Valley Wonk: The Indie Startup
Sith was a sure thing, thanks to George Lucas's hordes of rabid adoring fans. But not every big-budget film gets that kind of guarantee. Some don't find an audience, and others just plain stink. Smacks of the venture capital world, doesn't it?
Denny Miu was involved in a Sith-like project once and watched it turn into Waterworld. Miu's name came to Light Reading in the glory days of 2000 with Integrated Micromachines Inc. (IMMI), a player in the all-optical switch game. These were the companies germinated by the supposedly infinite need for more bandwidth, and egged on by Nortel Networks Ltd.'s (NYSE/Toronto: NT) $3 billion buyout of Xros Inc. (see Switch Startup Raises MEMS Questions and Nortel Buys a Monster Crossconnect).
IMMI shrank, then shrank its acronym to IMI, then became Touchdown Technologies Inc. (see IMMI Switches Strategy). But the IMMI that Miu built to 100 employees in one year is long gone. Now, Miu is fronting a venture called Gigamon Systems LLC, and he's thinking small: Six guys, all of them founders, none of them taking salary. One angel investor. No big budgets. Call it the indie startup.
"The VC world has changed. It's become Hollywood, if you think about it," Miu says. "Once upon a time, a few guys had careers making movies. Then Hollywood came along, and Hollywood has a recipe. It's actually very difficult for Hollywood to make small movies."
Silicon Valley venture capital is the same way. "You can't go to a VC and ask for a million. You can barely ask for $3 million. What they really want to hear is $8 million or $10 million," Miu says. "I was a struggling entrepreneur before 2000, so I say it's OK. You don't have to make a Production to make movies."
So Miu created Gigamon without much more than pencil and paper. He didn't want to do another all-optical startup but saw potential in one of IMMI's markets – test, where automated patch panels were needed to connect test equipment to network equipment. Slowly, Miu's crew drew up the idea for Gigamon's electrical crossconnect, one that can link any number of inputs to one output (or vice-versa), as opposed to the one-to-one linkups most crossconnects provide.
The angel investor ponied up less than $500,000 to get the company running, and beyond that, it's been a matter of friendships and connections. Miu didn't even know his five co-founders when Gigamon started; they got recruited one by one as he asked around about developing and marketing his product.
This is how companies used to get built, and it's something the industry forgot during the bubble, Miu says. "In 2000, it didn't make sense to bootstrap anything. There was so much money – why would you start in a garage?"
Gigamon is apparently onto something, because it was the only vendor that participated in eNet, the Interop show network, without paying. As Miu tells it, the network's dry run ran into problems as test boxes fought for access to certain parts of the network. One of the engineers happened to have seen Gigamon's box and convinced show organizers to give the no-name startup a free ticket into eNet. It was like winning on The Apprentice without enduring weeks of working with loudmouthed yuppies.
It was a lucky break, but it exemplifies the way Gigamon was built. Chassis building, software design – all those pricey necessities are coming to Gigamon for free or through deferred payments, again because of the founders' connections.
"I told the other five [founders], 'You're different from me. You have a lot of connections in the industry. And it's your first startup, so you still have friends!' "
Bootstrap operations have gotten popular in the wake of the bubble. Founders of Kamelian have bought back their intellectual property and are having a go as Amphotonix Ltd., for example (see Kamelian Reborn). And even the startups that find VC money are being told to keep it lean. Small spending. An economical product. Development in China or India (even Gigamon has tapped programmers from Asia).
And those big-budget flops? It's not as if the all-optical vendors were in the dark. They were getting told – by customers, not VCs – that these giant 1,000-node switches were vital, that any network without them was a Yugo with square tires. In fact, 1000x1000 – the size so many skeptics were laughing at – wasn't enough for some carriers. "They were saying they needed 10,000," says Rajiv Ramaswami, an Xros survivor who's now a Cisco Systems Inc. (Nasdaq: CSCO) VP.
So while Miu got worn down by watching IMMI disintegrate, he has no regrets about building it up.
"I think I did the right thing. I identified an opportunity and brought in money, built a team, built a strategic relationship with a customer, Cisco. And I was even working with my customer's customer," he says, to make sure the market was real.
The bubble turned tech into an industry of Hollywood yes-men, everyone telling everyone else that everything was OK while the budget spiraled out of control. Miu isn't the only one trying to rediscover the startup process, and the advantage of doing it in leaner times is that it's easier to tell which ideas are good.
"What customers don't have today is time. If you can get them to sit down and talk to you about their problems, it's real."
— Craig Matsumoto, Senior Editor, Light Reading