Two days before the board review, a VP pinged me to "align" our churn definition with the number Finance was using. The polite framing was consistency. The actual difference was that my definition counted accounts that stopped using the product but hadn't formally cancelled, and Finance's didn't. Theirs produced a churn number about a third lower. Guess which one was going in the deck.
I've watched this movie before. "Alignment" almost always resolves in the direction of the more flattering number, and then six months later someone in a planning meeting asks why the forecast keeps missing, and the answer is that we defined success as not measuring the thing that was failing.
So I wrote up both definitions, what each one includes and excludes, and which behavior each one hides. I sent it to the VP, and I copied the analytics lead and the data PM, not to ambush anyone, but because once a definition goes in a board deck it becomes load-bearing and I wanted the choice to be on the record as a choice. I said I'd support whichever one leadership picked as long as the deck footnoted what it excluded.
The VP picked Finance's number. Fine, that's their call to make. But the footnote didn't make it in, and now I have a reputation for "making things complicated" right before high-stakes moments, which is a clean way to relabel "insisted the number mean something."
I think the documentation was right and I'd write it again. The part I keep second-guessing is the cc. I told myself it was about the record. But copying three people on a disagreement with a VP is also a power move, and I'm honest enough to know I wanted witnesses. Did I protect the data, or did I just make sure that when the forecast misses, it isn't my name on the convenient number?