by Serguey Shinder
From about 2017 to 2020 I had, as far as I know, the best delivery record in our department. When I said a piece of work would be done by a date, it was done by that date, and usually a few days before. I was proud of it. My manager mentioned it in two reviews, and I think it was part of why I was trusted with larger things.
It was not a skill at estimating. It was a skill at never saying anything I might be wrong about. I would work out roughly how long something would take, then double it, then add a week for good luck, and give that as the date. When the work finished early, I either delivered early, which looked good, or found things to polish, which was easier to justify than it should have been.
I found out what that cost from a colleague in planning, over coffee, in 2020. She said, not unkindly, that they had a rule for my dates. They divided them by roughly two. They had worked out years earlier that my estimates were safe and therefore useless, and they planned around their own guess instead. Worse, two teams downstream of mine had spent part of 2019 waiting for work of ours that was finished and sitting on a branch, because I had said the end of the month and it was only the tenth.
So my record had been bought with other people's time. The accuracy I was proud of was not accuracy at all. An estimate that is never wrong has usually stopped saying anything, and the people who depended on mine had quietly stopped listening to it and replaced it with their own, which meant my name was on a number nobody used.
I had been treating an estimate as a promise, something I would be judged against. For everyone who received one it was information, the thing they used to decide what to start next, and I had been giving them the least informative version I could get away with, in order to protect myself.
What I do now is less comfortable. I give two numbers: the date I think is more likely than not, and the date I would be surprised to miss. The first is wrong about half the time, which is what it is supposed to be. When a piece of work finishes early I say so the same day. And I keep a plain list of what I estimated and what happened, which shows me whether my half-the-time dates really are wrong half the time, rather than whether I look reliable.
My manager at the time noticed the change and, to her credit, did not mind the misses. They were the first estimates of mine anybody had ever bothered to use.
– Serguey Asael Shinder
Leave a Reply