I’ve been using Accelerate a lot but not as a guidline to change the org but more like confirmation that we are heading into the right direction and focusing on right thing. It also helped me a lot during budging phase when questions about the efficiency and performance of tech teams popped out. Anyway focusing on key characteristics of highly performing teams is a way to do it imho.

I'm running into problems where people are using the four metrics as the way to measure change. Your approach sounds much closer to what I would expect.

I’m tracking all 27 capabilities laid out in https://www.devops-research.com/research.html

as capabilities

then identify capability gaps (for each where we need to improve?) and create initiatives out of bridging those gaps. These get prioritized and becomes our OKRs

Have you mapping those to business objectives?

Thats my general issue with some ways it’s applied: the capabilities aren’t abstract good things, they’re things that commonly are important to enable specific business objectives and the specificity of how they correlate ought to be the thing that drives prioritization of those capabilities

So the way I prioritized that, is roughly by “pain”. Not all capabilities have the same importance from where we seat. For instance, this quarter, we focused on dismantling our QA team to embed them in our development team and remove the handoff between department that was causing a lot of trouble/pain. We had OKR around ownership of deployment (Dev led), increasing testing automation skills, transfer of knowledge, … We learned a lot during the quarter and will be feeding these learning in our next quarter OKRs. So we start from where we are, focus on the pain, increment our way until we reach a “good enough” state and then shift focus to the next “best pain”. Our north star is CD and test in production (we present that as separating “deployed” from “released”).

awesome. That sounds great