I think it's a mistake to think of Google+ as a Facebook replacement or even as an attempt to replace Facebook.
Google has a basic problem: they make their money from search ads. Mostly from search ads when people are looking to buy something. So a search like "best 2009 used cars" represents a ton of money for them. A Facebook posting "hey guys, can anyone recommend a good 2009 used car?" is a disaster for Google. Not only is it probably invisible to their indexing engine, it connects people to information without search intermediation that you can attach ads to.
So Google has two problems: 1) people are moving out of its search-driven search space into one-to-many questions to selected groups of people 2) Google doesn't have a lot of rich information about the social context of people who are searching. Things Google might want to know: how connected is the searcher, how influential, how wealthy, what interests does she have, what content does she share...
Some people stay logged in to Google (GMail) all day, but this trackable identity can be lost if e.g. the user is reading GMail using iPad mail app, or if like me they only login to GMail web for a few minutes, check email, and then log out.
Google would much rather have a single, perpetual, always-logged-in identity. A social network with a notification bar encourages people to stay logged in, lest they miss a notifcation.
Google is a data-driven organisation. So let us imagine that through Google+ "all" Google gets is a complete map of the social graph of all the top tech influencers in the world, along with the kinds of things they like to share, and their interests (their "sparks" in Google+ terminology). And along with that, Google gets social graphs for a few million other people. Even if G+ then only limps along, Google now has much better data to analyse for its core search business.
So I think Google+ is mainly about providing Google with enormous amounts of data that it can analyse to determine social signals for search, to understand Q&A social behaviour, to find out how people are grouped and interconnected, and to have data to drive social driven re-ranking and display. But most importantly by far, it gives Google the beginnings of data to optimise ads for the social graph. Would you rather pay 1 cent to display your ad to someone who has zero tech influence, or $1000 to get your ad in front of the eyeballs of a tech influencer whose posts are reposted and retweeted thousands of times? Would you rather pay 1 cent to display an ad to random people based on search keywords, or $100 to display the ad to a "circle" of people who have demonstrated a sustained interconnected interest in your particular topic?
This post inspired by GigaOm: Why Google+ won’t hurt Facebook, but Skype will hate it found via @nicholemcgill
July 1, 2011 Google Plus