Eigenquestions: The Art of Framing Problems
coda.io
coda.io
You run a grocery store and you'd like to increase the frequency of customer visits, what do you do?
You are designing a game to teach kids math in a fun way, where do you start?
You'd like to shift your company's communication patterns from being primarily "synchronous" to primarily "asynchronous," what do you do?
First, I can brainstorm many questions:
• Is this a physical game or a computer game?
• Is this a solo game or a group activity?
• Is it a competition (with a winner and a loser) or a puzzle?
• Are we playing at school or at home?
• What kind of math are we learning?
• How old are the kids?
...
At this point, it feels like there may be a huge solution space. So I think the next step is to step back and look for relationships between the questions:
• There's a relationship between age and the kind of math.
• There's a relationship between physical vs. computer, home vs. school, solo vs. group.
• There's a relationship between solo vs. group and puzzle vs. competition.
Based on this, I think maybe we can simplify a 6+ dimensional solution space down to 2-3 dimensions:
1. Let's call the x-axis "What grade level are we teaching?" From that, we can make solid guesses about both age and what kind of math.
2. Let's call the y-axis "How large of a group is the game for?" I think this is the essence of the questions about where the game is being played, whether it's a competition, whether it's a group activity, etc.
This gives us a framework to start mapping the solution space. Near the origin of our graph (low grade level, small group) maybe we have a puzzle to teach arithmetic. If we go up the y-axis (low grade level, large group), maybe we have some sort of "math bingo" where the teacher shares basic arithmetic clues and students mark the corresponding spot on their board. If we go out the x-axis instead (high grade level, small group), maybe we create a math-heavy card game to explore probabilities. If we go out to high grade level, large group, maybe we need some sort of math-based scavenger hunt.
I don't think these specific ideas are actually very good, but I think the process helps take a very open ended question and start to dissect it along a couple of key dimensions to help focus on a more tractable problem; designing one game that works for all of those scenarios would be hard.
For a simplistic definition, the eigenquestion is the question where, if answered, it likely answers the subsequent questions as well. Great framing starts by searching for the most discriminating question of a set — the eigenquestion.
Then taking turns each person gets up and places a couple of the questions or features they have generated on a wall or whiteboard, briefly explaining what they mean while also placing it relatively near or far in conceptual space from the items that are already up there.
As you get further along, you invite people to start reorganizing the sticky notes talking about why they think some of these questions are more or less similar to others. As they explain their reasoning you start hearing the underlying latent dimensions, the organizing vectors.
When the process is all finished you have what looks like a cluster analysis. Then as a group we name those clusters. This qualitative dimension reduction exercise often has the result of getting a group of people talking about the question or feature space in a common set of reduced labels and gives them a common framework.
I found these techniques very enlightening to use with product groups as we prepared to use research techniques to capture a Voice of the Customer.