Linux Today: Linux News On Internet Time.

InternetWorld: The Truth About Patents

Apr 23, 2000, 17:23 (3 Talkback[s])
(Other stories by Bill Roberts)

By Bill Roberts, InternetWorld

Five Reasons They'll Have Limited Impact On The Web's Evolution.

The patent frenzy sweeping the Internet doesn't phase Jerry Kaplan, CEO of Egghead.com.

"Patents will not determine the near-term success or failure of any Internet company," he says. Kaplan's own strategy: He's secured a patent for the company's online auction method and hopes to license the technology to others--but at this point he has no plans to sue anyone.

Notwithstanding Kaplan's advice--and the fact that there are many Internet CEOs who agree with him--it's sure tough not to get panicked about patents.

Two widely watched lawsuits alleging patent infringement do nothing to soothe most fears. In one case, Amazon.com is battling to block Barnesandnoble.com from using Amazon's "1-Click" Web shopping feature. In another case, Priceline.com claims Microsoft is improperly using Priceline's reverse-auction technology.

Internet companies, many flush with resources from venture capitalists or successful IPOs, are aggressively getting patent applications into the pipeline. A few hundred Internet-related patents have been granted, and hundreds, if not thousands, more may be wending their way through the U.S. Patent and Trademark Office (PTO). Many insiders are wondering: Is this the point in Internet history when powerhouses begin to use patents to put a stranglehold on e-commerce?

Not likely. Patents are important, and every company needs a strategy to protect its intellectual property. But any fear of an impending patent war that would forever change the shape of the Internet or e-commerce is overwrought. "They're a small-to-moderate competitive tool people use to threaten others," Kaplan says. It's rare that any single patent gives the holder a lock on a technology or a business method.

Even Amazon.com's CEO, Jeff Bezos, admits there's much more to building a successful Internet business than lots of patents. "The vast majority of our competitive advantage will continue to come not from patents, but from raising the bar on things like service, price, and selection," he wrote to customers in a recent open message. In that letter, written after he was harshly criticized for his company's patent strategy, Bezos joined the chorus of people calling for reform of the patent approval process. He recommends the lifespan be shortened to 5 years or less, instead of 17, for patents that cover business methods and software--the predominant type of Internet patents.

Some patents now being issued are for the nuts and bolts of the technology itself--what patents have traditionally protected. Many, though, are a relatively new breed of patent issued to protect business methods that have been migrated from the physical world to the Internet--Priceline's reverse-auction format, for example. Several infringement suits have been filed, at least one of which has been settled out of court for $15 million. And several companies are offering patent licenses, for a fee.

Paul Esdale, vice president of corporate development at Open Market Inc., is among the executives pursuing an aggressive patent strategy. "Those who believe that Internet patents aren't relevant are just kidding themselves," says Esdale, who is trying to license technology for five patents involving e-commerce software.

But will patents dictate how the Internet evolves? Probably not, says Ullas Niak, an investment analyst at First Albany Corp. who evaluates Internet companies. Niak rates patents much lower in importance than a company's potential market, management, and other factors. "Patents serve mainly as a deterrent to competitors," he says.

Patents will have the same impact on the Internet as they've had elsewhere. They'll force competitors to pay attention to each other's innovations, explore licensing and cross-licensing options, and sue when necessary. They'll spur innovation and protect inventors, which are the same goals the patent system had when it was founded by Ben Franklin in the early days of the Republic.

With one difference: Because the speed of Internet innovation is so much faster than the pace at which patents are granted, licenses negotiated, and suits fought, many patents may quickly prove irrelevant. Patents take three years or longer to be granted, licenses can take a year or longer to negotiate, and infringement suits take years to plod through trial and appeals. Internet innovation, however, is measured in weeks or months. With this in mind, experts identify five reasons why Internet patents, although important, will only have a modest impact on the evolution of the Web and e-commerce.

It is a rare patent holder who wants to spend money and time to take a suit all the way to trial. Yet a court victory is the best gauge of a patent's strength. There's no telling how significant the Amazon or Priceline patents might be until they win or lose in court.

