I’ve been thinking about Michael Kennedy’s LSSC’12 talk, Set-Based Decision Making – Taming System Complexity to Ensure Project Success and ways to explain it to those who weren’t there at the conference.
Kennedy’s ideas go well beyond set-based design or pursuing multiple designs guaranteeing that at least one will succeed. He identified the Wright brothers as the first (probably, unconscious) practitioners of set-based decision making. While other aspiring airplane inventors spent a lot of money following the design-build-test-repeat process, the Wright brothers did something very different: test and experiment, repeat until feasible, only then design, then build. This approach was later independently discovered and (perhaps, unconsciously) practices by many innovators in the last one hundred years – the rapid design of and the market conquest by Toyota Prius is one of the best examples.
Set-based Decision-Making = Set-based Design Multiplied by Real Options?
In set-based design, we work on two or more designs of the same product simultaneously. The date when the design must be finished is fixed and cannot be moved. We don’t know which design will succeed, but we know that one of them will.
In set-based decision making, the same is done on the set of options that can be later used in designing the product. Instead of designing things, design options to design things! Instead of scheduling the date when the design must be complete, you schedule the date when the design will begin. And it begins by selecting some options from the set of feasible options and converting them into obligations. These obligations fix certain design parameters and a more deterministic design process follows, followed by building the product. The customer value is pulled through this workflow by market demand. Do not commit to any particular design options, or, in other words, don’t convert them into obligations until The Date. Once the design begins, the portfolio of options carries over to the next generation of the product.
To summarize:
Set-based design | Set-based decision making | |
The Set | designs | options |
The Date | when the design must be complete | when feasible options must exist and the design must begin |
The Decision | select the (most) successful design | select design parameters from the set of feasible options and commit |
What follows | build or delivery | design |