Asking people to recall what they did is better than asking them to speculate on what they will do. Yet a recall means you’re talking to the remembering self, as opposed to the experiencing self. The remembering self doesn’t know duration, and is triggered most by peak intensity whether good or bad, and final events. It would be better to observe how someone does that task, and ask for commentary about what they are doing and note emotional reactions to the work. It’s a good idea to record the interview as a way to debrief what happened and recall the experience. As the interviewer, you’ll also be much more aware of peaks and the end of the interview.
I’ve been guilty of thinking that talking to a handful of users is enough to know if our idea is valid or not. And I know of others in this field believe that, too. But this is the law of small numbers, where you’re most likely to hit an anomaly in these small numbers and find someone that really loves, or really hates, your application. Or their job. Or both. This will introduce a confirmation bias too early that your team should pivot, or persevere. On-site interviews help us with empathy and this needs to be paired with quantitative analysis. There are many great tools and practices to supplement the qualitative analysis in fast and inexpensive ways.
Asking a question like, “How do you like your work?” is too tough for most people to answer without some in-depth analysis. Coming up with a quick answer involves substituting a question like this with an easier one like, “Did I like the last thing I did at work?”. The question can also be influenced by factors like the previous question, the mood of the user, time of day, and even the physical environment where the interview is taking place. Be sure to ask specific questions and be careful of the relationship between the questions.
These are the big items that leapt out for me with respect to user research as I read the book. How about for you?