I really liked this post and it hit on a few problems I've encountered when trying to use OKRs: namely, what I'm going to call the "we know what we want to do, so just fucking do it" problem. I've been guilty of trying to contort things into an "OKR-like" view of the world when really, the answer was to just say "this quarter, we'll work on this task." It makes everything difficult when I'm then asked to turn this into a roadmap of things we'll do.

I do think it's worth remembering that John Doerr's implementation of OKRs at Intel actually did include binary goals. I remember reading the book and seeing "complete design and begin fabrication of chip X by Y date" (I'm paraphrasing because I don't feel like going through the book to find it). Yes, the higher-level objective was "establish Intel as the top chip maker" but that, by necessity, included "actually make the chip."