The HR startup has an HR problem

David Sacks walked into Dolores Park Café in San Francisco to talk to Lars Dalgaard, a venture capitalist, about what he should do with his life. Or rather, his money. Sacks, who is 43 and has thick gray hair and blue, protruding eyes, made his first fortune as an early executive at Paypal, then a second as the co-founder of Yammer, a social network for businesses, which he sold to Microsoft in 2012 for $1.2 billion. He played in poker tournaments, produced the film Thank You for Smoking, and became an early investor in Uber and Spacex. But by the fall of 2014, he was sick of jumping from hobby to hobby. He wanted in on a startup again.

Sacks and Dalgaard were business acquaintances—dalgaard, a general partner at Andreessen Horowitz, had once tried to buy Yammer for $300 million. After the requisite industry gossip, they got down to business. Dalgaard urged Sacks to take a look at Zenefits, a new company in which Andreessen Horowitz had recently invested.

Zenefits makes online software that automates health insurance, payroll, and other essential office drudgery—kind of a human resources version of Turbotax. It’s not a sexy idea, but with 6 million small businesses in the U.S., it’s enormously useful. The company was founded in 2013 by Parker Conrad, who realized he could streamline small businesses’ managerial needs, saving them hundreds of hours of mind-numbing paperwork—not to mention the cost of staffing an HR department—by putting everything online. Conrad was known to be a little frenzied and disorganized but fiercely intelligent. “From an investment philosophy … we look for the magnitude of the genius, as opposed to the lack of issues,” says Andreessen’s founding partner Ben Horowitz. “And in a way, [Conrad] was like the prototype.” Conrad had no background in health insurance but quickly learned the intricacies of the business as well as any veteran. “If you’re an insurance broker,” he said at the Techcrunch Disrupt conference in 2013, “we’re going to drink your milkshake.”

As Sacks saw it, the most compelling part of Zenefits was that it gave its software away, making most of its money through brokerage commissions that insurance companies paid when a client bought one of their plans. Those commissions recurred annually: Once a business signed up, Zenefits would profit for years to come. “It’s a glaring omission that no one had created [something like Zenefits] before,” Sacks says. He and Dalgaard talked at the cafe for three hours, stopping only when the waiters started putting chairs on the tables. A few days later, Sacks met with Conrad. He joined the company in December 2014 as chief operating officer.

Zenefits was everything Silicon Valley loved wrapped up in one company. It had a visionary founder. It tackled a stodgy industry ripe for disruption. The recurring commissions gave it a steady stream of revenue from the start. And Zenefits was the first in the health insurance software space, the Uber to its future competitors’ Lyft. The potential for greatness is what allowed Zenefits to expand from 15 employees to 1,600 in three years; raise $580 million in three fundraising rounds; and become one of Andreessen Horowitz’s biggest investments. Last year it was valued at $4.5 billion, which made it, in Valley parlance, a “unicorn” several times over.

“You’re looking for disruptive companies,” Conrad said in that same Techcrunch Disrupt speech. “We are going to mess stuff up.”

He was right, although not in the way he meant.

Conrad is 36, with a round face, a hefty build, and closely cropped, reddish-blond hair usually accompanied by a day’s worth of stubble. He grew up in New York City, the son of an environmental activist and a corporate lawyer. He went to, flunked out of, reenrolled in, and ultimately graduated from Harvard. People who’ve worked with him say he’s prone to emotional highs and lows that are amplified under pressure. In 2011 he and a former college roommate launched the financial management company Sigfig. Conrad’s partner pushed him out after a year.

According to interviews and talks he gave during happier days, Conrad came up with the idea for Zenefits in 2012 while trying to offer health insurance to Sigfig’s few dozen employees. He’d become familiar with the intricacies of the U.S. healthcare system while battling testicular cancer during his early 20s, and he even found navigating Sigfig’s health-care options cumbersome. “It was an area that was a real pain point for me personally,” he told Techcrunch last year.

Before a company in the U.S. offers health insurance to its employees, it has to comb through thousands of plans offered by hundreds of insurers, and then select the two or three options that best suit its employees, based on age, marital status, and location. The smaller the business, the less likely it is to have someone with enough time to sort all that out. “We have 10 employees. We’re not big enough to have an HR department,” says Scott Yates, founder and chief executive officer of Blogmutt, a content service startup based in Boulder, Colo. “We can’t do this on our own.”

