The best business schools in the world teach with case studies. It is a strange method when you look at it directly. You do not give students the answer. You hand them a decision somebody actually faced, with the full situation around it, the constraints, the people, the pressure, and you make them reason their way to a call. Then you show them what really happened.
It works because a case study is a decision with all of its context preserved. That turns out to be rare enough to build an entire pedagogy around.
I wrote last time about decisions being the one thing companies never keep. What I have been chewing on since is a narrower version of that. Not what you lose when a decision disappears, but what the next decision loses. Because a case study is not only a record of something that happened. It is raw material for the person deciding now.
So I started pulling apart what actually goes into one, and it is the same set of things that goes into any good decision inside a company.
What it takes to reach a real yes
When a company approves something that matters, a defensible yes draws on a few kinds of context, whether or not anybody names them out loud.
There is the business context. The numbers, the market, the strategic fit. What this does to margin, to the roadmap, to where you stand against the people you compete with.
There is the people context. Who is affected, who is asking, whose judgment you trust on this particular question, and who carries the risk if it turns out badly.
There is the situation context. The timing, the pressure, the alternatives sitting on the table, whatever makes this decision specific rather than generic.
Most companies can muster those three on a decent day. A capable manager holds all of it in their head and makes a call you could defend in a room.
But there is a fourth input, and it is the one that separates a case study from a guess.
The pattern of what you already decided
A case study is never only the situation in front of you. It is that situation placed against the ones that came before it. How those were decided, on what reasoning, and how they actually turned out. That comparison is the whole point. Without it you are reading an anecdote.
Now look at how your own company decides. The business context is there. The people context is there. The situation is obviously there, sitting in the meeting invite. And where the pattern of prior decisions should be, there is nothing.
How did we handle the last five vendor switches. What did we say the last time somebody asked for this exact exception. What happened the three times we extended credit past ninety days. Those questions have answers. Real ones. They just are not anywhere you can reach.
So companies decide with three of the four inputs and a blank where the fourth belongs. Every case gets solved with the last page torn out, which is the page that says what happened last time.
Why this is the expensive one
The first three inputs are recoverable. You can pull the numbers again. You can ask the people. You can assess the situation in front of you, because it is in front of you. All three cost effort and all three are available if you are willing to spend the effort.
The fourth is different. It cannot be reconstructed on demand, because it was never kept in the first place. Those earlier decisions lived in an inbox, a corridor conversation, a nod, someone's memory that has since taken a job elsewhere. The one input that would make today's decision genuinely wiser, which is the accumulated judgment of every similar call the company already worked through, is precisely the input that is missing. Not buried somewhere. Gone.
Which is how you end up with the thing I keep seeing. Companies do not get better at deciding. They get older at it.
A company ten years in has faced the same category of decision hundreds of times. It can draw on almost none of it, because it kept every outcome and threw away every reason. Each new manager starts more or less from zero, holds three inputs in their head, and makes a call that a hundred earlier calls should have informed and did not.
Picture a business school that made each incoming class work the same case, and forbade them from reading a single line of what the previous twenty years had written about it. Nobody would call that education. It is close to how most companies decide.
The part that actually bothers me
There is a version of this that is worse than inefficiency.
When the reasoning behind past decisions is gone, a company cannot tell the difference between a decision that was good and a decision that was lucky. From the outside, once the reasoning has evaporated, those two look exactly the same. Both produced an outcome. Both are now a fact about the past.
So the company cannot learn. It can only remember results, which is not the same thing at all. It repeats what worked without knowing why it worked, which means it keeps repeating it after the conditions that made it work have quietly changed. And it avoids what failed without knowing why it failed, which means it avoids a whole category of perfectly good options for a reason nobody can articulate any more.
Experience is supposed to compound. What compounds is not the number of decisions you have made. It is the number you can still explain.
Where this leaves me
A decision is not only something you record after the fact. It is an input into every decision that comes after it. The pattern of what you chose, and why, and what followed, is the fourth context. It is the thing that turns years of operating into actual judgment rather than accumulated age.
You can only use that input if somebody kept it.
Which lands me back in the same room I keep ending up in. Somebody asks not just what should we do here, but what did we do last time and why. And the room goes a bit quiet. And somebody reaches for a decision that ought to exist, and finds nothing there.
That silence is where I want to go next.
Part of The Other Books, an ongoing series on the decisions companies forget.
