When conducting an interview, there are a few things to keep in mind:
Don’t ask leading questions. Leading questions range from yes-no questions to questions that merely hint at a certain answer. They restrict the range of answers a person can give and therefore often result in you missing the big picture. “Please describe your feelings during the composition process” is a better question than “how frustrating is the composition process for you?” if a person hasn’t expressed any feelings about the composition process yet. If a leading question comes to mind, reformulate it to be non-leading.
Don’t ask to speculate. People are generally bad at predicting future behavior, so asking for solutions to problems, or asking to speculate what they would do in a situation tends to yield very inaccurate answers. Instead, ask people about past behavior or past problems. If they give feature suggestions anyway, ask why they’d like the feature and ask them to describe past situations that have come up in which it would have been of use.
You’re a listener. Don’t suggest solutions, don’t argue with what the interviewee is saying, and be friendly and accepting. If a person is objectively wrong, don’t correct them — these mistakes tell you about a person’s mental model and are extremely valuable in the design phase.
You’re a friend. You need the interviewee to be comfortable with you. Don’t bombard them with questions, ask like a friend and keep the conversation natural. Also, wear casual clothes.
Ask for concrete descriptions. Recent individual cases tend to be much more detailed than a general rundown, so start by asking about those and only afterwards ask about how those differ from the general case. Encourage storytelling and try to cover as much detail as possible.
Ask for demonstrations. If a person can demonstrate a task, even better. Demonstrations show you all the nuances and details of a task that you need, ones that a person could easily gloss over. Afterward, ask about how the demonstration differs from how it’s generally done and about what tends to vary.
Ask the obvious. You’ll make assumptions throughout the interview and you will feel like skipping some questions because you know the answer. Try to ask anyway. For example, your concept of what something is may slightly differ from the interviewee’s, and by asking for a defintion, you’ll get a clearer picture of their mental model.
Don’t be afraid of silence. If you don’t know what to ask, keep it quiet for a bit. It’ll give the interviewee some time to relax and reflect and you’ll have more time to think of a good follow-up question. If silence persists, the secondary interviewer will jump in.
Focus on the goal, not on your product. You’re trying to find out how the user accomplishes the goal you set out to target — focus on that. Even if your product doesn’t currently handle a part of the interviewee’s process or is currently set up for a different workflow, it’s still important that you understand this person’s entire workflow. You might uncover business opportunities or influence your product priorities going forward.