"Litigation is the best test of validity," says Alan Fisch, an attorney with Howrey Simon Arnold & White, a Washington, D.C., law firm. "During this process, the accused infringer has the opportunity to prove, by clear and convincing evidence, that the patent-in-suit is invalid."

Litigation, however, is the least common patent strategy. Companies are more likely to offer licensing programs, or to use patents for defensive purposes if someone tries to sue them. "Most patent owners want to license their patents," says Marc Brown, a partner in the Los Angeles law office of Oppenheimer, Wolff & Donnelly. "They don't want to put you out of business."

There have been a dozen or more Internet patent infringement suits filed so far, but none have gone to trial. "I've seen relatively little significant patent litigation in the Internet space," says Kaplan, who was CEO of Onsale Inc. when it merged with Egghead in November. At the time of the merger, Onsale had one U.S. patent [No. 5,835,896]. It has others pending for its online auction methods.

Kaplan says his main patent strategy is defensive in nature. Despite the occasional claim that Egghead infringes on some other patent, the company hasn't had to pay any licensing fees yet. Kaplan attributes that to his company's own patent. Egghead also plans to launch a licensing program after it gets more patents, but Kaplan doesn't expect litigation to be part of the strategy. Why? "It takes years and millions of dollars to pursue a patent infringement case."

Others agree. "Our long-term objective is to license the patents to people for whom they are relevant," says Nat Goldhaber, CEO of Cybergold Inc. It has two U.S. patents: One for the incentives-based sponsorship program it runs at its Web site and markets to others [No. 5,794,210], and one for privacy protection [No. 5,855,008]. It is negotiating licenses--in one case it's a cross-license--with a handful of sites for the sponsorship patent, and hasn't decided what approach to take with the other one, says Goldhaber. Cybergold hasn't sued or been sued.

Open Market may have some of the more seminal U.S. patents. It filed applications early and often for patents to cover technology for secure online payments and other networked commerce processes. Five have been granted [Nos. 5,909,492; 5,812,776; 5,724,424; 5,715,314; and 5,708,780]. It has negotiated with several potential licensees for more than a year, but has no deals yet. And despite the frustrating pace, it has declined to sue.

Open Market wants to foster, not thwart, the Internet's evolution, says Esdale. However, he admits, "others have decided that litigation is a prerequisite to meaningful licensing discussions, and we're watching them with interest."

Some experts say even Priceline and Microsoft will settle without a trial--even though, of all the Internet-related patent suits filed thus far, this one is most likely to go the distance, given both companies' deep pockets. Having touted its reverse auction [U.S. Patent No. 5,794,207] as a barrier to entry when it was wooing investors, Priceline may have more to lose than most if its patent isn't strengthened through a trial verdict, experts say. Priceline executives declined to be interviewed.

Given the choice of fighting in court or paying a license fee--even a stiff one--most companies, when sued for patent infringement, will pay the fee. The first thing an attorney does when a client gets an infringement notice is to evaluate the patent and determine if the client can avoid it through a design twist. Next, the lawyer looks into a license, says attorney Brown, who works with dozens of Internet companies. "Often, a license is available," he says.

Art Technology Group (ATG), which develops customer relationship management tools and other e-commerce products and services, recently settled a suit filed against it in December 1998 by BroadVision Inc. The fee: $15 million over three years. CEO Jeet Singh says ATG hasn't changed its opinion that it is not infringing upon the patent [No. 5,710,887], which covers aspects of one-to-one commerce. Settling the suit, he says, had little to do with the patent's merits, and everything to do with a desire to resolve the issue and get on with business in a fast-growing market. Paying even a stiff fee made better business sense than continuing to fight, Singh says. "It is to our advantage to put our resources into growing the business, rather than slogging through this complex and costly legal process."

BroadVision didn't really want to sell a license and has no intention of offering a licensing program, says Peter Downs, the company treasurer. It just wanted ATG to stop infringing, but "it was time to settle this thing," Downs says. "They're paying us quite a bit of money, and technology marches on. We need to move forward with our own R&D efforts."

