Defining a MVP by Setting OKRs

This is definitely better than not using any methodology, but the problem I’ve encountered with this approach is that OKRs are only as good as your current knowledge.

So you’re working towards a goal that’s as best as what you currently know. Of course, every iteration you do will erode or refine that understanding. This means your OKRs are a depreciating asset.

Because of this, I recommend focusing on very narrow OKRs: those only focused on the current experiment. The one constant I see is that you simply need to ship increments as fast as reasonable (and figure out how to do this better), rather than focusing on theoretical goals.