(FORTUNE Magazine) – You might be inclined to think that the most important thing about a technology startup is, say, oh, the technology. Venture capitalists don't see it that way--and the way they see it is remarkably important since they're the middlemen who turn today's bright ideas into tomorrow's Intels and Yahoos. To a venture capitalist, the crown jewels of a hot new tech company are not closely guarded software codes but the talented people who can quickly transform that technology into a revenue-generating product that leads the company to a blockbuster IPO or a lucrative acquisition. These stars, dubbed simply "the talent," aren't necessarily young and restless Gen Xers slaving away in garages. They're people who, in one way or another, have been around long enough to earn a few stripes. They're people like Keith Krach.

An affable 40-year-old engineer (yes, 40) with a Harvard MBA, Krach is exactly the kind of entrepreneur today's Silicon Valley VCs are drooling to give money to. He's worked at both a big company and a startup, and he's been in both management and engineering. Most important, he's helped lots of people make lots of money. All that helps explain why, for six months in 1996, Krach occupied an office at the venture firm Benchmark Capital and why he held a position you'll find only in Silicon Valley. His title: "entrepreneur in residence."

The so-called EIR concept has been around in various forms for about a decade; as competition for the talent has grown feverish, some VC firms are relying on it more and more. Technically, what VC firms do is fund startups in return for stock that might someday be worth a lot of money. But the average venture capitalist actually spends most of his time schmoozing with potential managerial talent for startups he's funded and building a treasured database of contacts. Last May, Benchmark hired David Beirne, the 34-year-old hotshot executive recruiter who, among other coups of persuasion, brought Jim Barksdale to Netscape and lured AT&T President Alex Mandl to a little-known startup now called Teligent. Beirne wasn't hired just to spot great new technologies or deals. "Ideas and money are commodities," he says. "If you have an A team of talented people, you'll win."

Keith Krach was in the Rolodex of Benchmark partner Bob Kagle for years. The two got acquainted when Krach was a rising-star engineer at General Motors. (He became GM's youngest vice president at age 26.) From 1982 to 1987, while Krach directed a GM robotics joint venture that went from scratch to market leadership and annual revenues of $200 million, Kagle would meet Krach for breakfast several times a year. In 1987, Krach became COO of Rasna, a startup that sold software for mechanical engineers. In nine years he helped position the company for a $200 million sale to Parametric Technology in 1995, a bonanza that produced 31 Rasna millionaires and returned many millions to happy venture investors. Soon after the sale, Krach left Rasna, and Kagle saw the opportunity to get him into Benchmark's entrepreneur-in-residence program

Krach had many other options. A few well-placed phone calls to other VC firms or individual investors would have secured financing for his own startup. Heck, since Krach walked away from the Rasna deal with more than $10 million, he could have bankrolled a startup himself. That would have enabled him to retain ownership of much of the company's stock. Krach could also have taken one of a number of CEO jobs available at existing startups, where the going rate is typically $200,000 a year in salary and options on 5% to 10% of the company's stock. Or he could have simply retired; moved his wife, Jennifer, and three kids to the south of France; and spent the rest of his days picking mushrooms. But Krach liked Kagle. He decided he might learn a few things by hanging out with venture capitalists for a while.

Besides, at the time Krach wasn't sure what kind of company he wanted to get involved with next. To hear him tell it, he still hadn't figured out what exactly he was going to do when he grew up. To the five partners at Benchmark, a 2 1/2-year-old firm formed by Young Turks from old-line venture capital firms Merrill Pickard and Technology Venture Investors, Krach's lack of direction didn't matter; he would eventually figure it out. As a Benchmark EIR, Krach would be paid $10,000 a month (Silicon Valley's concept of a minimum wage) to "incubate" an idea.

