If any other tech company had hauled in $948 million in IP income last year, the number would represent an amazing feat. But IBM Corporation isn't just any tech company. In fact, this is the first time in eight years that Big Blue
didn't clear the $1 billion mark. The decline continued into 2006. During the first half of the year, IP income at the Armonk, New York-based company dropped by 17.6 percent. Meanwhile, IBM continues to receive more U.S. patents than any other company on the planet (2,941 in 2005 alone). The company's overall portfolio is stuffed with a staggering 40,000 U.S. patents.
This isn't a case of corporate schizophrenia. The decline is part of Big Blue's big shift from software and hardware manufacturing to the software services industry. While competitors like Microsoft Corp. are trying to mimic the old IBM licensing machine--even bringing in former IBM licensing guru Marshall Phelps in 2003--IBM is moving toward an open source model. And where Big Blue goes, others may follow. Since IBM is the Patent and Trademark Office's biggest customer, the company often winds up functioning as a de facto PTO consultant. This recent shift in internal IBM strategy could affect the entire patent world--and that is precisely IBM's plan.
In the 1990s, under former CEO Louis Gerstner and IP luminaries like Phelps and Emmett Murtha, IBM turned its extensive patent portfolio into a golden goose, bringing in a high of nearly $1.7 billion in IP income in 2000. But just as the money came pouring in, the architects of the famous patent licensing plan started heading out--Murtha, director of licensing and business development, retired in 1997, Phelps in 2000, and his successor Gerald Rosenthal retired in 2005. At the same time, IBM was changing its core business. The company had already encouraged its large customers to buy their mainframes rather than lease them, because it realized that servicing the machines was more profitable than manufacturing them. But relying on IBM's own proprietary server software proved a hindrance to the service business, so IBM moved into the open source world. In 1998 the company adopted the popular Apache HTTP server software for its mainframe computers. Later that year IBM started supporting Linux, the open source movement's answer to Microsoft Corp.'s Windows. The company's turn toward free software forced a shift in IP strategy. The new IP leadership would need to walk the line between sharing its technology and vigilantly protecting its IP.
Changing how the IP group works was mainly a series of small reorganizations and reshufflings among longtime IBM employees--with one key shift. In 2004 CEO Sam Palmisano, who had replaced Gerstner two years earlier, brought in a 26-year IBM veteran named John Kelly and charged him with shaking up IP strategy. As the former head of IBM's Technology Group, Kelly had helped form an alliance between several competitors, including Sony Corp. and Infineon Technologies AG, who collaborated on advanced semiconductor chips.
Kelly started by reorganizing the IP department under the new Technology and Intellectual Property umbrella. The old department had been structured as one unit, and had only a loose tie to IBM's business executives. Kelly, who would now function as the direct line to Palmisano, split the department into technology, business, and legal groups, staffed by a combination of lawyers, inventors, salespersons, licensing executives, and other businesspeople. He seeded the business unit with a few new positions, creating a standards and open source group [see below]. The new jobs all went to company veterans, with one exception: Kelly hired Kevin Rivette, author of the seminal 1999 book Rembrandts in the Attic, as vice president-IP strategy.
On the legal side, IBM's chief IP counsel, David Kappos, became vice president and assistant general counsel-IP law. The 45-year-old now heads one of the largest IP departments in corporate America, a global team of 360 lawyers, patent agents, and staff.
Kappos sits at the end of a long blue hallway in an office decorated by a dog calendar, family photos, and a nineteenth-century Japanese firefighter's uniform, a memento of his two years in Japan as IBM's assistant general counsel for Asia-Pacific. The rest of IBM's IP team leaders are situated nearby--except for Kelly, who, when he's not meeting with executives in Armonk or flying to clients on multiple continents, keeps an office in Somers, New York. Kappos says that Kelly's level of access to Palmisano and the business side has made it easier for his group "to get things done."
"If a call has to be made at the senior executive level, we can do that now," says Bob Sutor, head of the newly created standards and open source division. "You can't paralyze your standards or open source process because you can't find someone to make a decision."
The faster pace is critical to keeping up good relationships with the open source world, an important component in IBM's new business model. While open source programmers continually improve the software itself for free, IBM generates revenue by charging corporate users for training and support services, as well as through sales of new hardware infrastructure to use the software. A healthy back-and-forth with developers--and the occasional IP swap--can only help improve the product. This in turn boosts IBM's bottom line, and compensates for the loss in IP income. In 2005, while IP income was falling, IBM's gross profit went up to $36.5 billion, about a 2.7 percent increase.
Last year the company sent out a well-publicized signal that it was serious about working in this new world: In January 2005 IBM pledged 500 patents to the open source community. Opening some of IBM's patent portfolio to royalty-free use was an idea that the legal team had discussed for years. "But it was only when John [Kelly] came in and provided us with the backing that the business side would go for it," says Kappos. Once the pledge was approved, it was up to Kappos to manage the internal negotiation between the open source and licensing teams within the IP group. The goal was to hang on to the patents that were related to proprietary IBM software, but offer up ones that could facilitate research in the open source arena. For example, Kappos says, "we left out the patents that were specific to IBM's ViaVoice speech processing [software], but included those covering general speech recognition technologies that could be relevant to open source speech software, such as those for the Linux toolkits."
Decisions about what to license for a fee and what to make royalty-free are now common. "We still license patents, but we participate in the newer world, which is open source," says Sutor, a Bob Dylan-worshiping mathematics Ph.D., who fits in well with the free software community. "It's all about [balancing] what we get money for in the short term with licensing and what we decide to offer up free because there may be long-term growth," he says.
Open source considerations are just one way that Kelly has helped foster a new level of creativity in licensing deals. These days, customers without sophisticated portfolios will come to the IP team seeking help in building their assets, says Bill LaFontaine, who became vice president of business development and licensing this past February. IBM will license IP to these customers and even give them access to its own inventors to help customize the new technology.