Most license fees are much more affordable. "Our philosophy is not to make the pursuit of business uneconomical to people," says West Shell III, CEO of Netcentives Inc., which has one of the more successful licensing programs to date. Netcentives, which develops software and services for electronic marketing, has two U.S. patents for customer loyalty methods [Nos. 6,009,412 and 5,774,870]. It is negotiating with several companies, including competitors, and has licensed its patents to AOL, Lycos, LookSmart, and American Express, says Shell, who declines to specify fees.

Shell says Netcentives is open to cross-licensing or paying fees to others if it must, but to date it hasn't. Netcentives and Cybergold compete, and their patents are for similar processes. Neither Shell nor Cybergold's Goldhaber will comment on whether they're negotiating a cross-license, which would be one of the first in the Internet industry.

Some firms will pay for a license even if there's a good chance the patent could be beaten, says Greg Aharonian, a professional patent-buster. He makes his living searching for documents and products, known as "prior art," that were overlooked by the patent applicant and the PTO examiner, and which would prove that the innovation in question does not meet the statutory requirements of being "new, novel, and non-obvious." Prior art can make or break a patent's value or validity. It's defined as any document or product that's publicly available before the patent application is filed and that anticipates the technology in the application.

One Internet-related U.S. patent Aharonian considers dubious is for compressing database information sent over a network [No. 5,253,341]. License fees have gleaned hundreds of thousands of dollars for the patent holder, TechSearch LLC, says Aharonian, because "the $50,000 or whatever they're asking for a license isn't worth fighting in court." A licensing fee is far cheaper than legal fees, and in many cases may be cheaper than revamping a Web site to avoid infringement, he notes.

If patent holders become too greedy, they could find the market evolving without them. This is especially true of business-methods patents in e-commerce. "Market forces will dictate how the holders of patents will be compensated, one way or another," says Drummond Reed, chief technology officer at OneName Corp., which is developing a platform for automated data exchange. OneName will announce products and a licensing program this spring. It has one U.S. patent for communication objects known as Web agents [No. 5,862,325] and three others pending. It is the only company Internet World interviewed that has pursued, and been granted, patents in other countries.

Reed likens market forces to a raging primordial river slashing through rock and earth. The river might make a small bend to accommodate an obstacle, he says, but its general flow won't be stopped. "If the holders of patents stake too strong a claim," Reed says, "the river will just go in another direction."

Attorneys tend to agree. Lawyer Brown urges clients to file for patent protection for their own innovations, but not to worry about other companies' patents. If the client is hit with a suit, the design of the process or technology can often be tweaked to avoid the infringement, he says.

The Amazon case is such an example. Amazon sued Barnes & Noble for infringing on its 1-Click buying method [U.S. Patent No. 5,960,411]. It won a court injunction, which prohibited its bookselling competitor from using one-click while the matter is settled. In her ruling, however, the judge said Barnes & Noble's proposed two-click buying button was not likely to infringe on Amazon's patent.

"There was this mass hysteria that the Amazon patent would destroy the Web," says Brown. "But Barnes & Noble goes right on with business." Amazon executives did not return calls to comment on the case or to discuss their strategy for eight other patents.

Brown thinks the one-click buy could benefit Amazon financially, since it will cause more impulse buys, but he admits that one can also argue that two-click is better for consumers, giving them a second chance to think about a purchase. Only time will tell, but Barnes & Noble could lose little or nothing. "By holding that piece tight, Amazon potentially becomes the poster child of the river-going-around-you analogy," says OneName's Reed.

The rare company that fights a patent suit all the way to a trial verdict may find that some patents can be busted. Whether they can or can't be busted will be less important than the perception that some can--a perception that may keep patent holders from getting too aggressive. Aharonian, who has examined prior art for hundreds of patents, says Internet patent applicants aren't doing adequate searches. And the PTO doesn't have adequately trained staff or resources to fully search prior art either, he adds. The result, says Aharonian: Because prior art for most Internet patents is not well searched, the patents are much more vulnerable in court.

