I'd wager on "beginning of the end". Shake up large projects until only the smaller, "quick wins" remain, then sell off the assets. It's easier that way from what I hear.