Thread by @johncutlefish

Fascinating q on an @Amplitude_HQ office-hours call today:

"At what stage is it important to worry about measurement?"

Specifically, they were asking about making information accessible & usable to/for front-line product development teams. So...(1/5) #prodmgmt #ux #design

My answer was that anytime you're making decisions under conditions of uncertainty ... it will probably help to have trustworthy/usable data. Imagine a team making 10-15 major decisions each year, each driving 10-15 medium decisions, each driving 10-15 small decisions..(2/5)

That is 3375 decisions driving 16,640hrs of team hours. I saw @gibsonbiddle give a talk recently about a string of decisions at @netflix ... and the punchline is that it took a couple tries. No one is perfect. So...(3/5)

What if you could improve decision quality/velocity by a meager 10%? Or cut an 8w effort short to 4w? Or make marginally better decisions on where to double-down? Or hit 9/15 instead of 5/15. That is huge. It isn't a stage you get to. (4/5)

If all you do with the data is use it to run retrospectives on prescriptive decisions and use that to slightly improve decision quality...you'll come out ahead.

So start. (5/5)