Tim O'Reilly: Close the Loop

So, but back on this notion of since we're on the subject of cars, another application that puts a bunch of these principles into work. Software above the level of a single device is Uber and rethinking workflows. Think about that wonderful experience. You've got a phone, it knows where you are. The driver's got a phone and knows where he is or she is. And then Uber has this memory, this brain, which is identifying that routing, telling everybody how to get together. That's a system. It's softer above the level of a single device, it's providing a whole lot of context without you having to do a whole lot when you first check it. Uber does a decent job of figuring out where you are. Maybe you have to correct it a little bit. They've done a pretty good job of applying these principles. But there's another principle that Uber brings up. Kind of a really interesting thing. When you use Uber, you're asked to rate the driver. The driver is also asked to rate you as a passenger. Now that's a really, really interesting thing. And Uber is pretty ruthless about weeding out people who get bad marks from their passengers. Now you think about your experience of how many people here have actually driven in an Uber. Fair number. Pretty good experience, isn't it? It's kind of one of those magical things like going into a square enabled coffee shop or going into an Apple Store. You go, whoa, this is really good. And you compare that to your experience with a cab driver. You have, the guy's playing loud music, he doesn't know where he's going. All kinds of terrible experiences. You get an Uber driver does that, they're out on their ear. And you think about the old school way that taxicabs are regulated. Well, we're going to test people, we're going to figure out who's approved to do this. Uber has actually done one better. There may be some regulatory issues there that need attention, but there's a lot of ways that we're starting to see reputation systems that we've seen on Internet applications start to have relevance in the real world. And I think that's, I'm not putting that as a principle, I'm just asking it as a question. I think it's one of the great, interesting unanswered questions in the government area. But my next lesson that I really wanted to highlight from Uber is to close the loop. You know, what makes that Uber experience so different is that you know when your cab is going to arrive. You know when you, let's say you want to go to the airport and you call a cab, you never know are they going to show up, are they close? You know, did they forget? You know, you're standing on a street corner in the rain, you don't know how long it's going to be. Think about Uber. You can sit in the restaurant, you can sit in the coffee shop, you get a text when the driver is outside. Uber has closed the loop. You want to know where they are, you can actually look on the map and you can watch their progress towards your location. Uber has closed the loop. And I got that framing from Chris Sacca, who's an early investor in Uber. He was also Google's head of special projects for a long time. And since what I learned from Google is to only invest in things that close the loop. And I think that's a really, really important principle. And if you're doing a startup, think about what loops you can change. That's another piece of how do you make things smarter so that you don't have these open ended systems where you don't really know what you're going to.