Think of the thing that keeps coming back. The same argument with the same person every few weeks, in slightly different words. The money that is somehow gone by the 20th no matter what you earn. The room in the house that ends up a mess within a fortnight of every clear-out. You have solved each of these before โ that is the strange part. You fixed it, it worked, and then it came back, which is the clearest possible sign that what you fixed was not the thing causing it.
A problem solving framework separates finding the cause from finding the fix, because most people jump straight to a fix and end up solving the wrong problem faster. The problem solving steps below start by tracing the symptom backward, then check the leading fix against the real incentives in the system before spending resources on it. Problem solving methods that skip the cause-tracing step tend to produce a longer list of patches, not fewer problems.
Write down exactly what's going wrong, when it started, and who it's affecting. 'Sales are down' isn't precise; 'renewals dropped for accounts signed after the pricing change' is.
Ask why, five times, each answer more specific than the last, until you hit something you can actually act on. A five whys pass usually ends at a process or an incentive, not a person.
Do this step now โ Five Whys โBefore you fix anything, map the incentives of everyone touching the process. A fix that ignores an incentive map gets quietly worked around within a month.
Do this step now โ Who's Paid For This? โDon't copy the last fix that half-worked. Run a first-principles breakdown of what actually has to be true for the problem to be solved, then build the fix from that instead of from habit.
Do this step now โ Down to the Bricks โApply the narrowest change that addresses the root cause, then watch the original symptom, not just the fix. If the symptom comes back, the cause wasn't what you thought it was.
Take a team whose deploys keep breaking the same billing integration every few weeks. The team's first instinct is to add another manual check before each deploy, which is a symptom fix. Running five whys instead traces the failure to a shared test environment that's rarely updated, which traces further back to no one owning that environment. An incentive map shows the two teams sharing the environment are each measured on their own release speed, so neither has a reason to maintain shared infrastructure. The real fix is assigning ownership of the test environment to one team's roadmap, not adding another manual check to a process that already has too many.
A typical structured problem solving process runs: describe the symptom precisely, trace it back to a cause, check the incentives involved, build a fix from that cause, then confirm the symptom is actually gone. Skipping the cause-tracing step is the most common way this breaks down.
Problem solving starts from something broken and works backward to a cause; decision making starts from a choice between options already on the table. The two often meet in the middle, once problem solving produces a short list of fixes to decide between.
Recurring problems usually get treated as one-off incidents, so each one gets its own quick patch instead of a shared cause. A five whys pass or a root-cause analysis is what stops the pattern, because it names the cause once instead of rediscovering it every time.
If fixing it would also prevent two or three other problems you've seen, it's probably close to the real cause. If the fix only prevents the one thing you're currently annoyed about, keep tracing.
Short daily drills on real decisions ยท free to start, no signup to try
Try a drill โ no signup โ