Latest Weblog Entry
John Zimmerman on Idea Generation and User-Centered Design
John Zimmerman, one of the three resident Design professors in the HCII, came to our capstone project class a few weeks ago (you can tell I'm behind...) to talk about ways to create user-centered design ideas. He espoused several interesting views that in many ways are fundamentally different from the more data-centric ways we've learned in the MHCI program.
John talked about several strategies you can follow to create innovative and appropriate designs:
- Draw from user intents, not from tasks, needs, or problems. What John means by "intents" seems to be the same thing Cooper means by "goals"; they are the underlying purpose of the user's desires and actions that remain constant even in the face of changing environments and technologies (unlike tasks, which vary). Understanding your user's real intents will free you to look at new ways of fulfilling them without straying from the path of user-centered design.
- Visit analogies and metaphors. In the course of designing, a common means of generating new ideas is to come up with metaphors, or solutions that are somehow similar to the ones you're considering. John suggested that you not just invent such metaphors, but actually get out of your office space and visit them. For example, a project team developing a document organization system for an operating system came up with the metaphor of tombstones in a graveyard. They went to a graveyard and examined the metaphor in its real setting to help flesh out the metaphor.
- Design for extreme users. Instead of imagining normal, middle-of-the-road people such as those commonly captured in personas, invent interfaces for the edge cases. If you're designing an MP3 downloader, think of the user who only downloads one song every five years and the one who downloads 50 a day. Alternately, there's the designing for extreme users technique (that I've mentioned before), which calls for looking for unusual types of users in addition to edge behaviors. John says he uses this technique to generate new ideas when he's stuck; obviously the interface that's appropriate only for an extreme edge case user would not be appropriate for the final design.
- Channel your personas. Actually those are my words, not John's, and I use them slightly tongue-in-cheek. John believes Cooper's concept of personas is almost right, but not quite. It's not enough to define a specific user and design an interface for them. John belives it isn't possible for a designer to design for anyone other than himself. However, it is possible to become the user; you can think like they do then design for yourself, rather than thinking like yourself and trying to design for them. If you're in a team, one person could assume the role of a persona while the others ask him or her questions.
- Live with the work. Put your designs up on the walls of your project space, so that they're always visible. Yet another benefit of radical colocation.
- Sketch on the way. Always be engaged in the act of design. Start doing low-fidelity sketches from the first day of the project; don't wait until you have all the user needs "right" before starting to think about the product itself. Most of these sketches may be discarded, but they will keep you focused on how what you're seeing influences the design effort.
- Design for the presentation. John feels strongly about this, although it is controversial. He suggests focusing on the final presentation of your work (whatever that might be) from day one, and avoiding any work that isn't feeding into that final presentation. His rational is that if you don't keep focused on the final results, you'll waste time looking at details that aren't important for the current stage of your project. If they won't be in the presentation, nobody will see them, and by the time they're relevant, they'll be obsolete. Basically you want to do the right work at the right time (very XP-ish in this regard) by keeping focused on what your audience's current problems are and what will excite them.
- Generate scenarios or stories of use. Oftentimes, individual sketches can't capture enough information to express an interaction design. Scenarios of use can fill in the details by describing the entire course of the interaction. Stories are often the best low-fidelity way to prototype an interaction design.
- Get yourself in the right mindset. To create new ideas, you have to learn how not to think. Though you have probably accumulated a sophisticated understanding of your users' goals, tasks, workflows, processes, etc., you need to be able to let this go at times and let your intuition guide your design thinking. Too much understanding of how things are can impede your ability to think about how things can be. It's also important to foster a team environment where it's ok to come up with less-than-brilliant ideas, since only by seriously considering the less good (if not downright silly) ideas can you get to the ones that are gold. To do this, you need to be able to trust your teammates; if you don't, you won't feel comfortable expressing half-baked ideas and this will impede creativity.
- Say "Yes, and..." During brainstorming and ideation, it's important not to start critiquing and shooting down ideas too early. Instead, build on your teammates ideas to take them to the next level. If someone shoots out a half-baked idea, don't think "what a crappy design idea", think "how can I take this further"? Brainstorming should be fun, not competitive.
- Stand up. You're less creative if you're sitting down. This might be a personal preference, but I certainly find I'm at my best when I'm pacing the floor; somehow the physical activity jolts my mental activity into overdrive.
Finally, John had a couple of things to say about user research. He warned us that our interpretations of the underlying goals and needs we uncover during user observations are generally going to be wrong. On the other hand, users' perceptions of their goals and needs are also wrong. Effective interpretation of user research lies on reconciling these two to arrive at something right. To do this, you need to do more than just observe; you have to show users the concepts you've created and get them to react. You can talk forever about abstract goals and get nowhere, but the instant you put a real, working interaction design (or some low-fi version thereof) in front of a user, she'll immediately be able to say "Yes! This is exactly what I need!" or tell you what's missing.
Understanding needs must be complimented by making things. User research and interaction design need to walk hand-in-hand throughout the product life cycle.