Thinking Lab · 4 lessons

Problem Solving

Break complex problems into manageable parts Most failed solutions solved a real problem that was not the problem. Time spent defining is almost never wasted.

Free — no account needed to read

Break complex problems into manageable parts

In one line: Spend the first effort on what the problem actually is, because that is where most of the leverage sits.

When you need it

• The same issue keeps coming back after being fixed.
• Everyone agrees on the solution and nobody has stated the problem.
• You are busy on something and cannot say what would count as it being solved.

The common misreading

What people think

A well-defined problem is one with a clear solution in it.

What is true instead

A problem statement with a solution inside it has already stopped the search. "We need a new tool" is an answer wearing a question mark.

After this module you can state any problem as the gap between what is and what should be, with no solution in the sentence.

The key idea

Most failed solutions solved a real problem that was not the problem. Time spent defining is almost never wasted.

Start with a hard question

Write it exactly as you normally describe it. Then ask whether that description contains a solution already.

What problem have you tried to solve more than twice without it staying solved?

Answer it before reading the lessons — then again after, and notice what changed.

The lessons

• Define the actual problem — State the problem as a gap between what is and what should be. If it contains a solution, it is not yet a problem statement. (6 min)
• Decomposition, root causes and constraints — Ask what produced this until you reach something you could change. Blame is not a root cause. (7 min)
• First principles and generating alternatives — Generate first, judge second. Doing both at once reliably produces two options and no insight. (6 min)
• Evaluate, test, learn, iterate — Write what you expect before you test. Unrecorded expectations always turn out to have been correct. (6 min)

Practise it

Short exercises. Under each is what a considered answer tends to contain — not a mark scheme.

Break it down

"Our customers keep leaving." Break this into parts that do not overlap and that together would cover the whole thing.

• Tends to contain: Which customers, leaving when
• Tends to contain: Acquired wrong vs served badly vs outgrown you
• Tends to contain: Parts that could each be investigated separately
• Tends to contain: Something measurable in each part

Find the real constraint

You want to change career but "there is no time". List every constraint, then mark which are fixed, which are choices, and which you have not checked.

• Tends to contain: Hours are a choice about allocation
• Tends to contain: Money and obligations may be genuinely fixed
• Tends to contain: At least one constraint that has never been tested
• Tends to contain: The binding constraint is often not the stated one

Five whys, honestly

A report is late every month. Ask why five times, and stop only when the answer is something someone could change.

• Tends to contain: Each why produces a different answer, not a restatement
• Tends to contain: It does not end at a person being careless
• Tends to contain: The final cause is something changeable
• Tends to contain: Notice where the chain wanted to stop early

Reflect

Which problem turned out to be a different problem once you wrote it as a gap?