In late 2003, TuVox board members met to consider whether they were thinking too small. The two-year-old speech software startup was responding to a challenge by a potential board member, an executive at a major enterprise software company. The company's goal of $100 million in revenue was just too modest, she told founders Steve Pollock and Ashok Khosla. She believed that TuVox, then based in Los Altos, California, had the potential to be a billion-dollar company.

That back-handed compliment turned what should have been a routine, upbeat meeting into a contentious three-hour discussion about the company's future. TuVox, founded just before the dot-com crash, was doing well. CEO Larry Miller had raised $6 million in a third-round funding and boosted quarterly revenues past the million-dollar mark, partly on the strength of a huge deal with MCI. The company was on track to meet its goal of $100 million in annual revenue within a few years. Yet the board members were beginning to think that goal might not be ambitious enough.

When a venture capitalist makes an early-stage investment, he's swinging for the fences, hoping that the portfolio company will mature into a billion-dollar exit. After a few years, the companies that gain market traction reach a moment of reckoning: the VCs and the company executives must decide whether to be content with modest success, or to make another bet that shoots for the moon. This is the story of how one company responded when that moment of truth came much sooner than expected.

Twenty Years Ago
TuVox' roots go back to the late 1980s, when Mr. Pollock and Mr. Khosla worked together at Claris, Apple Computer's software subsidiary. Mr. Khosla was the engineer behind MacDraw and MacPaint; Mr. Pollock was vice president of worldwide marketing. They struck up an unlikely personal and professional friendship, and pledged to found a company together someday. “We always said we'd love to do something together if the timing was right,” says Mr. Pollock.

After Claris, they went their separate ways, but both eventually achieved success at companies with a focus on speech and language technology. Mr. Khosla was vice president of engineering at Brightmail, which used a linguistic approach to anti-spam and antivirus software; the company was acquired by Symantec last May for $370 million. Mr. Pollock led the marketing team at call center speech recognition technology company Edify, which went public in 1996 before being sold to S1 in 1999 for $345 million.

The pair took a trip to Europe during summer 2000, just months after the Nasdaq had peaked. They met with friends and discussed what was “hot” in technology. As the journey progressed, their talks morphed from exploratory conversations to serious discussions about founding an actual company. They had focused on the emerging speech stack–a group of technologies being built around using voice to access and distribute data. Voice recognition technology had finally matured, and was poised to enhance and often replace interactive voice response (IVR), the old touch-tone system. Instead of pressing “3” to find out if a payment was due when calling their cable company, callers would soon be able to say “three,” or simply, “I want to see if my payment is due.”

The voice stack's standards were opening up, and existing voice players like IBM were slow to embrace several promising opportunities. Large and small companies were already creating applications to deal with voice calls, but development was complex, taking as much as a year. Mr. Pollock and Mr. Khosla decided to build a software package to help manage that process and cut time out of creating voice-related software.

By September 2000, they were testing the idea out on VC firms around Silicon Valley. “The market conditions were tougher then, [but] we got very positive feedback,” says Mr. Pollock. Before anyone would provide funding, though, TuVox needed to complete a laundry list of demanding tasks: create a prototype, develop a full-fledged business plan, find potential customers, and hire employees. The founders and their friends and families committed about $500,000 in a boot-strapped seed round, and then created the company in Mr. Khosla's Palo Alto, California, front room in early 2001.

VCs Take Note
In early 2000, just before Mr. Pollock and Mr. Khosla began pondering their startup, Paul Holland, a senior vice president of sales at Kana Communications, an enterprise relationship management company, bought a house on Harker Avenue in Palo Alto. The street was already home to a dozen or so current and former high-tech executives, from the CEO of IDEO to the former CEO of Oak Semiconductor. But fate put Mr. Holland, whose company had just completed its IPO, next door to Mr. Khosla.

By the time TuVox was founded, Mr. Holland had quit Kana and become “the neighborhood guy,” spending hours with his young daughters and with Mr. Khosla. The two talked across the fence between their properties, or in Mr. Khosla's front yard, and their families often had dinner at each others' homes. Mr. Khosla soon told Mr. Holland that he was leaving Brightmail to start a voice software company. “I had [tracked] speech recognition at SRI. I thought that voice was never going to happen,” Mr. Holland remembers.

In fall 2001, the TuVox founders went on the prowl for VC financing. “We said, ‘We're way under-funded. Who wants to be in on the deal?’” Mr. Pollock recalls. Around the same time, Mr. Holland was itching to get back to work, and in October 2001 he joined Foundation Capital. On his first Wednesday at the office, he walked into the conference room, and up on the market board were the remnants of a discussion about TuVox. Someone had referred the company to Foundation co-founder Bill Elmore, and the firm's partners had a 30-second prescreen chat about the company. Though they were intrigued, Foundation ultimately passed on investing in TuVox because Mr. Holland felt the company was too young for an investment.

