“I start by treating the request seriously, then inquire into the work it is meant to do: what prompted it, what has happened before, who sees it differently, and what would change if this went well. A neat issue tree or MECE framework can help organise the inquiry, but the real risk is that clean categories hide the relationships, history, and conflicting purposes that make the problem hard.”
“I learned at Deloitte and Accenture that the smartest analysis fails if you solve the wrong layer. Now I start by asking what belief or operating pattern is actually driving the problem, because once that is visible, the structure almost builds itself.”
“I spent years at BCG and McKinsey being paid to structure problems, then realized the hardest part is not the framework, it is naming the real question everyone is avoiding.”
“At Bain I learned the hard part isn’t the analysis, it’s getting the client to name the real question they’re afraid to ask.”
“I start by asking what decision this problem is supposed to produce, because a hard problem that isn't anchored to a real choice just spins. At BCG and KKR I saw too many teams build elegant frameworks that answered everything except what the CEO should do on Monday morning.”
“I stopped trying to solve hard problems and started asking what kind of problem I was looking at. A lead counsel on one megaproject told me most teams burn weeks on crossword puzzle problems — real, satisfying, and 100% irrelevant to the outcome.”
“I spent years at McKinsey delivering bright, rigorous recommendations that landed flat because the client’s people never owned the problem the same way we did. Now, when I structure a hard problem, I drag the team through diagnosis first: not a deck, but a conversation about prior commitments that failed, the bottlenecks they created, and the cause-and-effect chain nobody wants to say out loud.”
“At KPMG I learned that consultants love a clean answer, but the real value is in the structure you build before you answer. I now start any hard problem by asking what the system is actually optimised for, because once the purpose is clear the messy parts sort themselves.”