Thread by @lunivore

X: We want the Spotify model.

Me: Have you thought of just starting with one team that can ship independently?

X: Yes, we've got one.

Me: That can ship independently?

X: Yes.

Me: When did they last ship?

X: Erm...

"Spotify model" isn't just about restructuring the people. You also need to restructure work (which means your prioritisation and funding), skills (which means hiring, career progression and contracts), and how you deliver (which means your environments, licences, security)...

...and restructuring how you market and get feedback (which means communication, monitoring & analytics), how you learn as an organization (KPIs and OKRs and skill sharing including training internal trainers), and that's just a small subset. And you don't know which comes first.

And that's why it's emergent; and why you can't just adopt the Spotify model or any other end-state and make it "just work"; because your context isn't like any other, and you have no idea which of those will move most easily, until you try.

Try it. Lots. Learn. Do it again.

Also look at other models. The Spotify model provides coherence; a realistic reason for thinking that your attempted improvements might be a good idea. It isn't the only coherent model. And all of them are wrong, so any of them might fail.

First step: make it safe to fail.