How Do I Do It?

First, you must prepare by gathering your team and selecting the participants.

A typical day at the users' office:

(Holtzblatt and Jones, 1993)

9:00-9:30

Everybody meets and greets. Interviewers may give brief presntaition, whoever needs to attend, does.

9:30-12:00

Contextual interviews

12:00-1:00

Lunch with everybody

1:00-3:00

More interviews with different users

3:00-3:30

Interviewers do initial consolidation/review of findings

3:30-4:00

Meet with hosts to answer any questions and go over results

Send thank-you notes after you get back to the office.

Dos and don't of interviewing

Deal with the concrete actually happening only. Be wary of, and avoid if at all passxble the abstract, the generalized and the rationalizied. Don't let them describe, make them show you. Have them talk about what they are doing as they do it. Ask questions to confirm your understanding or to clarify your confusion. Standard interviewing is not helpful 'cause you don't get the information you need. The environment is a powerful stucturer and reminder of work. Acknowledge user as expert in the work. Ask open ended questions, and seek to generate a flowing dialogue. Explicity acknowledge your focus by stating (to your self) the design problem and what piece of work you want to look at. Expand the focus by asking about what you do not know. From my own limited experience, being a good observer is not something that can be learned from a book. You have to go out and do it, reflect on your experience, and if possible, get feedback from more experienced observers. Have a group visit the site and each interview different people. Have a team meeting before hand to make sure you're all on the same page.

What to look for, what to record

Each interview will begin with a brief introduction to yourself, your project, and the method you are going to be using in the interview. After that comes the heart of the issue "ongoing work inquiry". This is followed by a mini wrap-up. In the interview, you want to record the following:

In the introduction, give a bit more detailed description of how you want the interview to unfold, assure the user of confidentiality and get informed consent, especially if you are recording. Video recording is generally more trouble than it is worth. Audio is probably a good backup unless and until you are confident in your note-taking . A camera for taking pictures is not a bad idea, if the hosts don't mind. Let the user know that they are in control and the expert you are there to learn from. Don't let them get away with abstractions and generalizations, make them give you concrete, detailed actual examples. Get into a rhythm of them working/talking and you quietly observing and then the both of you conversing about the work. Don't feel you have to go the distance. If you get overloaded, suggest a break or end the interview, just as the user should. You also need to clue into subtle communication of users' needs too.In the wrap up you give the interviewee a summary of what you learned. This is your last best chance to validate your understanding. Answer any questions they may have, ask their permission to get in touch with them for post-hoc clarifications and offer profusest thanks.

Donald Cox, coxd@cpsc.ucalgary.ca
Last Modified: 28 February 1997