The group is also pushing into life sciences, where IBM provides IT infrastructure and data management for biology research. And it's getting involved in the automotive industry, where IBM provides chips for the increasingly sophisticated computer systems inside cars and trucks. "These are areas the Marshall [Phelps] team never spent any time on," says LaFontaine. "We're putting IP where we haven't in the past," he says.
But licensing is only a fragment of what this group does. Kappos says that his team handles the IP aspects of IBM's M&A transactions, like when it sold its PC business to China's Lenovo in 2005. Kappos assigned several IP transaction experts to the deal full-time for months. "I personally spent hundreds of hours leading up to the closure of the deal," he says. The group also makes decisions about whether or not to join standard-setting bodies, Sutor explains.
Then there's IBM's role as patent law and policy guru. The IP group closely monitors changes in patent case law coming down from the U.S. Court of Appeals for the Federal Circuit, advising IBM's senior business leaders on IP matters that may affect the business. The information flows the other way, too. When Palmisano or other executives hear from partners or clients about IP problems affecting the global business community, such as so-called patent trolls, they bring them to Kelly.
The IP group works closely with IBM's government relations group in Washington, D.C., and has been particularly active in IBM's efforts to lobby Congress on the current patent reform legislation. But Kappos isn't optimistic about legislating changes to the system. "It goes nowhere fast," he says. IBM seems to put more faith in changing policy at the patent office. "Sometimes we have problems in patent prosecution that we tell the PTO they need to fix," says Manny Schecter, associate general counsel-intellectual property law. (The same goes for the Japanese and European patent offices, adds Kappos.) The patent office also comes to IBM for feedback on the patent application process and ways that the PTO can improve it.
This year IBM announced three patent quality initiatives, suggestions on improving the system without legislative change, each of which they discussed with the PTO beforehand. So far, only one of IBM's ideas has grabbed the PTO's interest. In August the patent office gave the green light to IBM for a test run of community patent review, an idea that the company developed with New York Law School professor Beth Noveck. The basic concept is to open up the patent application review process to the Web--where PTO outsiders could give input on applications for free, like a Wikipedia for patent applications. The patent office itself first put forth a community project for gathering prior art as early as 1966, but Noveck points out that until now the technology hasn't been available to support the idea.
Marc Ehrlich, IBM's counsel for patent portfolio management, says that IBM had run the concept by PTO officials several times, but it never took hold. Then, last year, he saw an article online at Wired News about Noveck and her idea for a "wiki" patent review process. "Her proposal, though more aggressive than ours, had similarities," Ehrlich says. With Noveck, IBM is setting up an online forum where anyone with expertise in a patent area can access the site and argue which prior art was most relevant. The top ten most relevant pieces, as determined by the site's users, would then be shipped to the examiner to help judge the application. With Noveck's help, IBM has held a series of workshops on community patent review around the country.
Ehrlich admits that the process could run into trouble. The problem is that the engineers most knowledgeable about a patent under review are also the most likely to be working on similar projects for other companies, he says. If a company's engineers spend time reviewing other patents, and then that company gets sued for infringing those patents, the wiki review would be a liability.
Companies like Microsoft and Hewlett-Packard Company have signed on to the community patent review project, but Noveck also hopes the wiki process will generate prior art from more unusual sources, like graduate students or technology hobbyists. "We want to find those people and have them find us," she says.
Greg Aharonian, who makes a living running prior art searches and regularly rants against IBM in his e-mail newsletter, the Internet Patent News Service, doubts that many of the patents that IBM puts up for review will generate interest from the hobbyists that Noveck envisions using the service. "Without getting good money for [prior art searches], it gets boring real quick," he says. "Who is going to want to waste their time on them?" Aharonian points out that several of IBM's most recent patents--such as 7,089,354 for a "disk fragmentation system," or 7,088,465 for "printing more or less of a Web page"--do not cite any nonpatent prior art. "If IBM really cared about patent quality, why don't they submit more prior art for their patents?" he asks. Aharonian says that IBM submits substantial bibliographies of prior art for the patents it values, but none for the majority it uses to bulk up its portfolio.
IBM has supported patent quality initiatives in the past, such as the Software Patent Institute, founded in 1992, but many have failed. The Patent Institute was defunct by 2000. But with its new IP strategy, IBM claims that things will be different. Associate GC Schecter says that the company is simply interested in a more transparent, definite patent approval process: "If that means there's some small number of our patent applications that might be impacted by that, we're willing to live with that--in part because we believe we have a very high-quality patent process."
Turning an idealist's project like Noveck's into a business initiative is not new for IBM. That's exactly what it did with open source. When IBM first adopted open source technology, the company was considered a patent-hoarding monster within the open source community, which essentially defined itself against IBM. Then the company slowly began infiltrating open source groups around the world. IBM staff in Sutor's standards and open source group keep tabs on open source activities, and help to influence their thinking. "In the open standards environment, we're active in so many standards [groups] that we tend to have IBM people at high levels," says Ehrlich, who works closely with Sutor.
There are now IBM staff members in more than 300 standards organizations and 150 open source groups, including the Internet Engineering Task Force and the Open Source Initiative. Sutor himself was recently named in a "Who's Who" of the open source and free software movement by British technology magazine Computer Business Review. Sutor is a perfect fit for the hippie world of open source and for the new IBM IP group as it tries to transform from licensing machine to open source guru and patent quality devotee. "It is very unlikely that I'm going to do something today that will bring in value tomorrow rather than one or even five years from now," says Sutor. "Standards and open source don't work on a quarterly basis--this is a long-term investment." But that is just the kind of investment that this new team of old IBM'ers knows about firsthand.
