A number with no owner
I asked where a line came from and got a name. The name had left the company two years earlier.
I have inherited a lot of recurring reporting. A pack that goes out monthly, a set of figures that has been going out monthly for years, and my job is now to send it.
The first question I ask is where each line comes from. Sometimes there is an answer. Often the answer is a name, and the name left two years ago, and the figure has been produced every month since by a process nobody has looked at because it has never failed to produce something.
That is the state I want to describe. Not a wrong number. A number with nobody behind it, which is a different and much more patient problem.
Why unowned numbers drift
They are rarely born wrong. They are usually born correct and then the world moves underneath them.
A definition that was right when it was written stops matching the thing it describes, because a system was replaced, or a category was split, or a business unit was acquired and its data started arriving in a slightly different shape. None of that produces an error. The query still runs. The number still renders. It is simply now the answer to a question nobody is asking, sitting in a slot labelled with the old question.
If a person owns the figure, that person hears about the change through some other channel and thinks, wait, does that affect my line. That is the entire mechanism. It is not a control, it is not a validation rule, it is a human being who has a reason to connect two facts that live in different meetings.
With no owner, there is no such moment. The change happens, the number keeps appearing, and it can be wrong for years without anybody being informed, because being wrong in this way generates no signal at all. Errors of this kind do not get found. They get inherited.
The cheapest thing I know
Attaching a name to a figure costs a line of text and outperforms most of what gets built instead. Reconciliation checks catch a number that disagrees with another number. They cannot catch a number that agrees perfectly with a source that has itself stopped meaning what it used to.
The point of the name is not accountability in the punitive sense. I have never seen blame improve a reporting pack. The point is that a named person is a subscriber. They have a standing interest, so the news reaches them, and they are the only component in the system capable of noticing that a definition has quietly become inappropriate.
Two ways naming an owner makes it worse
Two ways, and both are real.
An owner will defend a number past its usefulness. Ownership creates attachment, and a person who has maintained a definition for four years is the person least likely to agree that it should be retired. I have watched an owned figure survive long after everyone had stopped using it for anything, precisely because it had somebody.
The worse failure is naming an owner who has no authority over the inputs. That manufactures a person who gets asked questions they cannot answer and cannot fix, which looks like ownership on an org chart and is closer to a complaints desk. It also lets everyone else stop worrying, which is the most expensive part. A figure with a nominal owner is less examined than one with none, because the absence of an owner is at least visible.
So the name has to come with the ability to change the thing, and that is a much larger ask than adding a line to a footer, which is why it usually does not happen.
Most reports I open carry neither a name nor a date. Both would fit on one line at the bottom, and I notice their absence far more than I used to, and I have still shipped reports without them.