Few venture capital firms have maintained as high a profile as Kleiner Perkins Caufield & Byers. The firm's portfolio includes such notable names as @Home, Netscape, Marimba, and Sun.
Kleiner Perkins claims that one of the reasons that it has been so successful is its keiretsu, orits network of portfolio companies. In an August 11 profile of Kleiner Perkins partner John Doerr in The New Yorker,Amazon.com CEO Jeff Bezos said that even though Kleiner Perkins had placed a lower valuation on Amazon.com than any other firm, he chose Kleiner Perkins to lead his company's first venture round because of the power of the firm's portfolio.
Interviews with Mr. Doerr, other venture capitalists, corporate investors, entrepreneurs, and academics suggest that a keiretsu is an extremely amorphous concept. At their best, these networks can help businesses develop relationships with other private companies and large corporations more efficiently. But at their worst, they are mere marketing hooks.
With venture capital in abundant supply, VC firms like Kleiner Perkins believe that networks are more important than ever in distinguishing them from their competitors. From 1992 to 1996, Coopers & Lybrand estimates, the median amount of capital raised per round increased from $2.5 million to $3.8 million; median premoney valuations during the same period rose from $8.1 million to $13.1 million. Since too much money is chasing too few deals, venture investors have to convince entrepreneurs like Mr. Bezos to take their money rather than somebody else's. But recently, two corporate venture investors have chosen different approaches to establishing portfolio networks. Their takes on the idea of a keiretsu raise important questions about the effectiveness of mainstream VCs' portfolio networks.
JapantownKeiretsu is a Japanese term that describes the set of interlocking relationships among Japanese suppliers and manufacturers. According to Mr. Doerr, "Keiretsu are rooted in the principle that it is really hard to get an important company going and that the fastest and surest way to build an important new company is to work with partners." In addition to providing its portfolio companies with regular updates on the strategies and plans of theirpeer group, Kleiner Perkins fosters coцperation by organizing a half dozen formal gatherings of portfolio company CEOs and other key executives each year.
Kleiner Perkins encourages its companies to help each other by forming "buy-sell, licensing, or endorsement arrangements," Mr. Doerr says. As an example, he cites Destineer, which Kleiner Perkins and Mobile Telecommunications Technologies incubated in 1990 to develop a one- and two-way nationwide messaging service that would use narrowband Personal Communications Services frequencies. Since Wireless Access, another Kleiner Perkins company, was developing advanced pager technology, Kevin Compton, a Kleiner Perkins partner, suggested that the two companies work together. Mobile subsequently invested in Wireless Access, and the two companies, along with Motorola and Destineer, codeveloped protocol, networking, and chip technologies that have been incorporated into SkyTel, Mobile's paging network.
But keiretsu is perhaps an odd label for this process. Keiretsu are regulated by Japanese law, and they are structured in such a way that coцperation between portfolio companies is virtually mandatory. In Kleiner Perkins's conception of a keiretsu, however, the venture firm does little more than encourage its portfolio companies to work together. When pressed, Mr. Doerr admits that his firm's use of the term is "inappropriate." He says, "If the entrepreneurs want, we can open some doors, but they walk through them and make the deals." One VC, who asked to remain nameless, sees even less structure in venture capital keiretsu: "I think the word keiretsu is just a sound bite with great PR value."
The universal networkIndeed, though Kleiner Perkins might use the most exotic word to describe its network, all the major venture capital firms foster coцperation among their portfolio companies. The Mayfield Fund, for instance, holds regular dinners for the top executives of its portfolio companies, sponsors a two-day retreat for entrepreneurs in the medical devices arena, and organizes aconferencecalled Internet Impact with the law firm Wilson Sonsini Goodrich & Rosati. Accel Partners also frequently brings the key executives of its roster together. At Accel's annual "CEO day" last October, the VCs invited 50 CEOs to rub shoulders with the likes of Greg Maffei, Microsoft's chief financial officer; Avram Miller, Intel's director of business development; and John Chambers, Cisco's CEO.
But as important as this cross-fertilization process is, the networks of venture capital portfolio companies can be somewhat nebulous. In the first place, since private companies often have more than one venture investor, they could theoretically belong to several different keiretsu. The question arises: at what point do keiretsu end and ordinary business alliances begin? In addition, although VCs may purport to invest in specific industry segments, the categories that they come up with are often catchalls--one of Kleiner Perkins's eight investment initiatives is "The Net."
Splitting crosshairsKen Fox of the Internet Capital Group (ICG) argues that keiretsu lack focus. "The only time portfolio companies get excited about working together is when they fit right into each other's strategic crosshairs," he says.
