A promotion round has just closed. Somebody good, who has been waiting two years for this, did not make it. Their manager has to explain why, and the explanation is sitting right there in the framework, so she reads it out.
At this level we look for evidence of strategic influence, not only delivery.
The person listens to that and asks a perfectly reasonable question. Who decided that. Why is strategic influence the thing that separates a five from a four, and why does it outweigh two years of delivery that nobody is disputing.
And the manager, who is not being evasive and would answer if she could, realises she has no idea.
She says she will find out. She means it. This is the story of what she finds.
The search
She starts with HR, which is the obvious place, and HR is helpful. They send her the framework document. It is a good document. Eleven competencies, five levels each, behavioural indicators for every cell, the whole thing laid out cleanly in a deck with the company's fonts on it.
What it does not contain, anywhere, is why. Not one line explaining what problem this framework was solving, what alternatives were considered, why eleven competencies rather than six, or why strategic influence sits where it sits.
The deck has a date on it. March 2019. It also has a file name that ends in v7, which she notices and finds slightly unsettling, because it means there were six earlier versions and every one of them presumably represented somebody changing their mind about something.
So she asks who put it together. The answer is a working group. She gets four names. One of them is now the CHRO, and two have left the company. The fourth is on the operations side and remembers being in some of the sessions but says, honestly, that HR ran it and he mostly nodded.
She goes to the CHRO, who is generous with her time and genuinely tries. And what the CHRO gives her is a memory, not a record. Something about wanting to reward people who think beyond their own function, because at the time the company was very siloed and there was a feeling that we were promoting good operators into roles that needed something else. She says she thinks there was a consultant involved early on. She is fairly sure the eleven competencies came from a model that was adapted rather than built from scratch, though she could not tell you which model.
Then she says the sentence that ends the search. That was six years ago, I would not want to swear to any of it.
She does find a name, and it does not help
Here is the part that surprised me the first time I watched a version of this happen.
She does eventually get an answer to who. The framework was signed off by the CHRO of the time and the then chief executive, in a leadership meeting in April 2019. There is even a slide in an old deck confirming it. Approved.
And it changes nothing.
Because the moment she has the name, the question immediately becomes the next one. Approved on what basis. Was the weighting deliberate or inherited from whatever model it was adapted from. Was strategic influence meant to be a bar for everybody at that level, or a signal for a particular kind of role that has since stopped existing. Did anybody consider what it would do to people in delivery heavy functions, where the opportunity to demonstrate strategic influence is structurally lower.
Nobody asked who because they wanted a name. They asked who because who is the only part of a decision people assume must still be recoverable. It is the handle you reach for. What everybody actually wants is the reasoning, and asking for the reasoning directly feels like a challenge, whereas asking who approved it sounds procedural.
So the name arrives and the room is no better off, which is how you know the name was never the thing.
Some decisions are load bearing
Most missing decisions affect one thing. A vendor was chosen. A discount was given. A clause went into a template. Losing the reasoning is annoying and the damage is bounded by whatever that one thing touches.
This is not that.
That framework has been running for six years. It has shaped every promotion decision in the company since 2019. It sits underneath every appraisal conversation, calibrates every hiring bar, and quietly determines who gets stretched and who gets parked. People plan their careers around it. Some of them leave over it. Thousands of individual judgments have been made on the authority of a document whose own authority nobody in the building can now explain.
That is a load bearing decision. Not one choice, but the foundation that other choices are made on top of, every week, by people who assume somebody once had a good reason.
And here is what makes it genuinely uncomfortable. The manager reading out that line about strategic influence is not being arbitrary. She is being exactly as consistent as the company asked her to be. She is enforcing a rule she cannot justify, in front of somebody whose career it affects, and she has no way out of that position because the alternative is to admit that the criteria are not anchored to anything, which is both true and unsayable.
Most organisations put hundreds of their managers in that position without ever noticing they have done it.
Who pays for it
The cost of that one unrecoverable decision is spread across more people than you would expect.
The person who did not get promoted takes away something worse than a no. They learn that the criteria are not anchored, which means next year's answer will not be anchored either, which means there is no version of their own effort that reliably produces a different outcome. That is the point where good people stop trying to understand the system and start either gaming it or leaving it.
The manager now knows something about her own company that she cannot unknow, and has to keep applying the framework anyway. Nothing in her week changes. She just does the job with slightly less conviction, which is not a thing anybody measures.
The CHRO is in the strangest position of all. She was in the room in 2019. She was probably one of the people who argued for it. And she cannot reconstruct her own reasoning well enough to defend it six years later, which is not a failure of character, it is what happens to reasoning that was never written down next to the decision it produced.
Whoever inherits it next will treat the framework as a given, because by then it will be old enough to look like a fact about the company rather than a choice somebody made. That is how a decision becomes furniture.
The same thing, everywhere else
Once you have seen it in a competency framework you start seeing it in places that look nothing alike.
The approval limit that lets a country head commit up to a certain figure and no further, where the number was set in a different currency environment by people nobody has worked with. The rule that this category of purchase needs three quotes and that one does not. The renewal invoice for two hundred software seats when about thirty people log in. The customer on the same discount for four years, where the volume commitment that justified it expired somewhere in year two. The indemnity clause in every contract you sign that costs a week of negotiation each time and traces back to one deal, one bad experience, one lawyer, in some year nobody can name.
All decisions. All still running. All generating consequences daily. And in every case the reasoning that made them sensible at the time, which is the only thing that could tell you whether they are still sensible now, is gone.
It is not a records problem
I used to think of this as a documentation failure, which is a comfortable way to see it, because documentation failures have obvious remedies. Write it down. Keep a register. Be more disciplined about minutes.
I no longer think that is what it is.
That company had documentation. It had a framework deck at version seven, a leadership meeting slide confirming approval, six years of appraisal data, and a shared drive with more files in it than anybody could read in a career. Nothing was missing in the sense of being lost. The material was all there.
What was missing was the ability to walk up to one specific decision and pull it out whole, with the reasoning still attached, on the afternoon somebody asked. That is memory rather than paperwork. Organisational memory is not the sum of the files, it is whether the organisation can retrieve the right thing at the moment of need, and no amount of storage produces that on its own.
A company with immaculate documentation and no memory looks exactly like a company with none at all, from inside the conversation where somebody asks why.
The measurement
I have come to think the few seconds before somebody says I will find out is the most honest measurement of this that exists.
You cannot argue with it. There is no dashboard, no assessment, no maturity score, nothing that can be presented in a way that takes the edge off. Somebody asked the company a question about itself, and the company could not answer.
If you want to know whether your organisation keeps its decisions, you do not need a framework for it. Wait for the next time somebody asks why a rule is the way it is, and watch what happens in the four seconds afterwards.
It will happen. Probably this month.
What I want to look at next is the bill for all of this. Not the moment of not knowing, which at least announces itself, but the cost that arrives quietly and continuously and never once shows up on anything anybody reviews.
Part of The Other Books, an ongoing series on the decisions companies forget.
