bubble holds a dashed, remembered interface.
An interview captures what's remembered, not what's actually used.

10. Three methods in practice

Interviews

Ask about past behaviour, not future intentions. Listen more than you talk. An interview is where empathy becomes practice: you borrow the user’s position for twenty minutes.

Do. Run a 20-minute interview about your case, in pairs if you can. Note five observations of behaviour, not opinions. Then fill in an empathy map (says, thinks, does, feels) and mark which quadrants you observed and which you inferred.

Read.

Usability testing

Give people a real task, watch, and don’t help.

How many people? The reference for qualitative testing is five per round: according to Nielsen, five users find most of the problems in one design iteration. Quantitative studies need at least 20 people. Regulated fields set formal minimums: the US FDA expects 15 participants per user group for medical-device validation.

One user is better than none: the first person already reveals about a third of the problems. The real trap sits at the two extremes, either not testing at all or over-investing in one big test instead of several small rounds.

A user acts on an interface inside an environment; a grey triangle observes from outside, an arrow leads to a chart symbol.
Watching a real task in use is what a usability test records.

Do. Test someone else’s prototype (or your own) with five people, or fewer if that is all you can get. Report the top three findings.

Read.

Card sorting

People group your content, so your structure matches how they think about it.

Small teal boxes scattered at first, then gathered into three clusters by an orange triangle
People's own groupings reveal the structure your content should have.

Do. Run an open card sort of 20 items with five people and compare their groupings. For a real study, NN/g recommends at least 15 participants per user group.

Read.

Many user-interface pairs cluster into groups, converge into one insight, then lead onward to a new question.
Observations become findings, findings become one insight, insight becomes a question.

11. Synthesis: from observations to insights

Raw notes are observations, not findings. Cluster your observations into findings (patterns), then add the implication: an insight is a finding plus its implication. An insight without an implication is trivia. Each insight becomes a “How might we…?” question for the Ideate step (Design Thinking 101, section 6).

Do. Make an affinity map of your interview notes and write three insights in the form “observations → finding → implication”.

Read.

A pipeline of labelled boxes — invite, schedule, session, repository, insights, report, plan — looping back to the start.
Research compounds in value only when it loops back into the team.

12. Research as an operation: the loop inside a team

In an organization, research is a repeatable pipeline: recruit, schedule, run sessions, store, synthesize, share, plan again. Its value compounds only if findings are stored and reused.

Do. Map how research moves (or doesn’t) in a team you know. Where does it leak?

Read.

Where this shows up in the work

Next in the course path

UX Design 101: From User Goals to Wireframes

In progress Course · 11 sections · about 8 min · Published
Start the course

Bring clarity and structure to your team

The same method, applied to your organization: a fixed-scope diagnostic with one price, one timeline and a deliverable described up front.

See the diagnostic