Mr. Fox thinks that his firm has discovered a better way. A private spin-off of Safeguard Scientifics, a publicly traded company (NYSE: SFE) that invests in high-tech ventures, ICG funds only early-stage intranet and electronic commerce companies. Furthermore, ICG's structure as a limited liability corporation is designed to ensure that it maintains its focus. Because ICG intends to go public, 55 percentof its income and assets must come from companies that it controls. (The Securities and Exchange Commission mandates that operating companies like ICG must own more than 25 percent of the companies that they control and be the largest shareholder in them.) ICG thus plans to invest 50 percent of its fund into three "core" companies, whereas traditional VC firms rarely place more than 10 percent of their capital into one company. Such large bets theoretically mandate that all of ICG's investments further the success of its core companies.
The company also plans to issue ICG shares to all of its portfolio companies. Sharing the ownership of the company will give the entire portfolio a concrete financial incentive to work together.
ICG has a plethora of financing options that could allow it to take a more strategic perspective than most VC firms. The company plans to raise capital from the public markets, the cheapest source of money available. As a corporation, ICG is also able to issue debt, and it has already been approached by banks eager to lend it money (even though none of ICG's companies has gone public). And unlike traditional venture partnerships, which must pay out their investments within seven to ten years, ICG can invest in companies for as long as it likes.
A mainstream VC calls ICG's investment thesis "financially foolish" because the leverage that ICG purports to have could be significantly undermined by its high-risk profile. Mr. Fox admits that ICG's concept is unproven, but he says that the worst thing that could happen to ICG would be that it loses its status as an operating company and becomes, in effect, a VC firm. Pointing to Kleiner Perkins's Java Fund, a separate fund that invests only in Java-based companies, he predicts that more traditional VC firms will adopt a focused approach.
The antikeiretsuOn the other end of the spectrum, Colin O'Brien,CEO of Xerox New Enterprises (XNE), thinks searching for synergies among venture capital portfolio companies can be a big waste of time. He dismisses VC keiretsu as a "marketing ploy" and says that he doesn't believe that either VCs or the CEOs of the companies backed by the VCs want to lose focus on their business by "helping somebody out."
After years of letting its best innovations run out the door, Xerox formed XNE two years ago to harness some of its advances (see "Postmodern PARC). Since then, XNE has taken two companies public--Document Sciences (Nasdaq: DOCX), which holds a market capitalization of $36.3 million, and Documentum (Nasdaq: DCTM), which holds a valuation of $448 million. XNE is aiming to reach a total of $1 billion in value by the year 2000.
Mr. O'Brien believes that his companies have a much better chance of achieving synergies than traditional venture-backed companies. With the exception of Documentum and Document Sciences, XNE maintains a majority share in each of its nine companies, and the engineering staffs of its portfolio companies have full access to Xerox's labs. Mr. O'Brien's group handles its companies' administration in return for a fee (approximately 1 percent of revenue), thus saving them the substantial headache of having to sort through issues like medical plans. That's really the extent of the formal network; Mr. O'Brien is more interested in making money than making friends. "I want them to make a success of what they've got," he says.
Different strokesBut though some networks are less elaborate than others, nearly all firms that cultivate startups use them because their benefits are incontestable. And it's only natural that different networks have arisen to meet different needs. All the debate about network structures amuses William Sahlman, a professor of entrepreneurial management at Harvard Business School. "Everybody thinks their model is better than everybody else's," he observes. "But models evolve to match people and projects."
Calling the American venture firm a "highly evolved form," he says there is no evidence that it is not working effectively and adds that it has certainly been more successful than corporate investors in working with startups. Still, the venture capital world traditionally runs in cycles, and only a significant downturn in the market will yield any meaningful results about the power of different network structures, he predicts.
In the meantime, Mr. Sahlman is content to watch and wait. "I believe it's useful to think about how to strengthen networks," he concludes. "But networks are going to exist whether you call them networks or not."
THE NETWORK IS THE INVESTMENTFour models for cross-company networking.
JAPANESE KEIRETSUStructure Horizontal associations involving banks and general trading firms as well as vertical associations revolving around parent companies. Subsidiaries frequently serve as suppliers, distributors, and retail outlets.
Characteristics Member companies cross-hold shares, are financed internally, invest jointly, coцperatively appoint officers, and engage in other joint business activities.
TRADITIONAL VC KEIRETSUStructure Informal network of private companies within portfolio of venture capital firm.
Characteristics VC firm organizes periodic gatherings to build bridges among portfolio companies' key personnel in order to accelerate marketing initiatives, development of key technologies, due diligence, and recruiting.
INTERNET CAPITAL GROUP Structure Informal network of private companies within portfolio of an independent operating company; network is built around two or three "core" companies that collectively receive at least 50 percent of the operating company's investment fund.
Characteristics Investment concentration on core companies theoretically mandates that all ICG's investments further their success. All companies receive shares in the operating company, giving them a direct financial incentive to coцperate in accelerating development of key technologies, marketing initiatives, due diligence, and recruiting.
XEROX NEW ENTERPRISESStructure Collection of subsidiaries of a corporate division that all have their own budgets.
Characteristics Subsidiary companies have access to Xerox's labs and administrative services in return for approximately 1 percent of their revenues. Coцperation among companies is not actively encouraged.