Data scientist has been deemed the sexiest job title of the 21st century, and there’s ample evidence to suggest that skills in the field are pretty much a license to print money. The only question is how much.
During a Tuesday afternoon panel at the ACM Knowledge Discovery and Data Mining conference in Chicago, four folks who’ve successfully made the jump from academia to entrepreneurship — and even into venture capital — shared some advice on how to maximize the return on a data science education. And although data scientists might have better prospects than most right now, it’s actually great advice for anybody finishing up a graduate program in any field or sitting pretty with a professorship and wondering what’s next. (And for a few more tips about starting big data companies, check out my recent post: Want to start a big data company? Here are 5 things you need to know.)
1. Reconsider that Ph.D.
If you don’t already have a Ph.D., think twice — or three or four times — about earning it. “Often, it does not make sense to get a Ph.D.” said Decide.com Founder (and former NetBot and Farecast Founder) Oren Etzioni. In fact he tries to persuade his students (Etzioni is also a computer science professor at the University of Washington) that they shouldn’t earn one. Unless you want to be a professor, they’re often just an unnecessary obstacle between grad students and a lot of money.
Etzioni told the horror story (arguably) of one student, a really good programmer and hacker, who he strongly advised against earning a Ph.D. The student did it anyway — although with another adviser and in robotics — and ended up joining Google four years later than he would have. Because he joined post-IPO, he missed out on a lot of money.
Furthermore, people who get Ph.Ds. immediately after finishing their masters degrees might being doing everyone — themselves included — a disservice. “I think 80 to 90 percent of thesis topics are totally useless,” said ChoozOn co-founder, Chairman and CTO (and former Yahoo EVP and Chief Data Officer) Usama Fayyad.
It’s not that the people are stupid, he clarified, just that they don’t really know the real-world problems that really need solving. His advice: work a few years and then if you still want a Ph.D., come back and do one on a topic that will really benefit the research and business worlds.
While someone looking for a first job after earning a Ph.D. definitely looks better than someone with a new masters degree, noted Claudia Perlich, chief scientist at Media6Degrees, someone with a few years of industry experience looks even better.
That said, Perlich spent six years earning her Ph.D. and, she said, “I sure don’t regret it.”
2. Do your first startup, then found your first startup
If you really want to build your own company, Etzioni advised, it’s usually best to work at another startup first. You’ll learn a lot about how startups operate and what things to replicate or avoid that it will making the founding experience that much easier. If that’s not an option, you should at least talk to a lot of people who have launched startups before and glean some tips from them.
3. Do your financial math
Because data scientists are in such high demand, the panel agreed, the only risk involved with leaving a cushy job in academia for one in the startup world is not doing it. It’s not as if you’ll be permanently out of work if the switch doesnt work out.
But, warned Carmel Ventures Chief Data Officer and former LinkedIn data scientist Ron Bekkerman, people making the switch should consider how much they value a sure thing versus a possibly huge reward. Being the founder of a company means a potentially huge payout, but the chances of that happening are pretty slim. On the other hand, taking a job at an established startup — like LinkedIn, with about 400 employees when he joined — means a smaller, but almost guaranteed payout.
Going to work for a small startup as just an employee? Well, that’s usually a blend of a low potential reward and a low probability of it happening.
That’s why negotiation is critical, Etzioni said. Find out what other people in the company are getting in terms of salary and equity, what percentage of the overall shares your stake comprises, and what employees at comparable firms are getting, and then use that as a baseline.
Or, Fayyad suggest, you can just make a bold promise and get it incorporated into your contract. He once had an employee tell him he’d make the company $35 million extra in one year, in exchange for getting to keep $10 million of it. He accepted Fayyad’s offer of $2 million, which is still a pretty sweet payout if you can pull it off.
4. Be a painkiller, not a vitamin
This is all about the importance of having a good idea. A few quotes from the panel drive home the point pretty hard, I think:
- Fayyad, who’s also an angel investor and chairman of Oasis 500, on building something that people need and will pay for: “We invest in companies that make painkillers, not vitamins.”
- Bekkerman on choosing an idea with a large addressable market of buyers: “VCs invest in technology, but they don’t really care about technology. … Are you going to be a billion-dollar company? That’s the only question they’re interested in.”
- Etzioni on the dearth of bad ideas and the importance of seeking feedback to ensure yours isn’t one of them: “You should be so lucky that someone wants to steal your idea.”
5. “Single founder: bad.”
In fact, the more co-founders, the better. “Single founder: bad. Many founders: good,” explained Fayyad succinctly. That’s because founders can often be blind to the reality that their product or business strategy isn’t that good.
“If you can get that team with you,” Fayyad added, “you probably have a good idea.”
Assuming a two-founder situation, technical founders should track down a business-savvy co-founder, and vice versa. And that doesn’t just mean hiring an employee, Fayyad warned: the second person must own a significant share of the company.
However, finding the right partner isn’t always easy, especially for technical founders. Etzioni, who’s also a partner at Madrona Venture Group, suggests speaking to potential investors, who often will know someone who can help round out the team. “Even if [VCs] don’t give you money,” he said, “they can give you helpful connections.”
6. Get ready to get stuff done
Everyone probably knows there’s a bit of a pace difference between startup life and research life, but Perlich — who joined Media6Degress after earning her Ph.D. and working for IBM Research — said it can still be a little jarring. Stuff happens in hours or days instead of weeks or months, and no one asks questions. They just do it.
“I had no idea what I was walking into initially,” she acknowledged.
Feature image courtesy of Shutterstock user Kolett.