A budget is a promise, not a forecast
The reason a missed budget feels like a betrayal rather than an error is that it is one.
Every organisation I have worked with has a line in its budget that everybody knows will be missed.
It is different in each place. Repairs and maintenance, set at last year’s figure in a building that is visibly ageing. Training, set at a number that would be generous if anyone ever booked anything, and nobody does. A revenue line that requires a rate of growth the business has never once achieved. The line goes in, it is approved, it is missed by roughly the same margin it was missed by last year, and the following autumn it goes in again.
I saw it across a book of small and midsize clients, and then in a much larger setting, across business units that had been acquired and were being brought onto one reporting structure. Same behaviour. It is not a competence problem, because the people involved were not stupid and could see the pattern as clearly as I could.
It took me a long time to accept the simplest explanation, which is that these numbers are not attempts at prediction and were never being judged as such.
What the number is for
A budget line is a social artefact. It records a commitment, and the interesting question about any given line is which commitment, and to whom. A schedule is the same kind of object and is harder to argue with, because nobody builds one intending it to be read back.
A cost line held flat is a promise to whoever provides the money that things are under control. It is aimed at a lender, an owner, a parent company, sometimes at a board that will look at it for ninety seconds. The number’s job is to be reassuring and defensible, and its accuracy is a secondary property.
An ambitious revenue line is a promise pointed the other way, at the team. It says what we intend to be, and setting it at the realistic figure would be read, correctly, as a statement that nobody expects much this year.
A generous training line is a promise about values, made to people who will never check the actual spend and would notice immediately if the line were cut.
Once you see the lines that way, the emotional temperature of the variance conversation makes sense. A missed forecast is an error. Errors are uncomfortable and survivable. A missed budget is a broken promise, and it is discussed in the register of broken promises, with the defensiveness and the blame allocation that go with that. People are not confused about what happened. They are responding accurately to the kind of thing it was.
Two years of fixing the wrong artefact
My response, for a long time, was technical. Better inputs. More granularity. Drivers instead of last year plus three percent. Build the thing properly and the misses will shrink.
The models got better. The outcomes did not move at all, because I was improving the predictive quality of an object whose function was not predictive. It is the equivalent of tightening the tolerances on a part that is not load-bearing.
Then I overcorrected, which I think is the more common failure among people who have noticed this. I decided budgets were theatre and stopped taking them seriously as constraints. That is worse, and it is worse in a specific and avoidable way: a promise, even a badly chosen one, coordinates real behaviour. Somebody in that organisation is defending that line. If you have privately classified it as fiction and spent against it accordingly, you have not transcended the game, you have become the reason a colleague has an unpleasant meeting in February.
What I ask instead
I do not have a better instrument to offer. I am not sure a better one exists, because the social function is doing something the organisation needs and no purely predictive object would do it.
What changed is the question. When a line looks obviously wrong to me now, I no longer ask what the right number would be. I ask who agreed to it, and what they got in exchange for agreeing.
Sometimes the answer is that somebody got a quieter meeting, or a lender’s signature, or a year without a conversation about headcount. Occasionally the answer is that nobody got anything, that the line was inherited and nobody present remembers agreeing to it at all. Those are the ones that turn out to be movable, and they are rarely the ones anybody is arguing about.