Short and Sweet Research Phase
I've finished writing my slides for my three hour workshop at Agile08. The theme behind the slides is making mental models fit within an Agile sprint. Some people's sprints are two weeks long. Other people's sprints are 6 weeks long. It varies. If you are living in the land of the shorter sprints, though, I can see how you'd think, "There is NO time for this mental model stuff. It sounds nice, but I've got deadlines to hit."
Luckily, a mental model is simply a diagram--a way of aligning understandings about the customer and the business. The amount of work that goes into the diagram doesn't have to take months. You could sketch the diagram in an afternoon based on what's in your head already. At the beginning of many of the workshops I teach, I ask attendees to sketch a model of movie goers in 10 minutes. It's more than enough time, and usually the models contain many of the same mental spaces as the model I create from research interviews.
The common mistake in sketching a mental model is taking the employee (problem-solver) viewpoint, rather than the customer (getting-something-done) viewpoint. We are all trained to look at things in terms of how we would fix it. The upper part of the mental model should be a plain-and-simple representation of the customer's world, using their own verbs to describe what they are doing and feeling. If you can do this, then the bottom half of the diagram will come easily, using your skills to list possible solutions.
There's an online appendix to the book that you can read right now discussing the time and resources aspects of making a mental model. In Appendix A, I describe several types of shortcuts that I've taken in the past. Besides sketching the model yourself in an afternoon, you could try sketching it as a group exercise with input from many varied people in your organization. Better yet, actually get out there and conduct some interviews. Ask people questions and concentrate on their answers. Jot down notes. If your notes contain the customer's verbs (and if you jot them on sticky notes), then you get to skip the transcript analysis and jump directly to building the model.
If money and the reputation of your organization are riding on your ability to see a new avenue towards your customers, then you'll want to spend a little more time digging into vague customer statements and getting to the root--the reason why. Fitting in-depth interviews and transcript analysis into sprints works if you can run a parallel track to development using those same sprint intervals. (If you're the only developer, and you can admit that you aren't the user, then cease development for a couple of weeks to learn a bit more about your customers.) If you can run a little research in parallel for a while, it helps to break it down into audience segments, and focus on just one segment per sprint. Interview participants one week and analyze transcripts the second week. For the next sprint, try focusing on a different audience segment. Alternately, for the second sprint, build a mental model for the first audience segment and align some strategic ideas beneath the towers so that you can get the development team working in parallel with you on the third sprint.
Mental models are not difficult to draw. The difficult part is getting into the mindset of the customer and stepping out of your problem-solving role for a while. If you can do this, then how much time you spend developing an understanding of your audience segments is up to you, your organization, your process, and your deadlines.
Postscript: Last week I presented a workshop at UX Week in San Francisco, and on the last day a fellow named Jure walked up to me during a break while I was gobbling down a chocolate-chocolate Haagen Dazs bar. He flipped open his Macbook Air, showed me the mental model he made about bloggers, and said, "We're a startup, so I didn't have a lot of money to talk to folks. Instead I just listed a bunch of behaviors and came up with this. Is that the wrong way to do it?" I looked at the model and, sure enough, it seemed to capture the mental spaces you might expect for bloggers. I told him, "Good work!" and suggested that if he wanted to increase his confidence in the model, he could interview people over the coming months, one at a time, and add things to the diagram, or adjust it as needed. I hope he feels reassured a sketch like his is a great way to get started.