It would be nice if you could clarify that idea a bit. Because we're not seeing 5+ GHz clocks from Skylake yet, nor do we know how it performs in the <3 GHz range. Or am I missing something here?
You were asking what Skylake needs to be considered an unmitigated success, not what we clearly know about the chips.
Intel's main priority is shaving power usage in mobile SKUs, so we should expect them to be very aggressive on this with every occasion. Haswell brought low idle power usage, which is necessary but not sufficient. Broadwell brought some power savings under load, but only just under 3Ghz, where it starts losing steam. Skylake needs to bring home a clear victory, significantly less power usage for same clocks period. (mind you, mobile SKUs)
As for the high clocks, we already saw early leaks on Skylake OC that looked a lot more promising (and more plausible) than Broadwell's early leaks. The pessimist will look at desktop Skylake 95W TDP and say it must be using a lot more juice than HW, while the optimist will think there's more room for faster SKUs. This time around I tend to lean on the positive side.
If Intel gets their most (clearly) efficient CoreM and their fastest factory clocked i7 out of Skylake... achievement unlocked.
PS: and a pony please.