Sync
Coming as someone who has worked with Ozzie's products over the years (both Notes and Groove)... he does have some great ideas but planting the seed vs. executing on the idea are two different issues. Also, in general, syncing/replicating data (whatever you want to call it) is actually a lot more difficult down at the nuts and bolts levels than most people realize. It is, however, the holy grail in a world where cloud connectivity, 100% of the time, no matter where you go for all three screens is not a reality. I'm not exactly surprised that Ray and Microsoft couldn't get it working - but that's no reason to throw in the towel IMO.
So what exactly is the job of a CSA... to plant the seed and get it half or mostly baked (which is what he did for Azure, but apparently not for Mesh) or is it to take it all the way into the market under his personal fiefdom (i.e. making his own new division inside Microsoft)? To me it would seem that having Ray, or someone around, to run the incubator... to span the divisions is absolutely crucial. Regardless of his apparent/alleged failures, losing him and not replacing him puts Microsoft squarely in the Nokia model - which is essentially a bunch of small independent companies inside one corporate umbrella. For a company that essentially does one thing (software) that is really... really not good.
/Fail, because not replacing the Ozz is a big one