After meeting with Granite Ventures managing director Tom Furlong in the spring, the TuVox founders began a long courtship with the firm. “It was a time when people weren't competing heavily for deals,” says Mr. Furlong. “That was the nuclear winter, so I had the luxury of just talking to them for six months.” The fundraising round lasted so long that Mr. Holland had another chance to invest in early 2002, but the company “still felt a bit raw.”

Mr. Pollock continued to scour his rolodex, making calls to old colleagues and friends. One call went to Adobe CEO Bruce Chizen, who he and Mr. Khosla had worked with in their Claris days. Adobe's venture funds happened to be managed by Granite. Mr. Chizen saw potential synergies between his company and TuVox, and his team assisted with due diligence that helped push the Granite deal over the line. In April 2002, Adobe and Granite split a $5-million Series B investment round in TuVox, just after the company launched its beta product.

Mr. Pollock had stayed in touch with Jeff Crowe, his former CEO at Edify, keeping him in the loop on TuVox. “I was intrigued by what they were doing,” says Mr. Crowe, “especially because it sounded very technologically challenging. I said, ‘Why don't you guys build the software, get a couple of customers, and get some funding? If you do all that, I'll join the board as an outsider.' And by gosh, they did it.” TuVox signed up gaming company Activision, and Mr. Crowe took a seat on TuVox' board in September 2002.

As with many early TuVox deals, the company sold Activision the tools to create voice applications for a technical support line, then helped its customer create the application. Over the next several months the business model evolved, as TuVox began to sell its own pre-built basic applications, along with tools customers could use to modify those applications.

In the first week of 2003, Mr. Khosla updated Mr. Holland. “I don't want to step in a pothole,” Mr. Holland told him. “If your [customer] pipeline is really firming, I'd like to help. Call me when you think you are ready to come back and present to us.” As the company grew, it became obvious that TuVox needed to hire an outside executive as CEO. In May, the founders hired Larry Miller away from supply chain software company WorldChain; before becoming president and COO of WorldChain, Mr. Miller had been a sales executive of call center company Aspect Telecommunications. During his first month on the job, Mr. Miller took a meeting with Mr. Holland in the TuVox conference room.

“Tell me why you think I should invest in TuVox,” said Mr. Holland.

Mr. Miller's response was honest: “Right now, I couldn't tell you. It's a company with a great team and a great product that should have a lot more customers. Give me about 60 to 90 days and I'll let you know.”

“We're likely to fund it,” said Mr. Holland. “When you think you have this thing figured out, call me. If you believe in your pipeline, then I am likely to believe, too.”

Mr. Miller spent the next several months getting to know his new company inside and out, and soon landed the large MCI deal. Sure enough, four months after their first meeting, he called Mr. Holland, ready to pitch for Foundation's capital. After tracking the company for almost two years, Mr. Holland was keen to invest, but his wife was slightly nervous.

“Are you sure you want to fund your next-door neighbor's company?” she asked him.

“I'd fund half the people I invested in if they lived next door,” he told her. “And besides, if anything bad happens, it's Larry and me, not Ashok and me.”

In November, Foundation led TuVox' $6-million Series C round, and Mr. Holland took a seat on TuVox' board.

Back to the Drawing Board
The next month, the board assembled for its monthly meeting: Mr. Holland, Granite's Mr. Furlong, CEO Mr. Miller, Mr. Pollock, and Mr. Crowe. After the MCI deal and its recent investment round, the company found itself at a crossroads. “There's that fragile period in the beginning where you'll be happy to make the numbers for that quarter,” says Mr. Furlong. “Once that happens predictably, if you've got a good management team, they lift their heads up and say, “Where can we take this thing?”

As Mr. Miller updated the rest of the board on the quarterly numbers, the ambitious predictions from the potential board member (she ended up not joining, so the company wouldn't disclose her name) were at the top of his mind. “We've heard from a number of experienced industry contacts that we should be thinking about TuVox beyond the $50-100 million long-term revenue objective,” he said. “We should be thinking in terms of $500 million to $1 billion and higher.”

“Do we believe that we should adjust our targets higher for the company?” asked Mr. Miller. The board members were all confident about the company's potential to hit $100 million, but it took some time for them to reach consensus that TuVox should aim higher. “We were all pretty consistent in our thinking,” says Mr. Crowe. “But you're always balancing how you are doing currently versus what trajectory you're on, how much money that is going to take, and how much that will dilute your investment.”

