The point, says Manny Schechter, an intellectual property lawyer for IBM Research, is that when it comes to patents, high-technology companies are, out of necessity, interdependent.
"Whereas pharmaceutical companies can get a monopoly--a single patent that controls a product set--software companies have to license each other," Schechter said. "If we didn't, the industry would grind to a halt."
Schechter spoke at the Nov. 30 Wharton Impact conference, "Managing Knowledge Assets: Changing Rules and Emerging Strategies," sponsored by the William and Phyllis Mack Center for Technological Innovation. Participants from industry and academia grappled with the knotty issues surrounding patents, intellectual property (IP), licensing and knowledge management.
In many industries, companies eager to capture gains from their innovations are filing patent applications at an unprecedented rate. At the same time, they are struggling to keep up with accelerating development cycles, shrinking lead times and changing legal interpretations.
The conference addressed issues such as: What strategies are companies using to protect their knowledge assets? How important are mechanisms such as secrecy, lead time and complementary assets (manufacturing, sales, service), and do they function as a substitute for, or supplement to, patent protection?
Mark Myers, senior fellow in the Emerging Technologies Management Research program at Wharton, recently retired from Xerox after a 36-year career in research and development.
"In the current period, having strategy around intellectual property, whatever your vision, is a key management concept," Myers said, describing himself as "decidedly pro-patent...Part of my bias comes from my career with a corporation that had a patent monopoly." Myers said he believes that great research laboratories such as the one he was part of (Xerox PARC) do not operate on the basis of secrecy, but on openness--and that intellectual property protection allows that openness to flourish.
In the eye of the beholder
Steve Andriole, a venture capitalist who was most recently CTO of Safeguard Scientifics, offered the perspective of a venture capitalist investing in software and information technology. "In a way, intellectual property and patents are in the eye of the beholder," he said, noting that while entrepreneurs looking to raise capital often ascribe great value to their patents and patents pending, venture capitalists are not that impressed.
"In the thousands of pitches I experienced personally, patents were considered one of the least important factors. Entrepreneurs would talk about the need to seize the marketplace within six months, but they didn't realize that this makes the patent process pretty much irrelevant," said Andriole, referring to the fact that patents take years to receive.
If patents are relatively unimportant, he was asked, what then could software companies do to prevent their value creation from being duplicated? "Not much," was Andriole's response. He emphasized that in software, the granting of a patent is a lag variable, whereas lead time and capturing market share during periods of exuberance are key.
Although VCs in high-tech might not set much store by patents, other industries differ, Myers pointed out. "In the biotech area, there is a lengthy open period in which companies have to expose themselves to regulatory processes," he said. "VC companies in that industry are in fact concerned about intellectual property, because companies need 'cover' during that period of time. And in optical communications and devices, which is another area I worked in, there's a perception of high importance surrounding intellectual property protection."
IBM's Schechter overviewed the dramatic change in his company's intellectual property strategy over the last couple of decades. In the 1970s--a mainframe environment in which IBM held the dominant position--the company, as it does today, licensed its software patents and other intellectual property to other companies.
But at that time, IBM granted licenses primarily to achieve what Schechter calls "freedom of action"--to get other companies to grant it licenses in return. Today, while IBM is still a giant, it has tens of thousands of competitors and now arranges licensing to make money. The company's royalty income from intellectual property licensing totals nearly $2 billion.
Schechter identified IBM as the holder of more U.S. patents--and more U.S. software patents--than any other American company. "Not by accident," he said. "We like the image it conveys of us as an innovative company. And we will license our software patents to virtually anybody, software business or not, on non-discriminatory grounds."
In fact, said Schechter, the leading U.S. holders of patents are all software companies that create licensing agreements with each other and balance out the relative value of their patent portfolios. Innovative technologies like DVD, PalmPilot, Handspring and TiVo have IBM patents that relate to them, he said.
IBM obtains its software licensing revenue in two different ways, he noted. In some cases, a company will approach IBM to set up a licensing agreement so it can use some of IBM's patents. "Those are the easy cases," he says. In many other cases, IBM must prove that another company infringed its patents to win licensing fees from that company for the continued use of the intellectual property involved. Schechter says that when a new technology surfaces in the marketplace, he and his team identify who the major players are and which IBM patents may be involved, then attempt to prove IBM's case.
Proving patent infringement might include using the following:
Product analysis: buying products and taking them apart. (However, copyright issues can arise when software distributed in machine-executable form is involved.)
Documentation: A company's own documentation may clearly reveal patent infringement, and such evidence holds up well in court.
Standards: A company may advertise compliance with certain standards, which can be purchased and may reveal which patents are being used.
