Data Scientists Guide to Making Money from Start-ups

How should data scientists think about starting or joining a start-up? We summarize the advice from a high-powered KDD-2013 panel of leading data scientists/enterpreneurs who share their start-up experience.

By Gregory Piatetsky, Aug 15, 2013.

Data Scientists are very much in demand, as KDnuggets readers certainly know, but making a jump from a scientist to an enterpreneur requires a rare combination of both analytics and enterpreneurial skills.

One of the more exciting events at KDD-2013 Conference on Knowledge Discovery and Data Mining, just concluded in Chicago, was a panel

Start-up“A Data Scientist’s Guide to Making Money from Start-ups”.

Panelists included 4 leading scientists/enterpreneurs who have shared their start-up experience.

The panel was moderated by leading data science researchers Foster Provost (NYU) and Geoff Webb (Monash U.), who also have experience launching a start-up.

Here are my (edited) @kdnuggets tweets and short notes from this panel – thanks to Usama Fayyad for comments on an earlier version.

Usama Fayyad:

  • Why I left Microsoft to launch my first startup: curiosity, see the impact of your work, unlimited upside
  • I have a list of mistakes to avoid when doing start-up, but with every start-up, we keep repeating the same mistakes

Oren Etzioni:

  • why I launched 1st start-up: to make an impact. Academia is like bridge – playing for small stakes and by the rules, startups are like poker – playing for big stakes, and with uncertainty
  • the only risk is not doing a start-up, not taking a chance. Start-up is not a risk, it is adventure

Claudia Perlich: Why I joined a start-up: the energy level, the fast pace of start-ups

Usama Fayyad:

  • how much can a data scientist earn from creating a start-up? Perhaps 10M in 5 years
  • how do I decide if a data science start-up is good: It is not the idea, it is the TEAM, and most important feature of team is Execution, as ideas will change many times.
  • my criteria as angel investor: Can this idea + team start-up double evaluation in 6 month ?

Oren Etzioni: learn from your first start-up, then immediately go to your 2nd start-up

Ron Bekkerman said that an average return from start-up was $1 million, but Usama Fayyad thought that for Data Science/Data Mining start-ups the average is much higher, perhaps $5-10 millions, as this is a new and hot area with strong demand.

Ron Bekkerman had several points on investing and expected returns:

  • the VC criteria – is what you are building nice to have vs must have?
  • VCs invest in technology, but not really because they care about technology – they really care about making money.
  • for early stage investment, VC is interested in only one question: are you going to be $1 Billion company?
  • A good VC wants a 10% chance of getting a $1 Billion company

Claudia Perlich: the most rewarding part of a start-up is being at the core of the company

Question: How can a start-up compete for talent with established companies?

Usama Fayyad:

  • How to recruit for a start-up: with passion, big vision, and excitement
  • Location does matter – in my last year at Yahoo, we did not hire any engineer from Bay area without out my approval. They were overpaid and very hard to retain.
  • there is little career risk jumping to a startup, as demand is too high and finding a replacement job in case of failure in startup is not problematic.

Someone on the panel said: If you want to pursue a PhD, the burden of proof is on you. You should also want to be a professor .

Ron Bekkerman: the critical success factors for a startup (not just data science) are

  • the size of the addressable market
  • goto market strategy
  • pricing
  • Most importantly, execution
  • the most important person in the start-up is the CEO

See also a write-up on this panel:
6 tips for turning that data science education into startup riches, by Derrick Harris, GigaOm, Aug 14, 2013.

  • Reconsider that Ph.D., says Oren Etzioni
  • Do your first startup, then immediately start your second startup (Etzioni)
  • Do your financial math – compare the expected earning from staying at your jobs and potential from a new start (Bekkerman)
  • Be a painkiller, not a vitamin (Fayyad)
  • Single founder: bad – need a team of founders
  • Get ready to get stuff done