This is an early version of how to conduct a rich listening session, where you find out the thinking going on inside someone’s mind as she pursues a larger intent. For example, “I’m presenting to a class this afternoon–how shall I approach this class, and how should it differ from the last class I presented to?” At this early stage, I still referred to the mental model diagram as containing “task-analysis.” I have since dropped that vocabulary, because it has other meaning.
(This post was originally published at Adaptive Path. Since Adaptive Path became a part of another corporation in late 2014, the post is replicated below for posterity.)
Six Steps to Better Interviews and Simplified Task Analysis
I spend a lot of time helping clients conduct task analysis to form mental-model diagrams. When teams first start analyzing the interview transcripts they’ve collected, they often run into a confidence issue. “How will we know if we get the task groups right?”
This question usually arises because the team doesn’t have the kind of details it needs to identify clear tasks. The problem isn’t in sorting; it’s in the data-gathering stage. If interviews don’t provide details, task sorting becomes much more complex. Fortunately, there are six simple things you can do to improve the quality of your interviews, and clarify task analysis.
1. FOCUS ON DEEPER GOALS
The whole point of drawing up a mental model diagram is to understand what is going on in people’s minds as they accomplish their goals. We want to know how people organize their effort and whether they do things similarly to another person.
The team might see a quote in the transcripts like “I keep in close contact with people each week.” Perhaps they decide that the task behind this quote is “contact people weekly” or “stay in contact with people.” Later, when you’re looking for patterns in the analysis, this task is too vague to group with any other specific task; it becomes an outlier.
This is because a task like “stay in contact with people” is too far removed from an explicit goal; it only hits upon the external action. The deeper goal might be “find out the progress of assignments” or “gather weekly sales totals” or “develop friendly relationships.” It is these deeper goals we need for our mental model diagram.
2. SLOW DOWN
In the commotion of an interview, it’s easy to forget that you need your subject to elaborate. This is especially true when a researcher is just learning to conduct interviews. It’s natural to hear a task, even a vague one, and decide to move on to other topics.
To eliminate vague tasks, the interviewer must ask follow up questions. In our earlier example, when the subject says, “I keep in close contact with people each week,” slow down and consider what she has said. Is it self-explanatory? How does it relate to the context of the conversation? Is it a new topic?
3. ASK WHY
Many of us know a toddler who has discovered the single most powerful learning tool: the word “why.”
“What is that?” asks the toddler, standing in the kitchen and pointing. You answer, “It’s a refrigerator.” “Why?” “Because that’s what we call it.” “But why?” “Because it refrigerates food.” “Why?” “So that the food won’t spoil. Some kinds of food must be kept cold if we want to use them in a few days.” “Why?” “If the food spoils, then it contains germs and fungus and bacteria, and that can make us really sick if we eat the spoiled food.” “Why?”
You get the idea. The repetition of the word “why” drives most of us crazy, so I’m not recommending such a simple approach for your interviews. But keep the concept in mind.
Keep digging into the background of a topic until the interview participant has no more to say about it, or takes you on another tangent. (Or sends you to your room for a “time out.”)
4. FORGET THE CLOCK
Latch on to the conversation you’re having and forget about clocks and timers. The most important thing is to collect in-depth data. Finishing the interview on time is secondary.
Of course, interviewers can’t help but be aware of the clock. We’ve promised to conduct the interview within a certain period of time; we’re all afraid that we’ll take too long and disappoint the interview participant.
Remember, people love to talk about what they do. Because of this oh-so-human tendency, you already have an advantage. You’ll usually be faced with a flood of ideas that you must sort through and fully explore.
Interviewing shouldn’t feel like pulling teeth. If it does, simply wrap up by saying these are all the questions that apply to this topic. Thank the participant, and go recruit someone else.
5. DIG FOR DETAILS
You’ll find that most participants will talk in vague terms at first because they don’t want to bore you with the details. You must get past this hesitation on their part. If you’re interviewing a rocket scientist, that person will be aware of your limitations on the subject.
Get the conversation rolling with a question about something harmless, like the weather, and then dive into the topic. Not many people have a significant other who is willing to listen to sixty minutes of blathering about how to solve certain problems, so participants are often pleased to find that you’re eager to delve into details.
Acknowledge your limitations, and encourage the interviewee to explain the mysterious points. You’re not looking for an explanation of the coefficient of gravity, but you are after the goals and steps the scientist employs. Goals and steps are easy to explain, even to someone with a limited grasp of physics.
6. SCRAP THE QUESTIONNAIRE
My last interviewing tip is to dispense with a list of interview questions. Reading questions in the order you wrote them is almost never in step with the current topic of conversation, so it’s tough to keep natural interaction going.
Instead, make a list of nouns or phrases that highlight the topics you want to discuss. These keywords will keep you on task, but they’re easier to read and to incorporate into conversational flow. You can check them off as you mention them without regard to numerical order.
CONFIDENT TASK ANALYSIS
To have confidence when building your mental model diagram, you need in-depth interview data. If your transcripts go into detail about each topic, you can create tasks with strong, explicit verbs. These tasks will be much easier to analyze for patterns, and your mental model will build itself.