Without further ado, here is part 2 of my interview with Walker. To start reading from the beginning please see A Conversation with Priceline.com Founder Jay Walker.
WALKER: Let’s switch to the other side before we go to the theory. On the other side are users of patented technology, most of whom don’t know which patents they are using. They have no way to run the kind of sophisticated outlook to say, well, if I’m using patented technology how do I know what it is? I can’t read claim lines, which takes a federal judge to interpret whether I’m actually am infringing or not. It takes a whole Markman Hearing to figure that out. And on top of that when I try to look through the patents that are already issued as a way to learn they’re not written up in a user-friendly language, and I’m often advised by counsels not to do that.
The other side of this equation says look, if I could license the entire U.S. patent database for one dollar is there any doubt that everybody would license it? No. Everybody would pay a dollar and license it. So the beauty here is not whether or not companies need to get to use the U.S. patent database, they do. It’s a question of price and tools. I need the right price. One that’s good enough that I’m willing to pay not too high a price where I’m saying, you know what, I don’t think I want to pay. And then you’ve got to give me tools to access the 50 million claims. It’s probably well north of another 50 million pages of spec. So it’s not good enough to just give me a cheap license or licenses. Or is it good enough just to give me tools to be able to understand how the teachings in this database could improve my product, services, and competitiveness. You have to do both. The Utility exists to do both. And that’s the underlying business theory of why people who today are not paying to license patents will. People who own patents who would prefer to license them for a large amount of money will license them per unit for a much smaller amount of money so long as they can control who they license with and they know they’re getting the vast majority of the money being collected on their behalf.
QUINN: Now I think I understand your system in order to attract both the people who have the patents and the people who need to license the patents. And it strikes me as what you are contemplating is different than any of the other market based licensing approaches that we’ve seen over the last few years. Because, as you know, you’re not the first folks to have tried to do this. At least as far as I am aware, it hasn’t been successful yet. So I remember when we first started talking I was a little skeptical. And then as I started learning more and I was like, you know, this seems to work. Because as I understand it the patent owner still has the right to engage in other licenses with other people for other purposes.
WALKER: Absolutely. But let me go backwards before we lose it. We are the first. Let me explain what I mean by that. Let’s assume there were jet airplanes on the market. And all the airplanes in the marketplace were expensive jets. And let’s assume for a minute that the only way you could get from point A to point B using an airplane was with a private expensive jet. And all kinds of services like Net Jets come around and Rent-a-Jet, and time share jets, and this jet, and that private jet service. And everybody was working on the private jet market. And there was a million people trying to come up with ways that you could use a private jet in your business. And somebody comes along and says, “You know what, that’s all a great industry. That jet thing is wonderful. I’ve got an idea called Southwest Airlines.” It’s a whole new idea. I’m actually gonna serve people who would never even think about a private jet. That’s just not what they were gonna think about. Like there are people who rent cars or they’re like taking boats or trains. We’re going to literally serve a market that nobody else is trying to serve. The vast unserved market. We would be the first people to try to do that. Because you would look at all the other players and say, wow, there’s a lot of creative companies trying to serve jets, by the way United Airlines has announced its own jet service, and so has—there’s all kinds. And by the way, somebody’s created a first class airline now and it’s all filled with first class customers. But nobody is down at the base of the pyramid here. Nobody is talking about how to create hundreds of thousands, or millions of licenses a year over time. Nobody’s anywhere near that .
So the fact that there are all these other companies who are trying to solve problems related to licensing and patents is true and interesting. But none of them are trying to do what we’re trying to do. Which is not to criticize them. I am not criticizing them. They are serving a completely different market. They’re serving a market with big players and big resources. Or they’re serving a market where people who know exactly what they want and have to buy the patent, or license but they know exactly what I want. If you go down every one of the alternatives that have appeared so far you will find that none of them are doing what we’re trying to do. And that’s again not a criticism, it’s just that they’ve all gone where the money is. And we’ve gone where we think the money is going to be. And they’re two different worlds.
QUINN: Yes. And they’re also different approaches. Because at least one of the ones that I’m familiar with, what it seems like they’ve tried to do is collect all of the patent rights for a particular technology and then try and sell licenses and say, okay, we’ve got all development rights you will need.
QUINN: So if you license this portfolio you could safely do what you want to do.
WALKER: If you know what you want to do. You have to know what you want to do.
QUINN: Well, that’s true, you have to know what you want to do. And that’s the way innovation works. You don’t know what you’re going come up with when you start down the path because innovation takes many twists and turns as it unfolds, so it is really impossible to know what you will wind up needing.
WALKER: Yes, I have no idea what I need. I run a tool and die manufacturing in Indianapolis, I don’t know what I need. I mean do I need this, do I need that?
QUINN: In my mind there is also a much larger, practical problem as well. I just don’t know how you can collect all of the relevant rights to any innovation. There’s more than 5,000 patents a week that issue, and improvements and incremental advances occur. Gaps in technologies get filled. So how can anyone say that they have collected all the rights that will be necessary?
WALKER: Those are wonderful businesses. Personally I’m not excited about them as an investor, but I hope they all succeed. We don’t have to succeed at the expense of anybody else. Literally. But they’re not in our world in any way, shape, or form and looking at them teaches you nothing about us. No matter how hard you look at them.
QUINN: From what I’ve seen about the various models I believe that is certainly true. The only looking at those other models would do is teach people that you are different and that your model seems to have a real possibility of succeeding in a large way.
WALKER: You know, all it would teach you is that we are trying to solve a different problem. So we’re trying to solve a much different problem. As long as you understand we’re trying to solve a different problem you can then make a decision about whether or not we have a thoughtful and likely solution to the problem.
