For example, there’s a taxi company in Tokyo, where the manager of the company looked around and said, “My business is growing. People love me. But they hate the dispatch center because they've got to call the dispatch center to find out where a taxi is. The dispatcher looks on a computer screen because all my taxis have GPS systems. He can find out where the nearest taxi is. He gets the directions from the customer; gives the directions to the taxi driver. If they don’t give them perfectly, the taxi driver wanders around, calls back to the dispatch center, ‘Now, were they front door, or back door, or elevator, or parking lot?’ And they've got to reestablish that conversation. It’s a bottleneck. Very often you’ve got to hang up on the customer and call them back, once you find a taxi. Everyone involved hates this thing.” “Now, my taxis already have GPS systems. My drivers already have cell phones. Every consumer in Tokyo, basically, has a Web-enabled cell phone, also. And I kind of know where those customers are. NTT (Nippon Telephone and Telegraph Corporation), the telecom company in Japan, can tell me roughly where those customers are. Why don’t I put my customers directly in touch with my taxi drivers? Why don’t I show them where the nearest taxis at that point in time are on their little Web-enabled phone screens, and bypass that dispatch center all together?” And that’s exactly what he does. And so he has eased the most severe bottleneck in his organization. This is a classic network technology. All it does is open up opportunities. It turns out consumers love it. The drivers actually love it, once they learn to stop answering the phone, “Yeah?” It doesn’t work in Japan, right? So once the drivers learned some new behaviors, they loved it as well. And this continues to accelerate the growth of the business. So one option we have, and the least common one, is to develop technologies internally, ourselves.