Enter the health insurance broker, who helps companies figure out what they need. Brokers are paid by insurance companies each time they sell an insurer’s plan. Sometimes they’re paid a flat rate, sometimes a cut of the plan’s premium. Blogmutt once used a broker. “We had an actual living, breathing human come talk to us, and it was great,” Yates says. “But he wasn’t always available. The promise of Zenefits has a lot of allure for a company like us.” Blogmutt signed up a year and a half ago.

Zenefits launched in April 2013 through startup incubator Y Combinator. Conrad had recruited one of Sigfig’s top engineers, Laks Srini, to be his co-founder, and he expanded Zenefits’ scope to include payroll, 401(k)s, Cobra, offer letters, and other Hr-related chores. Within eight months the startup was on track to make about $1 million in recurring annual revenue and had landed the investment from Andreessen Horowitz. (Srini declined to comment. Bloomberg LP, which owns Bloomberg Businessweek, is an investor in Andreessen Horowitz.)

That gave Zenefits—and Conrad—legitimacy. Andreessen Horowitz is one of the top VC firms in Silicon Valley, the big checkbook behind dozens of success stories, from Airbnb to Skype. Conrad was the type of person the firm was looking for. He could deliver a lecture on the inner workings of the insurance industry as easily as he could tell you what type of company Zenefits could become.

“I’m excited about meeting incredible human beings that change the world that we walk on, literally, and how we walk on it,” Dalgaard says. “I don’t think I’ve ever seen anyone— and I’ve seen a lot of people—think as comprehensively as [Conrad] did about the market he was going after and how to build his product.”

Andreessen Horowitz led two fundraising rounds totaling $82 million for Zenefits in 2014. Dalgaard joined the newly formed board. His job was to help Conrad develop Zenefits into a mature corporation. When Conrad set the company’s 2014 recurring revenue goal at $10 million, Dalgaard doubled it. “Lars sat there in his very Lars fashion and was like, ‘Why are you guys so f---ing bush league?” Conrad said at a software conference last year. Then, in Conrad’s telling at that conference, Dalgaard told him to add at least 100 sales reps to make it happen.

Zenefits started growing. It moved into a highrise in San Francisco’s Soma district and expanded from fewer than 20 people to roughly 500, many on the sales side. As far as Silicon Valley perks go, Zenefits was pretty sparse—no personal chef, no meeting-room-cum-ball-pit. Conrad had the offices painted gray and orange and named each conference room after a character from Star Wars. By the end of 2014, Zenefits hit Dalgaard’s $20 million recurring revenue goal and was making good on its promises to shake up the brokerage business.

According to former employees, Conrad oscillated between a conviction that Zenefits was conquering the world—at the time, about 2,000 businesses used its services—and a near-constant fear that it would never make its numbers and he’d be fired. Employees were afraid that any mistake would stop the company’s growth. At a talk last year organized by Khosla Ventures, Conrad said, “There’s a low-level panic that suffuses the organization, a constant pressure to keep moving faster and faster and faster.”

Soon, Zenefits ran into snags. The Utah Insurance Department banned it from operating in the state because the agency considered the free HR software to be in violation of a law against offering rebates to customers. Zenefits was also finding that some insurance companies weren’t technologically advanced enough to integrate with its system. “Most of these insurance transactions are still done on paper or with e-mailed PDFS,” says Sabrina Corlette, a research professor at Georgetown University’s Health Policy Institute. This increased the odds of errors—and Zenefits’ employees were often moving too fast to catch mistakes. “My dentist said they couldn’t run my insurance claim through, and it took three months to get someone at Zenefits to fix it,” says Brittany Keppel, a former office manager in Brooklyn who also had trouble getting Zenefits to sign her up for Cobra when she was let go from her job in January 2015. Blogmutt’s Yates says he deals with Zenefits problems every few weeks: “Names appear in records incorrectly. Dates are wrong. I don’t know where the breakdown is, but the biggest problem is there’s no quality check.”

Zenefits spokeswoman Jessica Hoffman responds: “Our focus on remediation over the last three months has included a quality initiative that has dramatically reduced error rates.”