Microsoft's Intellectual Property Group is building a financial model designed to value and predict prices for technology patents, allowing the company to better forecast and budget for intellectual property-related costs -- all inspired by a best-selling book about baseball.
Infringing a patent owned by another company can be extremely costly, as Microsoft discovered in 2007 when a jury ruled that Microsoft infringed on MP3-related patents owned by Alcatel-Lucent and awarded US$1.5 billion in damages. That ruling was thrown out on appeal and the two companies later reached a settlement that resolved all patent litigation between them. But the case underscores the significant expense and high level of uncertainty, and therefore risk, that technology companies face when fighting over intellectual property.
When it comes to mitigating this risk, companies have several tools at their disposal. They can develop their own technology and file a patent to protect it, they can buy an existing patent, or they can license a patent from a third party. The pricing model is designed to help Microsoft managers determine which one of these options is the most effective choice.
"If you're going to do that, you need to reduce those options to economic units that you can compare with each other, so you can do an apples-to-apples comparison," said Horacio Gutierrez, the corporate vice president of Microsoft's Intellectual Property Group, during an interview in Singapore.
"This is where we get really funky and we have all of our finance people totally excited that we're looking at the business of IP in a way that can be analyzed in the context of a financial structure," he said.
For Gutierrez, who formerly served as Microsoft's associate general counsel in Europe and once worked in corporate finance at an investment bank, the inspiration to build a pricing model for patents sprang from an unlikely source.
"I got this idea from reading 'Moneyball,'" he said, referring to a book about baseball written by Michael Lewis, author of "Liar's Poker."
Gutierrez was inspired by the book's description of the approach used by Oakland A's General Manager Billy Beane to choose players for his team, while still operating within a limited budget.
"While everybody was going through the discussion of talent selection the way they traditionally do it, which was with the old scouts who've been around the business for 50 years or so, saying 'This kid's got a good arm,' 'He looks like a ballplayer,' and 'He's got a cannon,' and all of these qualitative things, they started to have people come in and start to crunch numbers and realized that a lot of the things they were measuring, and taking as measures of success, were the wrong things," Gutierrez said
"For example, does it really matter how many home runs a player hit in one year? Or how many runs the person actually helps score? If you just focus on home runs you could have someone who's a disruption to the productivity of a team when he's just focused on hitting home runs. Yet, everybody evaluates talent and pays for it based on how many home runs they hit," he said.
Beane's novel approach to picking the right players for his team transformed the Oakland A's.
"They came up with a whole different framework and started to come in and acquire a number of players that were under-appreciated by everybody else, but based on the metrics that, according to them, mattered, they were incredibly valuable. So, for many years the Oakland A's have been able to have a fraction of the budget of the Yankees and yet consistently make it to the playoffs, which the Yankees can't do," Gutierrez said.
(Alas, that will not be the case this year -- as of this writing, the A's are an insurmountable 18 games behind the first place Los Angeles Angels, one of the best teams in baseball this season, in the American League West division.)
When Gutierrez took up his current position three years ago, there was no framework in place for Microsoft executives to value patents. "I kept hearing people say, 'Oh, you don't understand, it's very complicated,' and, 'No, we can't reduce it to numbers, but this is the way things are done here,'" he said.
"They were saying 'Trust me, the patents that we get are very high quality." All right, I think I trust you with that, but what would other people think about it? Have we looked at how frequently our patents are cited by other patents? How frequently are they cited by scientific papers? These are objective measures that while they may not necessarily give you a quality score, they are indicators of quality. They are being influential for people who are working in the field," Gutierrez said.
"Then it turns out there's this IEEE thing that takes into account, among other things, the number of times your patent gets cited. It takes into account how many other patents you cite, which means you've actually looked deeply into it, you've done your research, you've looked at the prior art. You didn't just throw it out there and see if it stuck on the wall," he said.
Drawing on metrics like these allows Microsoft to define the quality of a specific patent or portfolio, and give managers a degree of confidence that the quality ranking is meaningful. They can apply the same framework when Microsoft needs to make a decision about whether to acquire a patent portfolio from a third party.
"What's come out of that is we now have a taxonomy for the quality of portfolios, so when you're looking at the patent landscape in a certain field, say speech recognition, you're able to see what are the really influential patents, what are the really seminal patents and then when you're able to go and analyze the secondary market for patents, you can extrapolate some trends in terms of pricing, and you can say generally a portfolio like this should be valued at X, depending on which tier of the quality ranking you placed them at," Gutierrez said.
That price doesn't reflect the inherent value of the technology described in the patent. For low-quality patents, the price reflects the cost of defending against the patent in court. High-quality patents are different, with prices reflecting the cost of not owning it.
"The tricky thing with patents is that the value of the patent depends not so much on the patent itself but the revenue that is compromised for the company against whom the patent is asserted," Gutierrez said.
Microsoft's pricing model, which is still in the prototype stage, isn't perfect. The market for patents isn't transparent and prices for many patents that are sold or licensed are never revealed, but are kept secret by the companies involved in the transaction. But Microsoft has access to enough data, drawn from its own history of patent deals and information about acquisitions made by patent aggregation funds, to build a model that is useful.
"We acquire patents also routinely in the context of potential conflicts people may be asserting against us. We look at the patents and we tell them, 'This guy's actually got something, we better take care of it,' and then you negotiate," Gutierrez said. "Although [the model is] not perfect and it's not comprehensive, you have enough to have a sense of what the ballpark [price] is."
But there's still lots of work that needs to be done to refine the model.
"We can essentially tell you how many conflicts you're going to face, and how much it's generally going to cost, so people can budget and deal with those issues because there's been enough repetition where you can extrapolate and predict. We're not all the way there. We're only getting started," Gutierrez said.
"And when I say, we're not there, nobody is. I actually think we're ahead of the game, by most accounts," he said.