Start something with reasonable pace and massive scale
2above.com
2above.com
However, the ideas stand: To have a large success, you need to a) be doing something, b) avoid rushing things _too much_ [1], and c) pay attention to how small or insulated your market is. (You can't be "the next Facebook" by writing a landscape company search, even if it's the best ever.)
[1] This is tough to get right -- you need to launch before you're satisfied, but take longer than you really want to. The optimization of time spent really depends on the situation. For instance, if you're rehashing an old idea and selling based on the UI, it's worth going over with members of your target market before you release, to make sure what they _use_ is easy for them to find. But if your idea is new, you need to release ASAP to get market validation while you can still pivot if you're a bit off the mark.