From an office at Benchmark's digs on Sand Hill Road, a strip of land in Menlo Park, Calif., so crowded with moneymen that it is to Silicon Valley what Wall Street once was to the robber barons, Krach would spend his days playing venture capitalist--meeting with industry leaders and other VCs, attending Benchmark's Monday-morning partners' meetings, listening to pitches from eager entrepreneurs looking for funding, and generally ruminating on the future of technology. Eventually, if he felt like it, he could form his own company and Benchmark would fund it.

Nice life, huh? Being an entrepreneur in residence is indeed one of the dream jobs of the late 20th century, but it's hardly some kind of VC-subsidized welfare. Venture firms often get better returns from companies created by EIRs than they get by buying equity in existing startups. "These companies generate our greatest returns," says Andrew Rachleff, a partner at Benchmark, where half the companies funded are EIRs. "There's much less risk." Another firm, Mayfield Capital, estimates that its incubated companies deliver a 75% better return. "EIRs are a way to go in earlier in the food chain. If I can get someone who I know is a great entrepreneur, and a great manager who is going to do something fantastic, I can put a net around them," says Mayfield partner Michael Levinthal.

Rachleff points to two runaway successes at his previous firm, Merrill Pickard, to show just how well VCs can do with EIR companies. Memory-chip developer Rambus was founded by Mike Farmwald when he was an EIR at Merrill. Offered at $12 a share last May, Rambus stock opened its first day of trading at $22.50 and then sailed as high as $86 in August. "Rambus returned our whole fund [at Merrill]," beams Rachleff, recalling the $100 million Merrill made from the company. Similarly, Jeff Hawkins incubated Palm Computing, whose PalmPilot digital organizer became one of infotech's most successful consumer-product launches, after Bruce Dunlevie--now a partner at Benchmark--persuaded him to incubate his company at Merrill in 1991.

VC firms need those Rambuses and Palm Computings. Competition in the venture business is getting fierce. Once upon a time entrepreneurs had to beg and grovel to get funding. But recent record levels of venture money have turned the tables.

These days there's a ton of money chasing a relatively small number of promising deals. The number of VC firms in existence is at an all-time high. According to Price Waterhouse and Boston research firm Venture Economics, VC firms collectively invested $5.9 billion in technology startups in 1996 and $8.4 billion in 1997 (up from just $1.6 billion in 1993). There are also swelling numbers of so-called angel investors, wealthy individuals and tech-industry veterans (often one and the same) who relish the vicarious joy and respectable returns of doing early-stage funding of new tech companies. Jeffrey Sohl at the University of New Hampshire's Center for Venture Research estimates that angel investors committed at least $10 billion to technology startups last year. There are also more and more large technology companies eager to make sizable investments in startups. Microsoft and Intel alone invested almost $1 billion in them in the past two years.

The abundance of capital is thanks largely to the frothy IPO market of the past three years, which has enabled VC funds to post spectacular returns. The average annual return of VC funds was 48% in 1995, 40% in 1996, and about 36% in 1997, according to Venture Economics. Some annual returns are simply astounding, such as Benchmark's 100% on its first fund. Results like those, of course, attract even more money from the pension funds and college endowments that invest in venture capital funds as limited partners.

The downside of all this cash rushing into Silicon Valley is that VC firms find themselves in the humbling position of having to sell entrepreneurs on why their money is better than somebody else's. To do so, they pound their chests about successful companies they've funded, tout their partners' specific areas of expertise, and use the term "added value" a lot. For entrepreneurs, it's never been easier to raise large sums of money in Silicon Valley. To be sure, some would-be Bill Gates may still go unfunded. But anyone with a track record like Krach's will find investors lining up at the door.

Krach chose to be an EIR at Benchmark because he liked Kagle's promise of time to ponder his next move. The soul-searching took about four months. After a string of breakfasts and lunches at Il Fornaio, Silicon Valley's undisputed power eatery of the moment (sample mogul meal: antipasto and pollo toscana, at $20), and meetings in office buildings up and down Highway 101, Krach hooked up with, of all people, another Benchmark entrepreneur in residence. Paul Hegarty, ex-vice president of engineering at Next Software and a close associate of Steve Jobs', had come to Benchmark three months after Krach.

