Boeing’s venture arm is in a tailspin. Dell’s has gone blue screen. Applied Materials Ventures has short-circuited. These are just three of a growing number of corporations abandoning or severely curtailing the corporate venture capital investments that emerged in the late 1990s. The recent departures follow the likes of EDS, Hewlett-Packard, Bechtel, British Airways, Quantum, and AT&T, all companies that exited the market after their bubble-era investments failed to yield the expected financial or strategic returns.

But not all corporate VCs have abandoned hopes of profiting from startups. Many are looking back, realizing their mistakes, and retooling their approach. They’re hiring professional VCs to do the dirty work, cutting back to the most strategic areas of investment, and, in some cases, not making investments at all, yet still keeping close ties with the venture capital community as a way to build partnerships with the most promising startups. A few, including Xilinx and biotech Amgen, are even cautiously entering the venture investment market for the first time.

“A few years ago, corporate VCs were interested in accumulating more and more portfolio companies,” says David Ai, vice president of corporate venture capital for Hitachi. Today, “they’re finding ways to get rid of them.”

You can’t blame corporations that want to avoid venture capital altogether. While most dot-com histories focus on the startups that went boom and bust, many leave out the price that large corporations paid. The post-boom implosion hit corporate venture funds—which invest directly in startups and as limited partners in traditional venture capital firms on behalf of corporations like Intel—much harder than the VC industry as a whole. Corning, for instance, wrote down $47 million in venture capital losses before selling off its venture arm.

Last year, 166 corporations invested roughly $1.3 billion in startups, compared to approximately $17 billion invested by 472 corporate venture programs in 2000, according to the MoneyTree Survey conducted by PriceWaterhouseCoopers, Thomson Venture Economics, and the National Venture Capital Association. In 2004, only 2 percent of venture capital dollars invested came from corporations, versus 7 percent five years ago, according to VentureOne.

Competition among corporate ventures has diminished along with the stratospheric valuations of the boom, so for those corporations still in the market, there are plenty of opportunities to place capital.

The main reason corporate VCs are changing their strategies, of course, is because previous efforts have failed miserably. Many slow-witted corporate ventures flooded the market at the tail end of the 1990s, too late to make a profit. They made the critical errors of focusing attention only on strategic investments and concentrating on the technology instead of the financials of portfolio companies.

Many of these firms brought in people with no expertise in venture capital to vet startups, price term sheets, and manage emerging companies. And they didn’t bother to cultivate strong relationships with Silicon Valley venture capitalists, the backbone of any good deal flow.

In all, more than 90 percent of corporate VCs have done a terrible job with their investments, says Tim Rohner, author of The Venture Imperative: A New Model for Corporate Innovation, and managing director of the Bell Mason Group, based in Palo Alto, California. But things are starting to change.

No Risk, High Reward

IBM has an interesting approach to investing in startups: the company does have a robust venture group with about 30 employees, but few of them make actual investments. It hasn’t invested in a startup in years, and has greatly reduced its investments as a limited partner in venture funds. But that hasn’t prevented Big Blue from forging strong partnerships with promising startups as well as Sand Hill Road venture funds.

The company wants to play, and profit from, the venture capital game—it just doesn’t want to commit much capital. “In terms of gaining entrée to the venture community, being able to get closer to firms turns out not to be about the money—it’s about the relationships,” says Deborah Magid, a director of strategic alliances for IBM’s venture capital group.

Ms. Magid and her colleagues visit with venture capital firms and their portfolio companies regularly. When they see an opportunity, they call in the business development team, which grants a startup access to IBM’s industry expertise and wide breadth of partnership channels. If the partnership flourishes down the road, the mergers and acquisitions team steps in, and IBM acquires the company.

The process starts with developing the right relationships with the right VCs, and convincing startups to sign on to IBM’s platforms at an early stage. In the past five years, IBM has increased its stable of venture-funded startup partners from 20 to over 750.

“We want access as early as we can possibly get it,” says Claudia Fan Munce, IBM vice president of corporate strategy and managing director of its VC group.

