Groupon. I consulted briefly for them when they were ThePoint.com and I thought their original idea for ThePoint was a really bad one and they were going nowhere. Within a few months they had completely changed their business model and were profitable. Then they became massively profitable and Google wanted to buy them for $4 billion.
What I love about Groupon is that in the world of Google they've proven there is still a huge amount of revenue and profit available online.
They are also showing that revenue still counts for a hell of a lot and that pure growth without profit is not the only strategy and may even be a bad strategy.
I also love that they fly in the face of Google's attitude that software should do everything and the main role of employees is to write software. Groupon is a people company in the sense that people in their business deal directly with their customers (the businesses they market).
Groupon are also targeting a hugely unexplored space which is small to medium local businesses that target local customers. They're doing it with boots on the ground (or at least via a person to person phone call) in each city the launch in. So far "local" has been a cluster of servers in a data center and a group of developers. It feels like Groupon is doing local the way it should be done.
I also love that the Google/Groupon deal didn't work out because I think Google's culture and approach would have killed a wonderful business that we all continue to learn from.