How to Run User Interviews

by | May 10, 2019 | General, Research

 | 5 min read

Photo of two persons talking

User interviews are a very popular method of user research. They can easily be run in combination with other tests and are applicable in a number of situations. However, you need to keep in mind that there are things an interview can do and just as many things that an interview can’t do. So only use interviews in the appropriate situations.

You can use interviews to enrichen existing data that you gathered with user tests. You use them after usability tests to collect answers to questions that arose during the tests. Interviews are also very well fit to get a general overview of a topic. In the following case, you’re running exploratory interviews. You’re asking a series of very general and open questions, to get a feeling of an area you might not be well informed about.

Different kinds of interviews you can conduct:

Structured interviews adhere strictly to a given set of questions that you decided to use before you started the interview. Unstructured interviews give interviewees the opportunity to freely tell you how they feel about a certain topic. Unstructured interviews are right at the intersection of those two interview types. You’re thinking of a set of questions you need answers to but let users answer them freely. It is also possible for the interviewer to change the direction of the interview while they are doing the interview. If the interviewee says something particularly interesting, the interviewer may follow this thought by asking appropriate follow-up questions.

Semi-structured interviews are usually the preferred mode of interviewing participants as they allow interviewers to generate new ideas based on what the interviewee said. They give participants room to answer freely while at the same time avoiding the problems of completely unstructured interviews. They are hard to evaluate while it’s easy to lose track of what you wanted to find out in the first place.

Semi-structured interviews are a great way to supplement existing data, inform personas or user journeys and gain a general understanding of research areas. In general, you should use an interview if your research question or problem doesn’t need a direct answer but rather further exploration. Or if you want to find out how others relate to a certain topic. However, as has already been stated above, there’re also times at which an interview isn’t the research method you should prefer.

When and how not to use interviews

In general, you shouldn’t use interviews to gather data about users behavior or the way they actually use products. A certain problem arises if you ask your users about past interactions with a product or how they used it. If they can’t really remember how they did it, they won’t be aware of that fact. It’s because the human brain tends to fill in forgotten details with made-up details that make sense to us. Because of questioning users’ past behavior will probably result in inaccurate information and may even result in an answer that has been unconsciously rationalized by the person you’re questioning.

You shouldn’t ask users how they will use a certain product in the future or how they are planning to use certain products. They probably won’t be able to imagine how they would use a product–especially based on a prototype or description alone. Also, users are not designers. Capturing their opinions and experiences is certainly helpful. Having them create their ideal product or suggest ways to improve an existing product won’t get you very far. Simply asking them for their opinion won’t help you make design choices such as what color of button works best, where to put elements etc. If you want help designing or improving an interface, have users interact with a prototype.

During interviews, users can only give you accurate information concerning the present. Therefore interviews are useful for assessing users’ attitudes towards a product right after they used it. If you absolutely have to ask them about past interactions, try using the “critical incident” method: ask users about a time when something worked especially well or didn’t work at all and frustrated them. Extreme cases are remembered better.

Interviews are easy to conduct and can add valuable data to your existing research results. Doing them can give you valuable insights and pointer for new directions of research. You shouldn’t, however, rely on interviews as your only way of user testing.

How to conduct interviews

If you plan to run a semistructured interview, think about what your research question is before you start conducting the interview. Think of a number of questions you can ask your participants, that may help you answer your initial question. Use this set of questions as a general guideline during your interviews but try to always stay open-minded.

Try to look at the situation more like it’s a conversation between you and the participant and less like it’s an interview, you will uncover new insights and your opponent will feel more human. If you can, use your questions to get the conversation started and to give the interviewee a general guideline of what you want to talk about. To uncover the bigger picture try asking as many follow-up questions as possible. Asking users what they like or dislike, always ask why they like or dislike whatever they mentioned. This will either lead to them going into greater detail. This will give you valuable insight and can also indicate how important this issue is to them.

In case you’re asking for opinions, don’t force opinions that don’t exist. When asked for their opinion on a certain topic, humans will come up with an answer, even if they don’t really care about the topic at hand. Always give them the possibility to say that they don’t really have an opinion on the given topic. Keeping your prompts as open as possible will give participants the ability to answer freely. It will help them realize that their opinion truly matters to you and they aren’t reduced to numbers. If you have trouble evaluating these very open kinds of questions try doing a thematic analysis.

How to start an interview

Always try to set a casual atmosphere and make your interview partners feel at ease. Start your interview by greeting the interviewee with their name, explain why you’re interviewing them, and provide them with some context. If there’re other people/researchers present with you, introduce them as well. Tell your participants that you’re grateful for their participation and that their opinions are valuable to you and your team. Give them the opportunity to ask any questions they might have and start the interview with some easy “warm-up” questions.

Sometimes it’s good to ask questions even if you think you know the answer to those questions. During the interview, ask open-ended questions, avoid leading questions and don’t interrupt or correct your participants. If you feel like your participants start talking about irrelevant things, don’t block them immediately. Maybe your dialogue partner specifically wants to tell you something. Also interrupting someone is rude and you might gain some important information from letting them talk freely. Only try to gently get them back on track if they stray too far from the important topic.

If possible, try to conduct your interviews in person: remote interviews are generally finished more quickly: if you use questionnaires, answers to open-ended questions tend to be very short or non-existent. Whether you do interviews via phone, people will give you shorter answers than they would in person, agree more easily and lose patience with the interview situation more quickly. If you don’t rush your participants ask questions that show you are interested in their problems and listen closely and actively, you’re sure to gain some valuable insights to add to the data gathered during user tests.

Not sure how to combine the data you gathered in user tests with qualitative interviews? Try to connect your insights! Start your 30 days free trial now! is a fast-growing tool for quali­tative data analysis and UX research repository.


Mara Weingardt

Mara is interested in all topics around user research, user testing, as well as usability and UX.
Share This