Together they hit upon the golden startup idea--automating the purchase of office supplies and services at large companies. Hegarty developed a blueprint for software that would enable every employee in a business to order mundane but indispensable supplies--paper clips, file cabinets, pencils--by using desktop PCs to access a site on the company's intranet. In most companies procurement of such "nonstrategic goods and services" is an inefficient, paper-intensive process. "The size of the market is huge. Any global 2000 company has the need for this," enthuses Krach, "and with the Internet there's a paradigm shift where you can put this software on everyone's desktop." The idea may not sound especially sexy (in fact we're sure it doesn't), but for people who live and breathe technology, Krach's creation, Ariba Technologies, is about as racy as they come.

Ariba is what's known as an electronic-commerce company, currently one of the hottest areas in Silicon Valley. Its product is written in Java, the Internet programming language everyone wants to use for new applications, and is in a software sector Microsoft hasn't even contemplated entering. Best of all, by developing a new type of enterprise software, Ariba has coined a brand-new acronym to add to the ever-expanding lexicon of technobabble--ORM, standing for what else but operating resources management.

Joining Krach and Hegarty on the frontier of ORM software solutions were five other founders, rounding out a complete management team. Two of them--Bobby Lent, vice president of strategy; and Boris Putanec, head of development--were EIRs at Crosspoint Technology Ventures, a firm that has a relationship with Benchmark. The other co-founders worked with Krach at Rasna--Rob DeSantis, vice president of sales; Edward Kinsey, CFO; and Paul Touw, vice president of business development.

What VCs expect to get from their EIRs is the chance to invest in a company's early stages, when the stock's the cheapest and the potential for big returns the greatest. (Startup tech companies do a series of funding rounds, and the price per share typically increases with each round.) "We want the opportunity to invest $2 million to $3 million on your first round at a market price," says Rachleff. "If we do our job right, we're going to have an investment opportunity that's vastly superior to anything that's going to come in off the street."

Benchmark got what it wanted with Ariba. When the company officially formed in September 1996 and set out to do a first round of financing, Krach didn't even talk to other venture firms. He and Hegarty had been living in Benchmark's offices, rubbing shoulders with the Benchmark partners at Monday meetings and getting loopy with them at Friday-evening office beer busts. They felt they owed it to them. "Obviously Benchmark was the first group we approached, and their reaction was, 'We want to be the guys that fund you.' We said, 'Here's what we think is a fair price,' and they agreed to it," recalls Krach. Benchmark and Crosspoint didn't try to haggle a lower price for Ariba's stock, and went into the first round as the only investors, putting in $3 million and $2.5 million, respectively, for 19% and 16% stakes in the company.

In theory Krach had been free to skip down Sand Hill Road and take money from, say, Kleiner Perkins or Mayfield, leaving Benchmark out of the deal. Of course, no one actually does that. "The people in our EIR program are under no obligation to do anything with us at all," insists Benchmark's Kagle. But since venture capitalists make it their business to know everyone in Silicon Valley, burning bridges with a major firm probably isn't an enlightened idea. Nor a practical one, since free rent in one of the most congested and expensive real estate markets in the country is nothing to scoff at. It took three months after incorporation for Ariba to find real offices just south of Menlo Park, first in Mountain View and then in Sunnyvale. In the meantime 12 employees crammed college-dorm-room style into four Benchmark offices. Ariba's first two customer meetings were held in Benchmark's boardroom; the first conception of the product architecture was sketched with crayons on a paper tablecloth over lunch at Menlo Park's Quadrus Cafe, last year's Il Fornaio. "A lot of energy comes from that whole process," says Kagle. "There's kind of a special place in their heart for you because you have the offices they first got going in. They can talk to other venture guys, but they've been living with you."

