The question · Boundary
Where does the product actually end?
The visible product may be only one part of what has to work for the outcome to happen.
Earlier
I tended to draw the boundary around the thing being designed.
The product, platform, course, service, or experience was the most visible unit of design. If that object worked well enough, the remaining problem could appear to sit mainly in use.
Later
More of the outcome began to sit outside the object.
Capability, process, institutional conditions, incentives, adoption, and supporting systems repeatedly affected whether the designed object could produce the intended result.
The object still mattered. Its visible boundary simply stopped matching the boundary of the problem.
What this revealed
The boundary of the product is not always the boundary of the problem.
A useful object can depend on conditions that are not contained inside it. That makes the act of drawing the product boundary a design decision in its own right.
Now
Where does the outcome depend on something outside the object being designed?
I now tend to ask whether that dependency belongs inside the design problem — or whether it should remain an explicit constraint around it.