When I was there (2007-2008), there was a popular phrase among the management class along the lines of "We like to measure things in the billions here." Meaning that if whatever project you're working on is going to last (e.g. get more funding, get more headcount, not get killed, etc...) it better have some extremely large and impressive internal metric (or have a credible growth plan to get there). Billions of users, billions of downloads, billions of dollars in revenue, etc. If something doesn't have that kind of metric, start the countdown timer...