Huh? In what possible way can a marketing definition of "success" possibly make Windows irrelevant?
I'm pretty sure that, internally, there are a whole bunch of KPIs that strictly determine what will or won't constitute success not just for Windows as a whole, but for the various subsections of the ecosystem too. That's not the information you share with the world though, regardless of which way the outcome goes.
It can make Windows irrelevant if decisions (or the lack there of) are made off of that marketing definition. Too often, especially in large corporations, the dirty little details (KPIs) are often buried in order to save face (for the moment). It was pretty clear in the W8 blog that Sinofsky and crew were quite comfortable spinning and bending in order to justify what they wanted to do. If that trend continues we may be actually witnessing the downhill side of the Windows product lifespan (mind you a very long downhill side).
I frame this all in the same context of things like Ballmer discounting the iPhone originally. There's just no sense of urgency nor apparent desire to get things right. Sure you have to be confident in your public face but in private it should be a balls-to-the-wall effort to exceed what your competition has to offer and to get that effort out the door in a reasonable time frame. Microsoft has had flub after flub after flub in this regard and with W8 it doesn't look like that bothers them a whole lot.
I guess we'll just have to wait for the fat lady. If W8 turns out to be a dud too (especially in the devices category) maybe she'll be singing a tune for Ballmer's resignation soon.