Customer interviews are not a survey with extra steps
On asking about the past instead of the future, and why the most useful answers are stories.
Most customer interviews I am asked to review are surveys conducted out loud. A list of questions, asked in order, with answers recorded in a spreadsheet. The questions are usually about preferences and intentions: what would you like, how likely would you be, which of these matters most.
People are poor at predicting their own behaviour and generous when asked for opinions. The result is a large volume of confident data that points in no particular direction.
Ask about what happened
The most useful interviews are about the past. When did you first realise you had this problem? What were you using at the time? What made you start looking? Who else was involved? What nearly stopped you?
These questions produce stories, and stories contain the things that actually drive decisions: timing, triggers, politics, anxieties, the alternative that was almost chosen.
Opinions tell you what people think they value. Stories tell you what they paid for.
Listen for the words
Pay particular attention to the language customers use without prompting. The phrase a customer uses to describe their problem is almost always better than the one your team uses, because it is the phrase the next customer will be searching for.
Know when you are done
You are not finished when you have a certain number of interviews. You are finished when new conversations stop surprising you. In most B2B markets that happens somewhere between fifteen and forty conversations per segment, and it is usually obvious when it does.