Professor Foster Provost is Professor of Data Science at New York University and Professor of Information Systems and Andre Meyer Faculty Fellow at NYU’s Leonard N. Stern School of Business. Professor Provost studies data science and its alignment with business problems. He helped found companies like Dstillery (formerly Media6degrees), Integral Ad Science, and EveryScreen Media (acquired by Media6degrees). He advises other companies on data science and strategy.
Please give a brief introduction of yourself:
I am Faculty to member at Stern. My expertise is in data science and machine learning. I’ve been working recently at mining super fine grain data, largely behavior data, for tasks like advertising. I worked for five years, before joining NYU in 1999, for what’s now Verizon, working on problems like fraud detection and network diagnosis. Related to entrepreneurship, I have co-founded some startups. The most successful are Dstillery and Integral Ad Science. They both are in the online advertising space.
How does your research as a faculty member align with your aspirations as an entrepreneur?
They’re completely aligned. There are a lot of different researchers, and I don’t see one mode to be necessarily better than the other, however there is a mode that I have worked in for 25 years. It’s looking at real world problems that I have been convinced need to be solved. I look at state of the art technologies (mainly data science because that’s my passion) and whether or not problems can be solved with them or existing technologies. If they can that’s nice but that doesn’t produce a research problem or an entrepreneurial opportunity. Often you can’t just apply existing technologies so that creates a research problem. An awful lot of the research that I’ve done has essentially been entrepreneurial research; where it’s defining a problem as much as defining a solution, and similarly I have to marshal the resources to do the research project. Because that’s the way my search goes it actually aligns very well with entrepreneurship, because it’s exactly the same process one would have to take to be able to start a venture. Trying to solve some problem that you’ve been convinced that people have and then marshaling the resources to solve it.
Is there friction between taking something out of the classroom and into the business world?
There’s friction in everything I do — in the research as well. I’ve been very fortunate and I have a lot of publications, and it turns out they get read quite a lot. For example, I have three of the top five most read publications in Big Data Journal. When I talk to Ph.D. students about my problems getting papers published, they’re like “what do you mean?” If you want to do research on something that’s separate from what has already been defined by the research world as an important problem to solve, there’s a lot of friction in convincing the academic world to buy your product (in this case, a paper). To the question, “is there friction?” Yea there’s friction, there’s friction in everything that I do. But that’s often where the most value is created. I think that I have always been entrepreneurial, whether I’m an entrepreneur or not, I’ve always been entrepreneurial. That’s where the exciting work is.
What advice do you give to fellow faculty members that want to commercialize their work?
NYU has a very nice suite of options for faculty members to commercialize their work. Entrepreneurship is one of them. The most important thing for a faculty member is to think carefully about what they are good at, and what they want to be spending their time doing. I think it’s easy, especially reading about companies that sell for a billion dollars, to be seduced. The fact is that the probability is extremely small that your venture is going to sell for a billion dollars. But don’t go through life dreaming of doing one thing while you’re doing something else. I think that is the most important advice: do what you love, and don’t do it because you want to be rich. Do it because you’re going to love going through the process, and even if it doesn’t succeed it was still worthwhile to try. If you dream of seeing something you invented used out in the world, you should make it happen. If that’s not where you want to be, there are other ways to commercialize your research, or perhaps commercializing might not be for you. We’ve been blessed with people with a whole lot of different interests, and it’s a great job being a faculty member regardless of commercialization of research.
What do you think of the concept of a “gas tank”, and that you have to make sure that your “tank” doesn’t run out?
When I was working in industry, I thought I worked hard. I worked 10–12 hour days sometimes, and it’s nothing compared to being a faculty member which needs you 100% of the time. So you already have to try to figure out how to find time to do the other things that make life great. So you don’t want to get involved with something that’s going to take more away from your family life, if it is not going to be something that is actually going to be rewarding you (not just rewarding to you financially). Frankly if you’re the kind of person whose research can start a company, you can make more money doing something else other than a faculty member. You’re going to run out of gas fast if it’s not something giving you an intrinsic reward.
Tell me about Media6degrees.
I was approached by some entrepreneurs who were starting Media6degrees, and they were starting it based on one of my research papers. The story they gave me was that it wasn’t working, and they wanted me to help them get it to work. So I gave them some help. They came to me, and I loved the problem, and it was directly along the lines of the stuff I was thinking about. I agreed to spend the summer working with them to create what became Media6degrees, an ad targeting company using machine-learning technologies to target ads. I was essentially the founding data scientist and designed the initial targeting technology in collaboration with the original founders. Then I worked to build the data science team to replace myself. I’m nominally a co-founder, but what I wasn’t was somebody who was willing to give up his day job and work full-time for the company. Essentially, they call me the godfather, that’s my official title there.
How does Media6degrees relate to EveryScreen?
I was also a founder of EveryScreen. I partnered with the same venture incubation partners who put together Media6degrees. A different set of everyday founders for EveryScreen. Media6degrees then bought EveryScreen and then changed their name to Dstillery. In between the two, I did the same thing with what’s now Integral Ad Science, which was originally Ad Safe Media. In each case, I was the founding data scientist and then I worked to replace myself. I have an affinity for that super early stage entrepreneurial activity. I’m not cut out to run a 100-person business, or be involved in the running of a 100-person business.
What’s the best part of that “super early stage”?
Part of it is the same thing that I like about research. That is, trying to take this thing and look around at all the different angles until you finally find the way to see in such a way that you see the answer. This thing is really complicated because it’s simultaneously some technology, it’s some set of data, some set of potential customers, and some set of people that might be able to do this stuff. I like this very complex thing, and somehow you have to align all these different pieces, such that it crystalizes into a solution, which isn’t necessarily what you were thinking in the first place. You need to have this flexibility about what it is that you’re doing, so as you twist things around, you can get to something that could be great. I also like, with respect to entrepreneurship, that you also have to be very pragmatic about things. The fact is you’re building a company; you’re not doing a research project. In a project, although I’m very much driven by real problems, you do have the option to just do something because you find it really interesting. You don’t really have that option once you have another person working with you. In entrepreneurship you don’t have that option because other people are counting on the fact that this thing will eventually start making money to eat, pay somebody back, or pay somebody’s salary. Early on, you have this ability and opportunity to exercise this very complex alignment of lots of different things. Later on, it’s too late to do that. The business has moved and made a bunch of decisions and crystallized to a certain extent. You have so much less flexibility. You need someone who has CEO sensibilities at that point.
How did you balance the demands of your entrepreneurial activities and your position as a faculty member?
As a faculty member we have the ability to do a certain amount of consulting and summer work. One of the things I did was take no consulting and summer work. I think in the long run, I would have made more money doing consulting instead, but if my goal was to maximize the amount of money I was making than I wouldn’t have taken a faculty job. Secondly, I built research teams at the companies such that we end up doing research on problems that follow on to whatever it was we started the company with. I end up doing research with the teams that I created at the companies. Then it all comes together.