I like the 10% thing, but the problem is that a lot of the legitimately innovative products would never be authorised.
The 10% thing, to me, only works if it allows people to escape the internal political machine. Let's say you have an existing product, 1 mil+ LoC and ten years+ old, and some young upstarts want to try and rewrite it from scratch, of course the management is going to outright say "no" because:
A) It would take an entire department many years to accomplish the same task.
B) It won't generate new revenue streams even if successful.
C) It competes with an existing product.
But that's exactly the type of thinking that causes very successful companies to ultimately have the rug pulled out from under them when external developers do the same exact thing, and are able to deliver a more modern alternative in fewer lines of code (because it didn't evolved, it was designed that way) and with less bugs (due to better use of libraries and pre-existing code bases).
I guess what I am getting at is: that term "let people" has a lot of baggage. Ultimately you might just see people using their 10% not to innovate but instead just to do more of the same-old, same-old because that's all management will allow.
I've always wondered if big companies shouldn't try to simulate a "startup." Imagine this: You take ten developers and one manager. The big company continues to pay their normal salary, and they have six months, their own space, and a modest budget to produce whatever they want (no questions asked). The thing they produce is then trialed and if it proves to be successful, there's a personal financial incentive to the eleven people involve (e.g. they get a % cut of the profits for a period).
Essentially do a startup, but take the risk out of it (no financial danger), and keep the incentive (since they personally financially gain). Do one "startup" a year using a different eleven (10 + 1) people. As a bonus you'll likely see a morale gain due to just changing up what people do day to day.
Something can also be said for turning it all into one giant competition. See Steve Jobs with the Lisa vs Mac "war." It might have ultimately got Steve fired, but it also kept both teams on their toes and may have worked if he hadn't have gone overboard with it (and the Lisa in general). When he returned he continuously started projects which ate away at other products Apple produced (e.g. iPhone Vs. iPod, iPad vs. Macbook, and so on).