“It's better to aim high and come slightly under than to shoot for a much lower target and hit it,” Mr. Holland told his fellow directors amid the discussion. After they arrived at an agreement, Mr. Miller asked them the second key question: “What would we be doing differently today to prepare for that outcome?” Each board member had his own ideas, but there was one consensus recommendation: increase the sales and marketing budget for fiscal 2004, particularly overseas. “We had no marketing department in 2003. Steve was handling most of the work in his ‘spare time,’” says Mr. Miller. “We had less than $30,000, to participate in one trade show.” The company's international presence was essentially zilch.

Mr. Pollock hired Terry Shough, another Edify veteran, as TuVox' senior vice president of worldwide sales, and Mr. Miller set plans in motion to expand outside the United States. Yet all this expansion meant a higher burn rate for the company. By August, the board members realized what increased spending entailed: another round of VC funding.

Although Mr. Holland was a tech industry veteran, as a venture capitalist, he was still pretty green. He asked Foundation co-founder Bill Elmore for advice. “If I were you,” Mr. Elmore told him, “I would consider raising a very large round and go for a game-over strategy.” So by the August 2004 board meeting, Mr. Holland was ready to persuade anyone who needed convincing. Mr. Miller, it turns out, was concerned that a full-blown road show during a weak fundraising environment would distract from the company's momentum.

“Oh, come on, you guys'll have fun,” Mr. Holland told him. “This is going to be very different than your last round [raised just after the tech industry bust], which was a bit of a death march.”

“I've never associated the word ‘fun’ with raising venture capital,” said Mr. Miller.

In order to take the pressure off Mr. Miller and his team, the VCs proposed a scaled-down round, where he'd only have to present to a select few firms. Mr. Holland, Mr. Furlong, and Mr. Crowe, who had recently joined Norwest Venture Partners, each chose two firms to pitch. “We wanted to pick a small group of firms to show it to. The downside of that is, if you don't get your story right the first time, you have spoiled the opportunity,” says Mr. Furlong. To prepare, Mr. Miller and three executives honed the two-hour pitch by practicing in front of Granite's partners. With a few phone calls, the directors scheduled a whirlwind of meetings with East and West Coast VC firms spread over a Tuesday, Thursday, Friday, and following Monday in early September.

Because Mr. Crowe was already on the board, TuVox gave Norwest first crack at the deal. Though the firm's due diligence process is normally as tough as most VCs', Mr. Crowe had been with TuVox for two years, and knew the company as well as anyone. As soon as the pitch was over, one of Norwest's partners asked Mr. Miller, “Is 24 hours quick enough?” Mr. Miller laughed, but the term sheet was delivered the following evening.

The board members conferenced by phone at 11:30 p.m. that night–all except Mr. Crowe, who recused himself from the discussion. “We as a company hesitated. There were some deal points we pushed back on. They responded with a higher valuation and fixed most of those deal points,” says Mr. Furlong. On Thursday, less than 36 hours after Mr. Miller had first opened PowerPoint to give the pitch, TuVox closed a $15-million Series D round.

Before they could celebrate, though, Mr. Holland and Mr. Furlong had to make some painful phone calls–canceling the remaining VC meetings, and effectively shutting friends out of a sweet deal they'd hyped just days before.

Post-Deal Progress
TuVox has put the additional sales and market money to good use–the company's revenue for fiscal 2004 (which ended just after the Series D) was eight times its revenue during 2003.

Since then, TuVox has launched a customer program in Europe and is on the verge of closing a deal with its first Asia-Pacific customer, a major telecommunications company. In the U.S., the company signed a deal with TV Guide to help build software allowing the magazine's subscribers to use speech to change their name and address. TuVox has also signed up several healthcare clients, including Prescription Solutions and AvMed Health Plans, to create applications giving customers voice access to healthcare and prescription data.

There are no independent assessments of the market for conversational voice applications, but TuVox and similar companies are going after a piece of the multibillion-dollar market served by traditional call centers via live operators and touch-tone dialing. New York City-based Audium, a direct competitor, launched its first VoiceXML application before the TuVox concept was even hatched. It used this head-start to sign up large enterprise clients including GlaxoSmithKline, Coca-Cola, and two of the top five U.S. retail banks. As TuVox expands into Europe, one major competitor will be Germany's VoiceObjects, whose clients include its home country's largest mobile phone and landline providers, T-Moble and T-Com, as well as Postbank, Germany's largest bank.

To fend off competition, TuVox has taken recent steps toward partnerships with major voice-related companies such as Genesys, the call center software unit of communications company Alcatel, and tech heavyweights that play in the voice space, including IBM. Mr. Miller expects the company to be profitable during 2006, but the surest sign that TuVox is on its way up is that it replaced its sparse Los Altos offices with new digs for its 75 employees in Cupertino, just down the road from where the founders first met.