In Search of the Perfect Startup

It’s hard enough in this economic climate to start one new company. So why is H. Michael Zadikian, who sold his last startup for $500 million, determined to start four at the same time? An entrepreneurial adventure story, in two parts.

H. Michael Zadikian is searching for the perfect startup.


It’s a journey that started six years ago, taking him across countries and industries, from Silicon Valley to Texas. At that time, he was a restless young marketing manager for Cisco Systems Inc., convinced that he could do more than simply push out someone else’s products. Later, in 1999, his first attempt to colaunch a high-tech company, Monterey Networks, was sold to Cisco for $500 million — but it didn’t combine engineering genius, teamwork, and strategic mastery as well as Zadikian wanted.

During the height of the startup boom, he sold. Now in the depths of the bust, he wants to create the mother of all startups.

This time, Zadikian has retooled everything that he didn’t like about his first attempt. He won’t be a small niche player anymore — at the mercy of giants that can make or break his business. He will spend more time knitting together teams that genuinely work well together, so that factions and feuding don’t undermine his efforts. And if achieving his goal means starting a flurry of interrelated companies all at once, he’s prepared to do it. And well, that’s exactly the path he’s chosen. Think of it as a startup times four.

His new venture is called the Iris Group, and it consists of four optical-networking companies carefully constructed to tackle adjoining business challenges. Step-by-step, Zadikian wants to help telecom carriers manage high-speed data and voice traffic across every aspect of their optical networks. That is a huge problem, attracting attention from hundreds of young companies as well as industry giants like Nortel Networks and Cisco. Iris is still in its infancy and is less than a year away from delivering its first product to customers. But Zadikian and some powerful venture-capital backers think that Iris’s unusual four-in-one approach could pay off in a big way as the industry evolves.

“Our goal is to partner with organizations looking to change the way networks are built,” says Todd Brooks, an optics specialist and a general partner at Mayfield, Iris’s largest financial backer. “Making networks more efficient requires a bolder way of thinking about design. And that’s what Iris has done. It’s a nimble startup with the breadth of a big company.” Three of the four Iris companies garnered between $24 million and $30 million in venture financing from an investing syndicate with strong credentials in the optical field.

Each company within the Iris Group tackles a discrete piece of the bandwidth challenge. For example, one company focuses on the immediate problems of local carriers overwhelmed by voice and data traffic. Another tackles the long-distance hurdle of sending data across thousands of miles without interruption or costly conversions.


Even in boom times, making all the pieces fit together would take a mighty effort. Market dynamics are changing so fast in telecom that what looks like an effective strategy one quarter can be out of step just six months later. And lately, the optical-network business has gone from frenzied growth to harsh retrenchment. Industry leaders like Lucent, Nortel Networks, and JDS Uniphase have announced layoffs and profit shortfalls in response to slowing demand. And Cisco essentially shut down its optical-switching division, the very business it acquired from Monterey Networks.

But Zadikian is hardly deterred. His companies create products for next year, not next quarter. Eventually demand will bounce back, he maintains, and an industry-wide slump for the next few quarters will actually make it easier for the well-funded Iris startups to refine their next-generation products behind closed doors — to a degree that most other young companies can’t.

Zadikian has already recruited dozens of top-notch hardware engineers and seasoned executives for each of the Iris companies. Product-planning teams from different companies meet regularly with the network management teams. For some prospective customers, all the Iris companies present their solutions together. Zadikian keeps the marketing messages consistent by staying involved with each of the companies’ marketing efforts.

“We’re not on a wild innovation kick, breaking the laws of physics,” Zadikian explains. “We’re trying to balance entrepreneurial methods with the value of big-company experience.”

How to Reverse-Engineer a Big Company

It’s 12:05 AM Thursday morning. Zadikian and his colleagues, sluggish from choice wine and steak at one of Dallas’s best restaurants, climb into a Texas-sized Ford Excursion. But the workday isn’t over. Zadikian needs to finish a conference call before 1 AM. Cruising along Highway 75, he picks up one of his three cell phones and begins a conference call about hiring with the CEO of Coree, the Iris core packet-switching systems company in New Jersey.