What makes a good patent
Schechter also highlighted some key tips on acquiring, enforcing and licensing good software patents.
"Critical to having an effective patent strategy for software, and really for any technology, is, 'Think who the infringer will be when you write the patent,'" he said, "not 10 years later when you are trying to enforce it."
He showed the audience examples of three different ways to describe, or claim, a particular software invention when applying for a patent. Most claims for software patents fall into one of three types: method, apparatus or article of manufacture. He demonstrated how the type of claim a company chooses when it writes a patent will have a dramatic effect on what parties it will be able to sue for patent infringement.
For example, said Schechter, let's say the patentable invention involves a way of moving data from one aspect of a computer system to another, from a high-performance storage system to a low-performance one. If the invention is described as a method claim, then the infringer would be the end user, the person sitting in his or her office manipulating the data. Schechter said this is undesirable because it's hard to find everyone who uses your system.
In another scenario, if the invention is described as an apparatus claim, and Internet computing is involved, you could end up with no one to sue for infringement because "the processor could be anywhere, and the high-performance and low-performance systems could be owned by different parties in different places, made by different people." The optimum choice of claim wording depends on the specific invention and company.
Another tip for software companies, according to Schechter, is to stay in touch with their inventors. He called inventors "a resource which corporate intellectual property departments tend to miss...Inventors may not understand the legal aspects, but they know who's out there selling technology that smells like theirs."
In an interesting twist on its patenting and licensing stance, IBM, according to news reports, has announced a $40 million donation of its software tools to the public domain in an attempt to create an open-source organization aimed at developers.
"Clearly the issue of discoverability of how software works is a big difference of open-source software," Schechter said. "If a company wanted to know whether its patents were being infringed by an open-source software product, it would be easy to learn because the source code is human-readable. The flip side is, you really have to read the open-source licenses very carefully."
Semiconductor industry on overdrive
How is the current IP environment affecting the semiconductor industry? Wharton management professor Rosemarie Ziedonis said that semiconductor companies are at the low end of the list of industries in their reliance on patents relative to R&D costs. Because their product life cycles are six to 18 months, that's not surprising. Yet, as Ziedonis demonstrated, since the early '80s the industry has seen an explosion of patenting that outpaces other industries. What is driving this trend?
Ziedonis identified several factors such as the considerable patenting activity of a large, single company like Texas Instruments; the great number of new entrants into the field; and the burst of products related to wireless applications. But a more pervasive, overall cause, in her opinion, was broad legal change in the 1980s that functioned for many companies as a wake-up call for managing their intellectual property.
A more unified patent doctrine was coupled with several high-profile cases, such as the 1985 case in which Polaroid won more than $900 million in damages from Kodak for instant-camera patent infringement. At the same time, said Ziedonis, Texas Instruments was demonstrating that patenting could be a lucrative business strategy. It won victories with its integrated circuit patent as well as patents pertaining to how to transport wafers and how to encapsulate chips with plastic before connecting them to the motherboard.
"Other companies started thinking, 'Maybe I'm sitting on something lucrative too, or maybe someone else is about to charge me more for using their invention,'" said Ziedonis. The industry moved more toward accumulating patent portfolios for defensive purposes, identifying potentially patentable technologies early on, and offering bonuses and awards to employees for patentable inventions.
"Comparatively, semiconductor companies don't rely heavily on patents to profit from innovation," said Ziedonis, "but the flip side is that today, few of them can afford not to pay attention to patenting. That's one of the main changes that has taken place in the industry. And the legal changes dovetail into the competitive context." Like Schechter in his description of the software industry, Ziedonis characterized semiconductors as "a complex technological setting where to compete and sell your product, you need to buy rights from lots of design companies and providers."
Patenting, however, is expensive, Ziedonis pointed out. A single patent can entail a $20,000 application cost plus maintenance fees. Some companies are choosing instead to disclose their inventions in new online publication outlets, which some compare to the technical bulletins IBM used to generate. But Ziedonis does not think this phenomenon will replace patenting, since companies would lose their patent portfolios--their "pound or two of patents"--that many at the conference acknowledged is a valuable bargaining chip in legal negotiations.
While industries and companies vary in the extent to which they rely on patents and licensing, it is clear that having a definite intellectual property strategy is necessary--and potentially lucrative--for high-technology companies.
"Licensing its IP saw IBM through some very dark days," Schechter said. "In the early 1990s there were some years when the money we made from it literally exceeded the profits of the company."
To read more articles like this one, visit Knowledge@Wharton.
All materials copyright © 2002 of the Wharton School of the University of Pennsylvania.