The short answer: Two departments at one company argued about the same thing for years, and everyone treated it as friction between difficult people. They had been given targets that could not both be met: one was measured on volume, the other on margin, so every discounted deal was a win for one and a loss for the other. Every department was hitting its number and the company was losing margin because of it.
The first thing anyone told me about those two departments was that they did not get along.
Not as a complaint about the business. As context. The way you might mention that a building has an awkward staircase. Sales and service argued, and had done for a long time.
I was there about margin, which was thinner than anyone could account for.
It took a while to notice that those were the same subject.
What they were each asked to do
Sales had an annual sales target. Volume.
Service had an annual gross margin target.
Read those two sentences again, because that is the entire finding, and it took me longer than it should have to see it.
Sales was measured on closing. So sales closed, at whatever price closing required. That is not misconduct. That is a person doing precisely the job described to them, well enough to be rewarded for it.
Those deals then went to service to deliver. Service now had to produce a margin on work that had been sold too cheaply to produce one. Also not misconduct. Also a person doing exactly the job described to them, and failing, through no decision of their own.
Both targets could not be true at the same time. Not “were in tension.” Could not both be met, arithmetically, because one was being funded out of the other.
The argument was the evidence
Here is the part I think about most.
Everyone in that company could see the conflict. It was the most visible thing about the relationship between those two groups.
And it had been filed, universally, as a personality problem. Two teams that did not get on. Something to be managed with better communication, or a workshop, or by waiting for someone to leave.
Nobody had read the argument as data.
The fight was not noise sitting on top of the real problem. The fight was the problem, expressed by the only mechanism available to it. Two groups of people were standing on either side of a contradiction that nobody had written down, and the contradiction was pushing them into a room to argue about it every quarter.
A conflict that recurs, with different people in it, is not about the people. If you could replace everyone involved and the argument would still happen, then the argument was never theirs.
What is the folly of rewarding A while hoping for B?
I want to be clear that I did not find something novel.
In 1975, Steven Kerr published a paper in the Academy of Management Journal called On the Folly of Rewarding A, While Hoping for B. It catalogued organizations that systematically rewarded one behaviour while expecting a different one, and it has been reproduced and taught more or less continuously since. The reference is 18(4), 769 to 783, if you want to look it up properly. It is behind a publisher’s paywall, as is the reprint the Academy ran twenty years later as an Academy Classic, 9(1), 7 to 14. That tells you what the field made of it, even though neither copy is free to open.
So the pattern is not obscure. It is fifty years old and famous.
It keeps happening anyway, and the reason is structural rather than intellectual. Targets are set department by department. Nobody owns the interaction between them. Each target is sensible when read alone, which is the only way anyone ever reads them. A sales target is approved by whoever owns sales, and a margin target by whoever owns service. Both are reasonable documents, approved reasonably. There is no meeting whose purpose is to read all of the targets at once and ask whether a single person could satisfy them.
How can every department hit its target while profit falls?
This is what makes it survive.
Every department was green. Sales hit its number. Service was accountable for a number it kept missing, but that read as a service problem, and service problems get service explanations: execution, staffing, delivery.
The company was losing margin because its departments were succeeding. There is no line on a dashboard for that. You cannot build an alert for “all targets met, outcome worse,” because the reporting layer is built out of the targets themselves. The instrument and the fault share a definition.
That is why this needs an outsider, and not a cleverer insider. Seeing it requires standing outside every department at the same time, and nobody who works there is standing there.
How do you find conflicting targets in your own company?
It takes an afternoon and it is uncomfortable rather than difficult.
Put every department’s annual target on one page. One line each, no commentary.
Read them as a single instruction to a single person. Would that person have to do contradictory things to comply?
Then ask where the recurring argument in your business happens. Not the one-off disputes. The one that comes back every quarter with different names attached.
If the argument sits on the seam between two of those targets, you are not looking at a personality problem, and no amount of better communication will fix it, because everyone involved is communicating perfectly. They are telling you, at volume, that you asked for two things.
The fix was a line in a document
Sales got the gross margin goal as well.
That is the whole intervention. Not a reorganisation, not a mediation, not a new process. Once selling at any price stopped counting as a win, the contradiction stopped being generated, and the argument it had been producing stopped having anything to feed on.
I never saw the after number on margin, so I am not going to tell you what the change did to it. What I can tell you is that the arithmetic was demonstrable, on their own numbers, and that the change was made.
So what
If two teams in your business argue about the same thing every quarter, stop asking who is being difficult.
Go and read what you asked each of them to do, side by side, and check whether you have asked for two things that cannot both happen. The argument is not the problem to be managed. It is the report, arriving on time, from the only people in a position to file it.
If you would rather someone from outside read those targets with you, that is what I do for money, and I will tell you when the answer is that you do not need me.
Everything above is first hand. No client is named, and identifying details have been omitted or generalized. Every external source here is a published paper, cited in full above so you can find them whatever happens to the links.
Related reading
- The Deal Was Already Lost: what to check when a good argument has not moved in months.
- They Already Owned the Fix: the same pattern in a different failure, where the capability had been bought and never switched on.
- Being Right Is the Cheap Half: why seeing this is the easy part, and being believed is what costs money.