Thread by @johncutlefish

"what do we wall thing about being more data-informed?"

AMAZING

1/n cool, so I'd like to replace my feature-focused roadmap with a list of experiments and a target input to impact

"wait a sec..."

2/n yeah, and when we find better measures for things, we can iterate.

"hold on, we can't change wha...."

3/n we'll double down our investments in the things that are impactful, and teams can disband and reform around what's working

"hold on a minute...nooooooooo"

4/n yes, and we'll look for opportunities to prune and kill features that are no longer helpful to customers, and are getting in the way of us innovating on behalf of customers

"um, yeah, so, THAT, I'm not sooo....sure...."

5/n analysts, product managers, designers, and engineers can collaborate on doing exploratory analysis before jumping into building a feature

"have you seen their calendar ... um ... hmmm"

6/n it gets better. Instead of tacking on OKRs on to existing projects, we can tie in OKRs to more consistent and stable metrics. That will make things so much easier!

"yeah, um, so, about those projects...."

7/n and we'll mix and match the best approaches -- qualitative and quantitative -- to reduce uncertainty where it matters!

"yeah, the UX researcher is kind of busy ...."

8/n it'll be great ... we'll have a healthy data culture instead of using data to prove whatever point you want to prove, and back up whatever decision you want to back up

"OK, so I have this deck to present"

9/n we'll shift from using vanity metrics, even though the VMs go nicely up and to the right!

"but the slide only has three spots..."

10/n ...

tl;dr it is hard.

start somewhere.

it is something you need to practice.

and it isn't a math problem.

it is a human problem.