News article on 20% time. In a broader sense, these things can just be thought of as R&D – something that most tech companies need to stay competitive and innovative.
(Speaking as a layman:) “This stuff” sometimes turns out to be wildly popular: Google Maps, Docs, Scholar, and others all started out as Labs projects (full list of Labs graduates) and are now used by untold gazillions of individuals and also academics and enterprises – some of who click on ads, others who pay license fees, all who become more and more of a captive audience, spending time on Google’s services instead of Microsoft’s or Yahoo’s or Altavista or Webcrawler or Netscape or AOL or any of a myriad other companies that Google’s world domination scheme helped obliterate.
Other Labs experiments (Google SMS’s location-aware search, Google 411’s voice recognition, Goggles, translation things, Froogle/Products/Shopper, Places) contributed to the Android platform (and to ChromeOS in the future), helping make Android become the world’s #1 mobile OS, with Google’s mobile advertising system piggybacking directly on it. Though free to end users, these services all help contribute to Google’s consumer mindshare, making it that much more attractive to Google’s ultimate customers, the advertisers.
Taken individually, some of these projects seem pointless, but I suspect at least SOME of them helped play a role in Google’s rise to godhood.
I’m not as familiar with Microsoft Research, but I’ve seen many cool projects come from them… which unfortunately rarely seem to get marketed very well after the research phase. I’d blame the rest of MS for their glacial product development/release pace; they still function like a 90s tech giant, too bureaucratic and slow to keep up with the others. It’s also harder to see their impacts because a lot of their research is aimed internally, whereas much of Google’s is meant for public consumption.