As Google's product offerings evolve, the cost/benefit of providing customer service scales too. Providing personal support to customers who use search isn't scalable because Google makes fractions of a penny per interaction from these customers. Customer support for things like Google Fiber, Google Wallet, and Adwords is a whole different ballgame in terms of the amount of money Google makes on the average user so it is likely that there's enough margin there to actual scale customer support along with the growth of the userbase.
Anything of real value doesn't scale. Just look at Google Maps a lot of companies tried to make their own Map apps and failed. Why? Because google invested a ton of money into doing things that computers can't a present and require manual work this can't be replaced by any other map app without doing the same thing.
There's real value in recognizing what computers can't do and investing into doing it anyway and it can make your app irreplaceable.
Not to mention, scanning many thousands of books manually.
I thought the same thing till I saw this
http://hackaday.com/2012/11/16/google-books-team-open-source...
If we look at some of Google's biggest successes, it has been things that don't scale:
- The Ground Truth team in India, manually drawing the world's maps.
- Manual spam detection in search results
- Customer Support for most adwords accounts
- Customer Support for Nexus devices
I think the executives at Google have enough evidence that "do things that don't scale" can have a big payoff and it is really not that bad. They are clearly no longer too chicken shit to tackle problems that require customer support. The old Google only did things that match the business model of search, today's Google does what it takes to win.
I wouldn't bet against their ability to provide customer support. In the last 2 years, Adwords support went from non-existant to impeccable.