Our system of patent rights was intended to foster innovation. In practice, it’s often abused by companies, especially in high tech, as little more than a stick to harry and menace competitors. “There are industries where patents are used fairly to protect intellectual property,” entrepreneur Mark Cuban has written. “The technology industry is not one of them.”
Firms that are sued for infringement may scramble to build up defensive portfolios — acquiring other firms solely for their patents, in order to countersue. The result is a legal arms race from which, all too often, only the lawyers emerge victorious. Worst of all, litigation discourages would-be startups from introducing new products, which can mean less innovation in the end.
That’s bad for society and for entrepreneurs. But it’s also bad for the tech giants who rely on independent developers to bring their platforms to life. Without the two million titles in the App Store, for instance, Apple could not charge as much for its iPhone. In the sardonic lingo of practitioners, patent wars have a chilling effect on these ecosystems by spreading “FUD” — fear, uncertainty, and doubt.
However, big firms can do something about it, says McCombs Assistant Professor Wen Wen: They can give away their patent rights. It sounds counterintuitive, but that’s exactly what IBM did after being sued in 2003 for allegedly slipping some Unix source code into the open-source Linux operating system.
“IBM had the means to defend itself,” Wen says. “But the executives knew that wasn’t enough. Its business in hardware and services was based on Linux, and they worried that the legal risk would cause software makers to shy away from Linux.” IBM couldn’t sell servers if there were no applications to run on them.
To prevent that from happening, the company formed a Patent Commons in 2005, making its own intellectual property available to open-source developers. It also set up a legal defense fund for any firms or customers that got sued and vowed not to assert its own patent rights in the area.
The move was an about-face for IBM and a public relations coup. But was it more than that? Was sharing its patents also smart competitive strategy? Surprisingly, Wen says, no one had ever tried to measure the effects of IBM’s actions on the Linux platform. So she, along with Marco Ceccagnoli and Chris Forman of the Georgia Institute of Technology, set out to do just that.
What they found, in a new paper published in Management Science, is that IBM’s contribution of patents led to significantly higher market entry by startups with new open-source software products. “IBM didn’t do this out of altruism,” Wen says. “It did it to strengthen its own core businesses.”
A Clash of Titans
Wen says she became interested in this question back in 2010 as she watched a furious patent war break out in the smartphone industry. At the time, Google’s Android operating system was beginning to take market share from the iPhone, and Apple’s CEO threatened to “go thermonuclear” on the upstart. “I will spend my last dying breath if I need to,” an apoplectic Steve Jobs vowed. “I’m going to destroy Android.”
But Apple didn’t sue Google. Instead, it went after companies like HTC and Samsung that made the handsets Android ran on. Evidently, the idea was to scare off Google’s hardware partners. Jobs knew that if he could paralyze the ecosystem, Android would collapse.
In the end he failed, in part because, in 2011, Google acquired Motorola. Google CEO Larry Page was explicit about the motive for the $12.5 billion purchase: Motorola’s 17,000 patents in mobile communications, he said, “will enable us to better protect Android from anti-competitive threats.”
The Problem with Software
The experiences of IBM and Google are not isolated cases. Over the past 25 years, the number of patents issued in the U.S. has exploded — from around 100,000 a year to 325,000 in 2015 — without any evident economic benefit. Productivity growth has sharply declined, suggesting that all those claims are serving more as a barrier to entry than a stimulus to innovation. In fact, most startups don’t apply for patents, due to the cost.
A big part of the increase has come from the software industry, Wen says. “Software wasn’t really patentable before, but the courts changed that around 1996, and they extended it to cover even abstract ideas and algorithms in 1998. After that you see a huge increase in litigation.” A GAO report found that nearly 75 percent of defendants in patent lawsuits are software companies.
Why is this field different? One issue is that innovation in IT products, and especially software, is highly cumulative; that is, new ideas build heavily on prior advances. That results in dense patent thickets, which are hard for entrepreneurs to navigate — especially since the claims in software patents can be vague, and the boundaries on protected IP are often ill-defined and overlapping.
“Patent thickets raise market entry costs,” Wen says. “You have to do all the legal research. Then you have to invent around them — which is often like reinventing the wheel — or else get licenses from all these different rights holders. Just having to negotiate all those deals is a big burden, and then you have to pay royalties.” And after all that, you still risk being ambushed later by a troll looking for unearned revenue.
So while new software can be cheap to create — in the sense that a few coders in a loft can build the next big thing — the cost of commercializing it may be fatal. “It’s not just a lack of money and legal staff,” Wen says. “Because startups don’t have their own arsenal of patents, they can’t deter lawsuits the way a large firm can, by threatening to countersue or by negotiating cross-licensing agreements.”
What Goes Around …
That’s where IBM’s idea of sharing its intellectual property comes in. By clearing a path through patent thickets, a big firm can make it easier for others to bring new products to market that complement its own.
To find out if IBM’s gambit worked, the researchers looked at innovation by startups in 33 market segments related to the Linux operating system — everything from database tools and graphics software to email, spreadsheets, and word processors. In each market, they counted the number of patents IBM contributed to the Commons, weighted by the number of claims in those patents.
The effect was clear: The larger the Patent Commons in a segment, the more new open-source products were introduced. And the benefit was especially large in markets where innovation was highly cumulative—that is, where there were more overlapping claims and thus higher infringement risk: A 10 percent increase in pledged patents led to a 5.5 percent rise in new products.
Besides lowering barriers to entry, this strategy also benefits innovators in indirect ways, Wen says. “If IBM or Google stands up and says, ‘We’re not going to sue anybody,’ it puts pressure on other firms.” In the Linux case, Red Hat, a major distributor of Linux, soon followed suit. And once it’s clear that incumbents want to build a platform instead of trying to profit directly from their IP, well, there’s less FUD.
For IBM, the strategy of opening up its intellectual property proved successful: It resulted in more new software products, thus strengthening the platform on which the company had staked its fortunes. While the findings of this study are specific to software markets, Wen says the approach is now being adopted in many other industries that may have platform aspects, such as microprocessors and electric cars.
Even in business, apparently, sometimes it really is better to give it away.
Opening Up Intellectual Property Strategy: Implications for Open Source Software Entry by Start-up Firms is published in Management Science.