
AI & UX Research, BEST PRACTICES, UX METHODS, UX QUALITY
In-Depth Interview vs. Usability Test: Which Method When?
13
MIN
Nov 19, 2026
In a formative usability study in the US, healthcare professionals work with the prototype of a blood-based wound care product, and the test protocol assumes that they draw the blood for the application themselves. At exactly this point, many participants say something no test script had anticipated: someone else does that here. Drawing blood requires special training, and hospitals have dedicated staff for it or even external companies.
When you compare an in-depth interview vs. a usability test, you find that the two methods answer different questions. An in-depth interview shows how people work and under which conditions. A usability test shows whether they can handle a specific product. Which method fits depends on the question you want to answer, and that question is missing in many UX research projects.
At this point, the study mainly exposed a wrong assumption in the study design. We will tell you how things continued further below. First, we look at what the two methods deliver and why so many in-depth interviews run without a real goal. Finally, we show how you can combine both in a sensible way. The examples come from our work at uintent, including human factors studies in medical technology.
Key Takeaways
In-depth interviews capture how people work and under which conditions. Usability tests show how they operate a specific product.
Interviews deliver self-reports, tests deliver observed behavior. Both types of data have their own blind spots.
Many so-called in-depth interviews have no clear research goal and too many questions. That leaves no time for real depth.
If the need is still unclear, an in-depth interview usually fits. If there is a prototype and a known context of use, a usability test fits.
A usability test can uncover wrong assumptions about the work context. Then it pays off to step back into interviews.
For qualitative usability tests, five people per user group are often enough. Interview studies usually need more conversations to reach saturation.
What is an in-depth interview and what is a usability test?
An in-depth interview (IDI) is a one-on-one conversation that usually lasts 60 to 90 minutes. Its purpose is to get to know the participants and their environment better, meaning their everyday work and the conditions under which they make decisions. It is based on a discussion guide with open questions. How many questions the guide contains varies from study to study. What matters more is leaving enough room to ask follow-up questions and to pursue an unexpected thought.
A usability test works differently. Participants receive tasks and complete them on a prototype or product while the research team observes. Often they think aloud, so their reasoning becomes visible alongside their behavior. The result is a set of concrete findings: Where do people get stuck, which steps do they misunderstand, where do they take detours?
Christian Rohrer of the Nielsen Norman Group places UX methods on an axis between what people say and what they do [Rohrer, 2014]. The in-depth interview sits clearly on the "say" side, the usability test on the "do" side. Self-reporting has a built-in limit. It only goes as far as people are aware of their own behavior and are willing to share it [Rohrer, 2014].
Which questions does an in-depth interview answer, and which does a usability test answer?
The in-depth interview answers questions about the why and about context. The usability test answers questions about the how, specifically for a concrete product.
| In-depth interview (IDI) | Usability test (UT) |
Guiding question | Who are the users, how do they work, what do they need? | Can users handle this product? |
Typical insights | Needs, goals, motivations, pain points, habits, existing workflows, evaluations | Comprehensibility, ease of use, navigation problems, errors, detours |
Type of data | Self-report | Observed behavior, supplemented by thinking aloud |
Prerequisite | A clear research question | A prototype or product and realistic tasks |
Typical timing | Early, while the need is still open | As soon as there is something to try out |
Limitation | Results can stay abstract | Reveals little about needs and work context |
The biggest strength of the interview is exactly what the test lacks. It makes context and needs tangible. In a usability test, we see that someone fails at a step. Whether that step even occurs in their everyday work is something we rarely see there.
The flip side: interviews can stay vague. People often describe their routines the way they should be, and questions about a hypothetical product lead to hypothetical answers.
The usability test has an advantage that no interview can replace. It uncovers problems the team was not aware of beforehand, because the method relies on observing more than on asking. It is still not free of influence. Task wording and moderation shape the results, and so does the maturity of the prototype. If a task already uses the word from the menu, you have done the searching for the participants.
Why so many in-depth interviews run without a real goal
Many studies are called in-depth interviews and turn out to be a questionnaire read out loud. We see this again and again in projects. The name promises depth, but the guide leaves no room for it, because participants get so many questions that they cannot really think about any of them.
The pattern behind it is usually the same. There is no clearly defined research question. So the team collects everything stakeholders have always wanted to know, and the guide grows with every round of alignment. A fictional but not unrealistic example: with 40 questions in 75 minutes, you have just under two minutes per question, before subtracting the welcome and the consent form. There is simply no time left for follow-up questions or for pursuing a side remark that suddenly turns interesting.
That is why an in-depth interview needs a goal before it needs a guide. This can be one or two research questions from which the guide questions are derived. A research question might read: How does wound care work on the ward today, and who takes on which step? Every question in the guide has to serve a research question like this. Anything that does not gets cut, even if it sounds interesting.
This check is uncomfortable because it takes questions away from stakeholders. It protects the quality of the results, though. You can read more about what good research quality depends on in our article UX Research Quality: Why Good Intentions Aren't Enough.
When does an in-depth interview make sense, and when does a usability test?
An in-depth interview makes sense as long as the need or the work context is still unclear. A usability test makes sense as soon as a prototype exists and it is clear who uses it in which situation. Four questions help with the decision:
1. Is it clear which problem the product should solve and for whom? If not, start with in-depth interviews.
2. Do you know the work context from observation or conversations, or only from assumptions? If there are only assumptions behind it, you first need interviews or an on-site context analysis, for example through a Contextual Inquiry.