Now we have some big advantages in trying to solve this problem. Number one, we are business system inventors and this is the kind of problem we typically solve. Number two, we’ve solved other big problems. Number three, this problem actually is one I have lived. I own hundreds of inventions that I’ve been unable to generate a dollar of licensing revenue from so I am a living user of this problem. And I’ve had the luxury of being able to fund a world class team of people using technologies that have only just appeared on the horizon specifically semantic search and big data analysis tools that in the past were extraordinarily expensive are now only expensive enough so that we can afford to buy them and build them and then each user can pay a small portion to access the system .
So with new tools, our timing is right. We are the right company with the right people, with the right experiences, with the right capital structure. Literally you bring those thing together and you could say, all right, these guys could actually solve the problem of how to create an ASCAP for patents that is neutral and therefore balances the needs of inventors and users without ever using the court system.
QUINN: Right. That’s exactly right. That’s the only way the model works.
WALKER: Right. It’s the only way. It’s huge. Without a huge volume of low priced licenses the licensees would rather infringe and take their changes. And the licensors are forced to go to court in a way that they can’t afford.
QUINN: Right. So now going back to the semantic piece of the puzzle. I understand why that’s important, but to educate people who are not familiar with how your system works, why is the semantic piece so critical?
WALKER: Well, to understand semantic search you have to understand the different problems first. And the problem you have to understand is you don’t infringe a patent, you infringe a claim. Right? Let’s say there’s 2 million active—2.1 million active patents, let’s use an average of 25 claims a patent. That’s 50 million claims. So the question somebody has to ask is does the sub component of my current product infringe theoretically or possibly on any of 50 million claims? That’s the problem. Now, one way to solve that problem is to try to take the exact words that is in your product specifications and apply it to the exact words in 50 million claims. But you and I both know that’s a recipe for nothing.
ELLENTHAL: Like the keyword search used by Google.
WALKER: Right. Claims are written by claim construction specialists who of course are trying to claim the broadest possible conceptual space using the most broad words that the spec will support . So if you’re doing a key word search you’re gonna learn almost nothing. Or put it this way, you’re going to leave out 98+% of the possible infringement arguments that an attorney might make reading a claim prior to a Markman Hearing when looking at your product as best as they understand it. What semantic search does is it takes the key words and phrases that are typically used in your sales literature, in your technical literature, in articles about you that you have uploaded or we have found and abstracts them into a conceptual framework much like the famous Watson does when it plays Jeopardy. Watson doesn’t do a key word search it does a semantic search on puns and innuendo and obscure references, and double entendres because only a semantic search could decode a Jeopardy answer to deduce a question. Now Watson is the famous example of semantics but there are many people doing exactly what Watson’s famous capabilities are doing.
ELLENTHAL: It is effectively an advanced form of search that goes deeper than key words matching to get at concept and context. And advances in data science and computer science is allowing us to process giant data sets and tease out useful commercial relationships within the data. And that’s what our semantic search system does.
WALKER: And then you go one step further which says we are not trying to determine infringement, which is a legal conclusion.
ELLENTHAL: Right. And to me this is really fundamental to our approach, in contrast to what else is out there in the marketplace.
WALKER: Right. So we’re running this semantic search but we’re not running the semantic search to find infringement, which, by the way, would require a human brain and a Markman Hearing. We’re running semantic search to assign a probability score that a claim and a practice have in common. And once you would find a probability score you can say look it’s not clear whether the ‘809 patent as implemented, you know, as issued specifically relates to your method of coding a ball bearing here, but there is enough statistical overlap that it’s not worth arguing about. Let’s figure out how to price that uncertainty within a package of licenses which is efficient so that we don’t ever have to argue about infringement as a binary state but we can simply look at an accumulation of probabilities across a wide variety of relevant patents so your semantic search in the middle of a process between statistical probability and actual practice. Now you see the whole picture.
QUINN: Yes. Then you package a bundle licenses to the patents and the portfolio that statistically relate to what it seems based on this search, this semantic search?
WALKER: Yes. And for a thousand bucks a month for a hundred licenses. So if I’m giving you licenses for a thousand bucks a month and you’re paying $10 – $15 a month per patent for a license, are you better off taking a license or are you better off basically saying I’m gonna gamble I’m not gonna be sued, I’m not even gonna read any of these patents that might improve my products and services, I’m gonna be an intentional bad actor in the system for a thousand bucks a month. And if your answer is you’re gonna be that potential bad actor in the system congratulations, there’s nothing we are gonna do for you. But I think over time you will realize that’s a foolish decision on your part. As a business you’re better off being able to look at the solutions of everybody in the marketplace in a free and clear way than you are to simply say you know what for $1,000 a month I’m opting out.
QUINN: Right. I agree. Now this also requires you to get enough licenses so that it is attractive. So it’s cyclical. So what’s your pitch to those people who you’re trying to say, well, why don’t you give us some of your patents or a piece of your portfolio and allow us to add it to the utility?
WALKER: I don’t use any of the words you would use. I don’t say give us anything. I would say why don’t you make me the nonexclusive agent, for some of the patents that you feel you would like to get revenue for? Any patents you don’t want to get revenue for or any patents you’d like to litigate you keep those. You basically give us no rights. You’re just literally entering into a nonexclusive arrangement with us that you can revoke on a moment’s notice.
ELLENTHAL: On top of that you can give us a list of exclusions even when you list your patent in our catalog and say, listen, don’t license the following the companies, or don’t license any companies in a particular industry or a particular size. So you are in complete control as the patent owner of where your patent may end up being licensed through our system.