Jerry Kaplan: If you’ve ever been online and if you haven’t, I don’t know what you’re doing watching this video. You know that many websites are tracking and studying your behavior and in a way they help you by presenting products and information that they think that — they believe — based upon your browsing history and other characteristics are going to be of great interest to you. But there’s also a darker side to that activity. While that may add great convenience to you, the truth is that that also permits them to look at questions like what do they estimate you’re willing to pay for that product? Now a lot of people think mistakenly that you’re supposed to charge the same price for a product to everybody. That’s not the case. You can’t discriminate based on certain criteria — race, religion, sexual preference. But it’s perfectly fine for me to charge this guy more than that guy because I think he’ll pay more and just look at airplane tickets as a perfect example of that sort of thing. Now here’s the problem. We’re taking those kinds of decisions in these websites. Amazon itself is a fantastic example of this and we’re incorporating very sophisticated machine learning algorithms that are designed to manage the overall behavior of the group of people who are visiting that website.
In order to optimize profitability for the companies that are running those websites. And they will cut you the least slice of pie, the smallest slice of pie that they can to get you to send you to do what they want you to do in order to maximize the profits of the corporation. Now you may have been on Amazon and you may put things in — I use what’s called a save for later or something in your cart. And you come back the next day and good news, you know, this book is three cents less or that’s two cents more or this is a dollar more. But there aren’t people doing that. This is a machine learning algorithm. And what it’s doing is analyzing time of day and the characteristics of what you bought in the past and how you’ve responded to different kinds of incentives. And where you came from and what kind of browser you’re using as a major factor. Anything it can in order to adjust the price to just the point where you’re going to buy at the highest possible price. You as an individual have freedom of choice. It’s a free country. You can buy it. You can not buy it. That’s great. But we as a group as a set of customers purchasing from Amazon or some other site we adhere to certain statistical properties. So as a group, we don’t have that freedom because it can be managed by the entity on the other side. Whenever there’s an information asymmetry like that, they know what you’re likely to buy by what your characteristics are and they can optimize the yield on site based upon that. They’re at an advantage over you. Amazon is a wonderful company, but it is basically one giant machine learning algorithm. It is designed to do what’s called arbitrage. It knows what it can buy things for. It knows what it can sell things for. And it can adjust the profitability in that zone in order to maximize sales, in order to maximize profits.
And it can do so in a way that is far more efficient than has ever been possible in retailing before. So when I think of Amazon, the fact that they’re selling goods is incidental. I think of it like a stock-trading program. Buy low; sell high. Buy here; sell there. There’s a spread. These really are arbitrage systems and you are the mechanism by which these companies maximize their profits.