Re:Ask HN: What are the best unknown books you have read?
by smogcutter on Sunday January 19, 2020 @09:04PM
(#22095653)
Sounds interesting. I’m sure the book covers this, but isn’t this how you wind up with perverse incentives? Once managers figure out that you’re measuring X as a proxy for Y “thing we actually care about” they’re going to start optimizing for X, with potentially unpleasant results.
[ Parent
| on HN ]
Re:Ask HN: What are the best unknown books you have read?
by crimsonalucard on Sunday January 19, 2020 @09:37PM
(#22095820)
This goes back to the whole correlation does not equal causation thing in statistics.
You're likely thinking of a correlated number, the measurement number must be a causative. Direct improvements on any causative number will by definition affect the outcome.
Say we're looking at developer retention. A correlated number is like scrum points. A causative number is like salary.
Of course, while the data on the causative and correlative effectiveness of both metrics is dispute-able... you still get my point.