Once venture capitalists own a significant chunk of equity in a company, the game changes. The objective is not to keep other investors away but to find ones who will pay a higher price than you did for the stock, thereby raising the theoretical value of your investment. So after the first round, talking to other VC firms isn't just okay, it's encouraged. Ariba decided to do a second round of funding in June 1997, and Kagle recommended that Krach meet with various VC firms whose calls he'd been receiving regularly since late 1996. He also thought Krach should seek out corporations and other later-stage investors that tend to offer higher prices, since he felt Ariba was now worth a lot more.

For eight months sales manager DeSantis had been talking to large companies across the country about the impact Ariba's product could have on a company's bottom line. The seductive pitch, supported by research from Killen Associates in Palo Alto, was that a 5% drop in purchasing costs may result in a 20%-plus increase in profits. Having seen little more than that, three high-profile customers--Cisco Systems, Advanced Micro Devices, and Octel Communications--put in orders for more than $1 million worth of software, several months before the product would be generally available. Ariba looked like a hot deal.

After conversations with everyone, Krach shunned VC suitors and went with later-stage investors (even though this was only Ariba's second round). It was a lucrative move. On a good day VCs hope to double or triple their startup's valuation from one round of financing to the next. In this second round, however, Amerindo Investment Advisors, Van Wagoner Capital Management, PeopleSoft, and Visa bought Ariba stock at $12.50 per share, or six times the price paid in the first round. Ariba's valuation, the private company equivalent of market cap, had jumped from $16 million to $113 million in nine months.

In fact there had been so much investor interest in Ariba that the company actually turned down an even higher price, fearing the effect of an unrealistic valuation. Amerindo, a San Francisco pension-fund manager, initially offered a price eight times the first round. Emeric McDonald, the Amerindo analyst who spearheaded the investment, liked what he saw when he ran financial models projecting Ariba's revenues and comparing them to publicly traded enterprise-software companies in which Amerindo owns big stakes, like PeopleSoft and Siebel Systems. "Ariba had significant and important customers and had proved the concept. It is the first enterprise application that is entirely Java-based, and I think this market has the potential to be really enormous," explains McDonald.

Ariba took Amerindo's money, though not as big a chunk as McDonald had hoped. Other investors were so eager that Ariba ended up doing what many companies do when they find themselves the belle of the technology investors' ball--it took in more money. Having planned to raise $6 million in the second round, it accepted $13.2 million instead. Amerindo paid $4.5 million for 4% of the company. Compare that to the 19% stake Benchmark bought for $3 million, and you see the advantage of getting in on the ground floor. The value of Benchmark and Crosspoint's investments in Ariba have already increased by a multiple of four (not six, because additional shares were issued in the round and each firm put in additional capital to maintain its percentage.) Crosspoint partner John Mumford thinks that's just the beginning. He hopes ultimately to get a 100-fold return. "You've got to get that out of your winners to make up for the craters you create," says Mumford.

With a 100-fold return, Ariba wouldn't be a mere winner; it would be the Silicon Valley equivalent of Jurassic Park. You have to wonder--can a 1 1/2-year-old company that's had just two rounds of funding already be considered a potential blockbuster? So far, the signs look good. By investing in Ariba, Amerindo and Van Wagoner are betting that an IPO will take place sometime this year--both buy equity only in companies they believe will go public in the next 12 to 18 months.

Even if Ariba doesn't prove to be the next PeopleSoft, Benchmark and Crosspoint may still make out like bandits. When a company in its portfolio goes public, a VC firm generally distributes 80% of its stock to its limited partners, while the general partners get 20%. Those shares may be sold in the public market after a statutory post-IPO waiting period (at least three months). Even if Ariba's share price reaches, say, only $30 after the IPO, Benchmark and Crosspoint would each still make at least $45 million. Ariba employees could cash in some of their options to dash out and buy new Porsches. And Krach would reap another $10 million-plus and go on to the next chapter in the storybook life of a Silicon Valley entrepreneur.