I remember a large group in the room, and nobody was going to budge.
We were meeting at the Cathedral of Our Lady of the Angels to help shape the Los Angeles Universal Preschool master plan. The room included early-education advocates, political and civic leaders, businesspeople, experts, and others who cared deeply about what we were building.
Strong beliefs met strong personalities. Everyone had an idea of what the preschool system should look like. Nobody wanted to compromise.
We were stuck. And we could not afford to stay stuck.
At the same time, I was taking a marketing class at Anderson and learning conjoint analysis, a technique for understanding how people choose among combinations of product features.
A light bulb went on.
What if a preschool system could be modeled the same way?
Instead of product features, we had system choices: teacher qualifications, full-day or half-day programs, ages served, how quickly to ramp up, location, and other decisions.
The field was different. The structure of the problem was similar.
So I borrowed the tool.
We used conjoint analysis to find patterns underneath people’s preferences. What I remember as many apparent alternatives resolved into three major clusters.
Then I connected another dot. GIS could show where unmet need was greatest. Financial modeling could show what each system configuration would cost over time.
When I presented the results back at the Cathedral, the room went silent.
I asked volunteers from each of the three groups to put forward their preferred configuration, and I modeled all three. Then I asked the larger group to choose a preferred option and a backup.
People could see their preferences represented. They could see where others clustered. And they could see the financial consequences of the choices in front of us.
The mathematics did not decide what people should value. It gave us a structure for seeing what we valued, where we agreed, where we differed, and what our choices would cost.
That was enough to break the gridlock.
I remember the central system-configuration decision becoming much clearer after that. Once we cracked that nut, the rest of the planning became more manageable.
The lasting lesson was not conjoint analysis itself.
Conjoint analysis came from marketing. GIS came from another discipline. Financial modeling came from another toolbox. Preschool planning was another world again.
But problems do not care where we learned our tools.
Do not ask what field a tool belongs to. Ask what problem it can solve.
The more widely you learn, the more dots you collect. The creative part is recognizing when ideas that were taught separately belong together.
That day, I did not need to have the winning answer. I needed to help the room see the choices clearly enough to find one.
Collect dots. Connect dots. Create something new.