I kayak on the waters around Manhattan--which is a ton of fun but can also be a bit precarious when it comes to navigating ferry traffic.
A few weeks ago, we were with a group of paddlers and came face to face with a ferry boat heading in our direction. We wound up between it and the pier that it wanted to turn into coming around a bend. The ferry boats in NYC are on pontoon-like structures, and so the worst thing you can ever see as a kayaker is straight through to the other side of the boat. That means you are dead set in its forward path. On the street, sometimes you wind up doing that little back and forth shuffle with someone walking in your direction and you bump into each other. No biggie. On the water, if you're a kayak, you can't afford to bump into a ferry.
So how do you signal what direction you're going to? Point with your oar? Some kind of lighting system? No, you just start turning... and fast! Since you can turn quicker than the bigger boat can, by committing to a direction, you force the other boat to move the other way to avoid you. Instead of trying to guess, they clearly know where you're going and where they need to go.
The same maneuver works in product development. The worst situation you can ever get caught in is when your users just feel so-so about your product, and you're kind of left guessing what the next feature or overall direction should be. That's why a lot of people advocate dramatic product moves early on in a product lifespan to get good feedback. If you start out with no social features whatsoever, you'll hear early on whether or not people want that. When you suddenly make things public, you'll either see site usage go up, or see your fans completely throw up all over it. Actually, either is a good outcome, because like in the kayak example, big moves make it clear what direction you need to go in to avoid disaster.