Zadikian wears many hats — marketing vice president, chairman, interim CEO for a while. Midnight conference calls are more routine than extreme. New projects arise so quickly that he can’t always keep them straight. “It’s hard to get out of an architecture discussion and into a conversation with a controller candidate,” he says.


But that’s Zadikian’s style. “I’m interruption-driven. The adrenaline rush makes it tolerable,” he explains. His email inbox overflows with 400-plus messages, not to mention the trash-cum-archive with 11,686 items. He seldom takes notes — because going noteless and paperless is a way to stay focused on the larger mission, clearing away the clutter of too many demands that can derail a startup.

That’s how the companies are run. Think of Iris as a way to reverse-engineer a big company. Trim the overhead and distractions. Build deep solutions. Serve different kinds of customers. And do it all as a small, nimble enterprise.

Zadikian, 38, is an Armenian immigrant who moved from Syria to California with his family when he was 16. He entered high school fluent in Arabic but barely getting by in English. After two years of learning English as a second language, he knew that computing was the hot field to be in. He argued his way into the University of California at Los Angeles, persuading an administrator that his smarts and determination outweighed what his lackluster SAT scores showed. He was right — he graduated magna cum laude and went on to do his master’s in computer science there as well.

He worked his way through college by writing front-end software for the Sabre airline-reservations system. He taught himself on the job, a habit he continued when he started his companies. At Cisco, he taught himself marketing as the company shifted its focus from universities to corporate customers. He quickly rose to go toe-to-toe against IBM’s router business, the success of which eroded the foundation of Big Blue’s market.

Eager to try his hand at a startup, Zadikian became vice president of marketing in 1995 at SourceCom, an early player in the effort to deliver broadband-type access. But naïveté about what carriers wanted, and SourceCom’s unwillingness to sell to companies that knew carriers better, ultimately prompted SourceCom to fold.

Unwittingly, he took another big jump in 1997, at the Goldman Sachs investment conference in Sundance, Utah. At the resort, he struck up a conversation with Patrick Nettles, president and cofounder of Ciena Corp., one of the first major optical networking companies out of the gate. The conversation introduced Zadikian to the world of optics, and he was hooked. At first, he read arcane documents, such as Bell Corp.’s “Standards and Specifications” and “SONET: Protocol Reference Guide.”


Then he and his SourceCom colleague Zareh Baghdasarian started Monterey Networks — literally — in a Motel 6 in Monterey, California. They wired their hotel room for an Ethernet connection and charged their laptop computers on the lighter in a moving van. A few short circuits later, they moved their families to Richardson, Texas, in a hot optical corridor.

Monterey Networks grew rapidly, specializing in managing bandwidth in the optical spectrum. Its key product, the wavelength router, was a “hot” box that conducted information in the world of color and light — looking a lot like a 7-foot-tall refrigerator with the entrails sticking out. But with a limited product line, it was destined to become part of a bigger company.

Just two weeks after Monterey Networks was sold to Cisco, Zadikian was restless again. He invited Mayfield’s Brooks to his home in a Los Angeles suburb. Watching roller skaters whiz by, the two men sipped Starbucks coffee and talked about a way to create multiple, independent companies in the optical-networking industry.

In many ways, Iris is built on personal faith. Brooks bet on Zadikian’s potential to anticipate the next market, not because Zadikian was the best technologist or CEO. Now nearly half of Brooks’s portfolio companies are Iris companies, a huge investment of time and commitment to Iris’s success. Iris’s first employee, ex-Sprint technologist Rob Butler, joined as chief network architect before the company was even incorporated. He accepted the job in a hotel bar (Iris had no office) with a personal check from Zadikian for his relocation expenses.

Such an investment of faith meant that the company had to have the right structure.

When Brooks met with Zadikian, the two decided that Iris should be more than an R&D outfit, an incubator, or the forced relationship of a keiretsu. It should be an ecosystem of at least four independent entities that own stakes in one another. True to that philosophy, each Iris company has its own office and is funded and run independently. The idea is that customers get more value if all the companies succeed — and if the products work seamlessly with each other. But if one ship sinks, the armada sails on. “We didn’t want business dependency,” Zadikian says. “The key criterion was that wherever you made the cut, the remaining picture had to be meaningful.”