3. Is there a prototype or product on which realistic tasks can be completed? Then a usability test is possible.
4. Do you want to know whether something can be understood and operated, or whether it is needed at all? The usability test is for the first, the interview for the second.
Teams skip the second question most often. Assumptions about everyday work feel like knowledge, especially when they come from subject matter experts or have been floating around the project for a long time. Our case study shows what happens when they turn out to be wrong.
What happens when a usability test exposes wrong assumptions?
Back to the study from the beginning. The product under test was a blood-based wound care product, and the participants were healthcare professionals in the US. The study design took it for granted that they draw the necessary blood themselves. During the test, it turned out that this was not true for many of them. Drawing blood requires special training, and in many hospitals dedicated staff or other companies take care of it.
So more than a single task was off. The question was how well we actually knew the working reality of this target group. The consequence: before testing the usability of the product any further, we would first get to know the people and their work processes.
Madeleine then conducted in-depth interviews with healthcare professionals together with Wolfgang Waxenberger. The focus was on context analysis: What does the workflow look like in practice, and where are the barriers for a new product? Many assumptions the project had brought into the usability test could be checked there. And a lot turned out differently than expected.
For us, this case holds the core lesson of the whole topic. The usability test did nothing wrong here. It did what it does well, which is making problems visible that nobody had asked about. The problem simply sat one level deeper than a test can measure. And because the study was formative, the course could still be corrected before expensive design decisions were built on wrong assumptions.
How many participants do interviews and tests need?
For qualitative usability tests, a rule of thumb has held for years: five participants find about 85% of usability problems [Nielsen, 2000]. Nielsen himself recommends spreading a budget for 15 people across three test rounds of five each and revising the design in between [Nielsen, 2000].
The rule comes with a condition that often gets overlooked. It only applies to users who are similar and use the product in comparable ways. With clearly distinct groups, each group needs its own participants [Nielsen, 2000]. In medical technology, that is the norm, because nurses and patients, for example, do very different things with the same product.
Interview studies follow a different logic. Here it is about saturation, the point at which new conversations hardly bring up new themes. A systematic review finds that saturation is typically reached after 9 to 17 interviews [Hennink & Kaiser, 2022]. With 60 to 90 minutes per conversation plus analysis, an interview study is often more effort than its reputation as "just talking" suggests.
If you need reliable numbers instead of problems and themes, quantitative methods are the better choice. We explain how to read and interpret them in Quantitative UX Methods: How to Read and Interpret Numbers Correctly.
How can you combine in-depth interviews and usability tests?
The two methods complement each other so well that combining them almost always delivers more than using either one alone. The order depends on where the project stands.
The classic route leads from interview to test. First you clarify the need and the work context, then you build a prototype on that basis and check it in a usability test. That way, the tasks in the test are realistic because they come from real workflows.
Our case study shows the reverse route. A usability test brings gaps in the understanding of the target group to light, and in-depth interviews close them. For the project plan, that feels like a detour. In fact, this is exactly where research does its job.
A third variant combines both in one session. An interview part before the tasks captures the context, and one afterward collects evaluations and reasoning. That saves recruiting effort. The risk is obvious: if you squeeze 20 minutes of interview and 40 minutes of testing into one hour, you get a usability test with a warm-up conversation, but no in-depth interview. That is perfectly fine, as long as everyone involved knows which questions this session can answer and which it cannot.
Conclusion: First the question, then the method
In-depth interviews and usability tests do not compete with each other. The interview reveals how people work and what they need. The test shows whether a product works in that everyday reality. Both become weak when nobody has defined in advance what the study is supposed to find out.
Our recommendation for your next study: write down the research question in one sentence before you talk about methods. If the sentence starts with "How do … work" or "What do … need," it is a case for an in-depth interview. If it starts with "Can … handle," it is a case for a usability test. And if you notice that you are only guessing the answer to one of these questions, start there.
In the workshop The Fundamentals of Successful Continuous Discovery with a Dash of AI, we show how teams continuously test needs and solutions against each other instead of relying on assumptions gathered once.
Which assumption about your users most recently turned out to be wrong? We are curious.
FAQ
What is the difference between an in-depth interview and a usability test?
An in-depth interview uses open questions to capture how people work and what they need. A usability test observes how people complete concrete tasks on a prototype or product. The interview delivers self-reports, the test delivers observed behavior.
How long does an in-depth interview take?
An in-depth interview typically lasts 60 to 90 minutes. How many questions fit into it depends on the goal. As a guideline: fewer questions and more time for follow-up questions.
Can you combine an in-depth interview and a usability test in one session?
Yes, an interview part before or after the test tasks is common and saves recruiting effort. It rarely becomes a full in-depth interview, though, because there is not enough time for it. Plan the session so that it is clear which questions it is meant to answer.
How many participants does a usability test need?
For qualitative usability tests, five people per comparable user group is the rule of thumb, and they find about 85% of the problems [Nielsen, 2000]. Several test rounds with revisions in between deliver more than one large test. With distinct user groups, each group needs its own participants.
Does a usability test replace a needs analysis?
No. A usability test shows whether a product can be operated, but says little about whether it fits into everyday work and is needed at all. If a test shows that workflows differ from what was assumed, that is a signal for in-depth interviews.




