While there’s no doubt that early access is valuable, there is still no clear method of measuring the ultimate success of a corporate VC program, because strategic returns are tougher to quantify than financial ones. Financial returns from corporate VC investments can simply be compared to the company’s other business units, but strategic returns involve a complex relationship between factors like partnership revenue, market presence, channel partners, and intellectual property.

Part of the payoff for IBM was explained with its late-February announcement that it would be dedicating 1,000 sales specialists to work directly with independent software vendors and channel partners focused on the small- to-mid-sized market. Rather than invest itself, the company would gather its VC friends and “make the introductions,” says Buell Duncan, IBM’s general manager, ISV and developer relations. The goal: an IBM-led high-tech delivery system that capitalizes on thousands of companies worldwide. In building its dream team, Big Blue is starting with its inner circle—those that have registered with IBM.

IBM is not the first company to play down its investment and focus its venture strategy on relationships. Microsoft, for example, has all but shut down outside investments over the last four years. The company learned its venture capital lesson the hard way, taking $9 billion in write-downs during 2001 and 2002 against telecom infrastructure investments made in the late 1990s.

Since then, corporate vice president and .Net strategy guru Dan’l Lewin has led a team that brought Microsoft’s partner strategy into the depths of Silicon Valley via hundreds of meetings with VCs. The earlier Microsoft gets to know these startups, the more next-generation IT applications are written using .Net and other Microsoft platforms. For a company that makes about 95 percent of its revenue through partnerships, getting buddy-buddy with early-stage companies can also yield future dividends.

“One-time returns and risking capital are not natural and not good uses of shareholder capital,” says Mr. Lewin.

Of course, not all corporations have the luxury of taking such a risk-averse approach. Only corporate behemoths like IBM and Microsoft can command the attention of major VCs without paying to play. After all, VCs know that these companies are the ones most likely to help bring in revenues, and eventually, to purchase their portfolio companies.

Two venture community entrepreneur-scholars are collaborating with the National Venture Capital Association to establish a definitive benchmark for corporate venture performance: Edward Roberts, the MIT Sloan School of Management professor who has studied corporate VC for decades, and Val Livada, a frequent MIT visiting lecturer who is the president of Weybridge Partners, a consulting firm specializing in commercialization strategies. If they succeed, corporate venture professionals might finally have objective numbers to bring to their CEO—and their shareholders.

Outsourcing the Finances

Some corporations have realized what Sand Hill veterans knew all along: that they are not professional venture capitalists, and if they want to make the most out of their venture program, they have to hand the reigns over to the professionals.

Soon after forming their venture arm in 1991, Adobe Systems executives realized they didn’t know much about the venture capital world, but they did know their future success would hinge on the relationships they built with the startup community. So they decided to outsource the day-to-day, nitty-gritty venture capital activities to the pros, hiring traditional VC firm Granite Ventures.

Granite currently manages a $100-million fund in which Adobe is the sole limited partner. Adobe shares deal flow with Granite, and the two coordinate on due diligence to get the Adobe perspective on every investment.

“Granite handles the important aspects of company-building when it comes to term sheets, board seats, and overseeing the health of the investments,” says Fred Mitchell, vice president of venture development at Adobe. “Our focus is in helping the startups grow strategically.”

Adobe provides startups insight into the company’s thinking and gives them an inside look at Adobe technologies. This helps portfolio companies formulate ways to benefit from the Adobe platform and the company’s position in the marketplace.

To ensure that the knowledge transfer takes place between the startups and the corporate parent, Adobe brings the entrepreneurs face-to-face with its management team every week. They share ideas about how Adobe can add value to the startup, and vice versa.

For instance, Granite/Adobe recently invested in NetCell, a company creating a systems-level solution for high-definition (HD) video. This is a perfect fit for Adobe’s video editing product Adobe Premier, which is just starting to tackle HD video, says Mr. Mitchell.

“The challenge with every corporate VC program is that we don’t get credit for financial returns in the stock market,” says Mr. Mitchell. “This means we have to focus on the best ways to add value to the corporate parent.”

The success of the Granite/Adobe partnership could serve as a model for other corporations. Already, Granite manages a separate fund for Texas Instruments, while NEC is a majority stakeholder in Convergence Partners, a Sand Hill Road VC firm.

‘It’s Just Like Pruning Your Trees’

Though it may seem as if many corporations are abandoning the VC space altogether, some corporate VCs say they are merely shedding excess pounds. The companies are looking to leave behind bad investments and take a fresh approach with a streamlined, more strategic portfolio.

“It’s just like pruning your trees,” says Mr. Ai of Hitachi.

Dell, which sold its venture portfolio to secondary fund Lake Street Capital, insists it will remain active as a corporate investor. “Dell is not no longer in this space,” says Dell spokesperson Amy King. “We will continue to make investments in companies whose technologies complement our own.”

Ms. King says that Dell sold only some of its companies—those that fell outside the company’s strategic business lines, servers and storage.

Still, the fact that corporations like Dell are jettisoning part of their venture portfolio demonstrates that corporate VCs gorged too heavily on bad investments during the past few years. They are now eager to wipe the slate clean by getting rid of problem companies that eat away valuable time and resources. In a nutshell, they want to focus again on core investments without being weighed down by the folly of the past.

“Some of the largest companies out there are still very active VCs who wanted to free up people’s time,” says Gretchen Knoell, a Lake Street partner and co-founder. Selling off part of their portfolios “had nothing to do with where they were going in a corporate direction. It was just about opening up more time and availability for their staff.”

While some funds are cutting back or getting out, other corporations are entering the venture market for the first time. When Amgen established a new fund in November, it joined biotech rival Genentech, as well as a host of other relatively new big pharma funds, in trying to harness the innovation coming out of the biotech industry.

While some say it may be naïve for inexperienced companies to jump into the venture game, these biotech firms beg to differ. What they lack in venture capital experience, they argue, they more than make up for in scientific expertise—in the form of thousands of top scientists at their disposal who can help them identify and vet the most promising startups.

In the case of Eli Lilly, for example, investments don’t necessarily have to lead to partnerships, collaboration, or even an acquisition. Lilly Ventures may invest in a startup purely because it has an interesting technology that could potentially be very valuable, even if Lilly corporate has no particular use for that technology.

“I think we can learn something just by watching some of these startups develop,” says one managing partner of Lilly Ventures.

Steady Does It

Even in an environment rife with ups and downs, there are still bread-and-butter corporate VCs that are successful simply because they refuse to make major changes to their venture strategy. In stark contrast to the firms that jump in and out of the market every four years depending on how well the economy is doing, Intel Capital’s success can be attributed to one major factor: consistent investment, no matter what.

Intel has invested $4 billion in about 1,000 companies since the early 1990s, maintaining a standard investment pace through thick and thin during the past several years. Moreover, Intel’s executive team—from the CEO down—views Intel Capital as critical to the company’s overall well-being. Clearly, this is not a sideline business susceptible to the chopping block at the first sign of trouble.

Intel Capital has made one significant change over the years: it has expanded its venture investments rather than narrowed them, establishing new venture arms in Israel, India, China, and Russia. International deals made up 40 percent of its investments in 2004, up from just 5 percent in 1998.

“Corporate VC is a very transient lifestyle with people coming and going all the time,” says Mr. Livada of Weybridge Partners. “There’s no continuity in learning and sharing knowledge. When you are Intel and you’ve been at it for more than a decade, that’s pretty good.”

Beyond staying the course, Intel has found the right companies by taking a hard look at both the strategic and economic benefits of every potential investment.

“A lot of corporations have gotten into trouble looking only through the strategic lens,” says Keith Larsen, director of strategic investments at Intel Capital. “Though their investments may be strategic, they may not be financially attractive. If a startup goes out of business, was it ever strategic?”

Of course, many corporations are now asking the very same question about their internal venture units—and they don’t like the answer. Still, for those bold enough to stay in the venture business, big rewards may be lurking around the corner. The past can’t be undone, but the smartest firms have learned from their mistakes and are more confident than ever that they can improve their performance.