The Big Customer

How to devote more time to the client and help it define its needs

By CHERYL WINOKUR MUNK

Everybody knows that putting all your eggs in one basket is typically a strategy to avoid. But for some small companies, it's often the best -- and perhaps the only -- way to get started.

The eggs-in-one-basket route seemed the way to go for one fledgling technology company starting out in 1997. Improv Systems Inc. (http://www.improvsys.com/), Beverly, Mass., had one product -- a largely untested chip-processing technology called Jazz PSA -- and was looking for a large backer to help expand the use of its technology in multimedia products. But its founders didn't plan to stick with having just one big customer. The idea was to woo a few more big customers and grow from there.

Anytime you have a small company, says Cary Ussery, Improv's chief executive, president and co-founder, the best way to get things going is to find one or two lead customers who have multiple projects. "If you don't have a large direct-sales channel yet, you can't really go out after hundreds of customers at a time," Mr. Ussery adds.

What Improv ended up with was a customer that represents 25% to 30% of its revenue and has an 8.5% stake in the start-up: Philips Electronics NV of the Netherlands. In addition to buying into Improv, the electronics giant has engineers from its various divisions team up with Improv engineers to experiment with new chip designs using Jazz PSA.

The strategy of relying heavily on one customer is common among all sorts of start-up companies, for clear reasons. For one thing, a small company finds it easier to handle one or two accounts than hundreds. It also gives the company time to build relationships and develop a reputation in a particular industry. And it allows the company to grow without having to fund the costs of a large sales force.

What's more, if you have one dominant relationship, you can devote more time to the client and help it define its needs. You also have an automatic market for your product. "You're not doing R&D for the product that you then have to market and sell," says Robert Waxman, a certified public accountant and partner with Corporate Finance Advisory, an accounting firm in New York. "That's a big leg up."

But the eggs-in-one-basket strategy also has an obvious drawback: "If your major customer walks away, you have a major problem," says Jeffrey Chazen, tax partner at Eisner LLP, an accounting firm in New York. "If you had one customer and it was Enron or WorldCom, where would you be today?"

Although Improv hasn't had to deal with an Enron or WorldCom, this is a lesson that it, too, has learned. The economic downturn and subsequent cutbacks in mid-2001 at Philips have forced Improv to lay off one-quarter of its staff and search for smaller chip-design customers as well.

The primary risk of depending on a few customers, says Mr. Ussery, is that you are dependent on the fate of those customers for a major portion of your business. "This is particularly true when, like us, you start with a targeted product line and group," he says. "Your fate is tied to their fate; the recent down market in communications equipment significantly impacted our projects with these customers."

From the Beginning

Even before Philips, Improv relied on a single primary customer. The company spent two years developing its technology, and then went after its first big licensing client, Geneva-based STMicroelectronics NV, which signed a licensing agreement in 1999.

STMicroelectronics was an ideal first customer for several reasons. For one, it is a top semiconductor vendor with a broad product line and signficant experience in consumer and communications markets. The company also was active in designing new products in the emerging markets Improv was targeting, namely voice-over-Internet protocol, which is used to transmit voices over the Internet, and streaming media. What's more, STMicroelectronics had strong internal research-and-development teams that understood the potential for Improv's technology and was willing to give early guidance for its development.

For most of 2000, the licensing agreement with STMicro contributed as much as 35% of Improv's revenue, compared with about 10% today, says Mr. Ussery. But eventually STMicro decided not to license as much from Improv, says Mr. Ussery. The company does much of its development in-house now, he says.

"Improv has a very interesting technology that we've been and are continuing to experiment with," says an STMicro spokesman. "None of our current products use their technology, but based on our customers' needs there is the possibility we may deliver it in the future."

Improv's second client was Philips, which licensed Improv's technology for use in networking, telecommunications and other application areas. The relationship quickly became an important one for Improv.

At one point, Philips represented more than 80% of the start-up's revenue. And in November 2000, the companies announced that Philips had become a minority investor, injecting $9 million and taking what was then a 9% stake. (The company has since issued additional stock to new employees, decreasing the percentage of shares outstanding that Philips owns.)

For Philips, Improv's Jazz PSA was an opportunity to buy into a new low-power, low-cost technology that could be applied to a number of multimedia products. When Philips sees a technology that it could apply to its own business, says Robert Payne, vice president and general manager of Philips Semiconductors, the company sees the technology as an investment opportunity as well.

For Improv, the relationship with Philips was an opportunity to expand its resources and get its technology into more consumer products, such as voice-over IP phones and personal digital assistants, faster.

Mr. Ussery says working with a large backer has been an important source of new business and has allowed the company to refine and improve its technology. In addition, he says, having a primary customer makes things simpler: For instance, once one business agreement is signed and the relationship is defined, it dramatically simplifies subsequent legal and business discussions.

Having such a dominant customer has required a lot of work on Improv's part, though. For one thing, keeping such an important relationship means going above and beyond what the customer has paid for to make sure the products are successful. "You support the hell out of them," Mr. Ussery says.

In the case of the chip-design collaboration with Philips, the companies have gone as far as setting up an intranet so their engineers can easily swap ideas on new chips for a variety of products. There have been no product announcements yet, but Mr. Ussery says products are expected to be out next year.

Economic Strain

Tying its future to its relationship with Philips has been nerve-racking for Improv, with the economic downturn putting a strain on the two companies' relationship. Philips had selected Improv's technology for five different projects within its U.S. telecom division, Mr. Ussery says. But when the market completely dropped out on telecom and Philips had major cutbacks, the projects ended up getting cut in mid-2001. Improv, in turn, had to let 16 people go, bringing its staff down to 45.

The loss of some business from Philips reinforced to Improv the importance of diversifying, which it is now trying to do. So the company has sought to work with smaller companies, which during the downturn have been more willing than larger ones to do new design work. Most of these companies are one-product start-ups like Improv, hoping to grow from one big idea. At this point, the small clients can't afford to cancel a chip contract with Improv, because it would mean canceling the development of that one idea, and in turn endanger the company's future.

Big Is Better

Nonetheless, Improv ideally would prefer to replicate the relationship with Philips with several other large companies at once, so it would have a buffer when one isn't doing well. A major reason for targeting larger companies is that they typically have more than one project that needs doing, so one project with a large company often leads to others. Also, larger companies have more internal resources to work with Improv in developing and refining the technology, Mr. Ussery says. "With a small company, it's 'I need this fixed today or I'm going to go out of business,'" he says.

Ideally, Improv would like 75% of its revenue to come from large companies and 25% from small semiconductor companies. Right now about 60% comes from small companies and 40% from large companies.

Improv isn't yet profitable. It had a loss of $3.5 million in both 2000 and 2001, and expects a loss of $7.3 million this year, though Mr. Ussery predicts it will be profitable next year. Revenue grew to $3.5 million last year from $1.7 million in 2000, and Mr. Ussery expects revenue this year to be flat.

--Ms. Winokur Munk is a reporter for Dow Jones Newswires in New York.

Write to Cheryl Winokur Munk at cheryl.munk@dowjones.com

Updated October 28, 2002