I said last time I wanted to look at the bill. Not the moment somebody asks why a rule exists and the room goes quiet, which at least announces itself, but the cost underneath it, which arrives continuously and never appears on anything anybody reviews.
The bill has the shape it does because of something in the literature, so I have to start there. Briefly.
Where the maps end
There is a serious body of work on decision making and most of it is good. Herbert Simon argued that an organisation is essentially a decision making system and gave us a structure for the act itself. Behavioural economics told us what goes wrong inside the decider. Engineering went furthest and formalised it into a standard, ISO 15288, which defines a decision management process as a structured way to identify and characterise and evaluate a set of alternatives and select the most beneficial course of action.
Read that definition again and notice that every verb in it happens before the decision exists.
That is the pattern across all of it. Between them these traditions have mapped how a choice should be structured, who should be involved, how uncertainty should be handled, and which biases distort it. Then the decision gets made, and the map ends.
There are exceptions and I should name them rather than pretend. After action reviews are a real discipline. Software teams run retrospectives. Some engineering groups keep architecture decision records, which is the closest thing I have found to a genuine second half practice, and which emerged from engineers getting burned rather than from any theory.
But those are optional, done by teams that happen to care, and connected to nothing. The first half is a discipline with standards and qualifications behind it. The second half is a few good local habits that stop when the person who championed them moves on.
Why it stopped there
Two reasons, and neither is negligence.
Outcome quietly became the substitute for reasoning. If you can see whether a decision worked, you appear not to need why it was made. That substitution feels harmless and is not, because outcome and reasoning are different things, and the gap between them is exactly where learning would happen.
And the discipline that might have owned the second half already existed, filed under clerical. Records management, retention, archives. Real expertise, universally treated as a compliance function. Nobody connected the craft of keeping things to the craft of deciding things, because one sat with the lawyers and the other sat with the executives and they were never in the same conversation.
The bill
The cost of an unmapped second half is not one number. If it were, somebody would have found it, and there would be a budget line and a person responsible.
It is invisible for a structural reason. It gets paid in small amounts by many people, none of whom can see the others paying.
Twenty minutes looking for a decision that was already made. An hour of a colleague's time helping look. A meeting where something settled last year gets reopened and settled slightly differently, because nobody in the room knew it had been settled. A manager applying a rule she cannot explain. A request deliberately made smaller than it needs to be, because a clear yes is not available and a smaller ask is a smaller thing to be exposed on later. Four months of a new joiner rebuilding an understanding that could have been handed over in an afternoon.
None of that has a cost code. Nobody submits a claim for it. It is absorbed into salaries that were being paid anyway, which is exactly why it can run for a decade without anybody objecting.
And it compounds. Every decision that goes unkept raises the cost of the next one slightly, because the next one has less to stand on. A company ten years into this is not carrying ten years of the same annual cost. It is carrying an increasingly thin foundation under an increasingly large number of decisions.
The part that actually costs money
Here is what convinced me this matters more than it looks.
When the second half is missing, the symptoms never present as a decision problem. They present as something else, and companies spend real budget treating the something else.
Work stalls in odd places, so it gets read as a project management problem and somebody buys another tool. Two functions turn out to be working from different understandings of the same agreement, so it gets read as a communication problem and somebody runs a workshop. Good people leave saying nothing here makes sense, so it gets read as a culture problem and somebody commissions a survey. Execution keeps drifting from intent, so it gets read as an alignment problem and somebody reorganises.
Each of those remedies is expensive and each is aimed at a symptom. The condition underneath is that the organisation cannot reliably find, trust or reuse its own decisions, and no tool or workshop or reorganisation touches that. Which is why the same symptoms come back in eighteen months in slightly different clothes and get treated again.
The wasted minutes are the interest. The expensive part is a decade of solving the wrong problem while the actual condition compounds underneath.
Where this leaves me
If your problem is how to structure a difficult choice, the literature will serve you well. A century of good work is available and you should use it.
If your problem is that the choice you made two years ago is now unrecoverable, there is nothing. Not because it matters less, but because it never looked like a research question and got filed somewhere nobody strategic was looking.
That is the gap I have been working in. Record, Retrieve, Rely, Reuse.
Next I want to take the first of them seriously, because Record sounds like the easy one and is the one almost everybody gets wrong, usually by assuming that having written something down somewhere is the same as having recorded it.
Part of The Other Books, an ongoing series on the decisions companies forget.