A lot of Internet research was done in the 1980s. Some of it resulted in patents, and a lot just ended up in scientific journals, university and corporate libraries, and other places, Aharonian says. If it is public record, it is potential prior art, he adds. He says the number of non-patent prior art citations appearing on software patents has gradually declined since 1994, which indicates that applicants are not doing rigorous prior art searches. One problem, he says, is that the non-patented research isn't easy to dig up. "You have to go through back volumes of stuff page by page to find it," he says. Aharonian, based in San Francisco, spends a lot of time in the libraries at Stanford University and the University of California at Berkeley.

Companies can help their own cause by placing as much information as possible into the public realm, says James Pooley, a patent attorney and partner at Gray Cary Ware & Freidenrich LLP, Palo Alto, Calif. He points out that IBM has published research and other material for decades. Much of it never becomes patented, but it does become prior art and makes it more difficult for competitors to prove that a similar innovation is new, novel, and non-obvious. "Make sure a lot of stuff gets into the public domain so there's very little room for someone to maneuver in issuing a patent," says Pooley. He suggests that companies work together to create libraries or museums of such material.

So-called business-methods patents are most susceptible to busting. Critics, including Aharonian, argue that the PTO shouldn't be granting such patents. The PTO, however, has decided that putting a physical-world process--like a reverse-auction format--on a network can be new and novel, and therefore worthy of a patent. The business-method patent as a type was upheld by the federal court in a precedent-setting, non-Internet case nearly two years ago (State Street Bank & Trust v. Signature Financial Group).

Still, Aharonian and others believe that overlooked prior art will show that many Internet business-methods patents do not meet the requirement of being new, novel, and non-obvious. Many believe even Amazon's 1-Click patent is frivolous and would not withstand the scrutiny of a thorough search of prior art.

Nothing about patents operates on Internet time. That's to the advantage of the company alleged to be infringing. If the Internet moves like a raging river, then the patent process is as slow as a glacier. "By the time the issue gets resolved, the chances are the businesses will have succeeded or failed on their own merits rather than on the value of the patents," says Egghead's Kaplan.

The fact is, given that it takes about three years to be granted a patent, the technology it covers could be obsolete by the time it is awarded. "Patents often have a potential legal life that surpasses the protected idea's practical life span," says attorney Fisch. It's possible an Internet startup could begin and fail before the PTO issues its patents, he notes. Fisch points out that Bezos' argument for a five-year patent lifespan implicitly admits that the innovation is likely to become obsolete in that time period anyway. That being the case, he adds, it doesn't matter how long the life of the patent is.

Esdale admits that Open Market has learned a tough lesson from its licensing efforts. "In the patent area, the time frame is more akin to traditional industries. The Internet moves faster than anything any of us has been involved in." He says there's a tremendous amount of research that a potential licensee must do once the patent is presented to them. It can take three months or longer before there's a second contact between the two parties. "The cadence of patent licensing is not as fast as other things we know," Esdale says.

Meanwhile, the pace of Internet innovation doesn't slow down. While lawyers negotiate licenses, or file and counter-file suits, innovation moves ahead. Reed expects that when the Web agent platform OneName is developing takes off, there'll be new innovation to move many business methods onto the platform, thus antiquating a previous business-method breakthrough. "How much we could enhance the reverse-auction process with Web agents is unclear," he says, "but it wouldn't take much to argue that you've got more innovation."

That doesn't mean that companies should avoid securing their own patents. Everyone interviewed agrees that an Internet company needs a patent strategy. Most urge companies to file for a patent if they think they've developed innovative technology or business methods. Attorneys say they meet too many young companies that were too busy during their startup days to consider applying for patents, which is shortsighted. "Companies should develop their own program of creating intellectual property and protecting it so they have their own cards to trade if someone comes knocking," Pooley says.

Brown agrees: "An average patent application costs about $10,000. That's a drop in the bucket to setting up the Web site. If you have a new, enhanced approach to doing business and are going to spend $100,000 designing the Web site, it's a no-brainer. File for the patent."

Related Stories: