Sramana Mitra: Is DataXu a 30 percent services and 70 percent products kind of company?
Bill Simmons: Currently, the majority of our customers are managed service clients, but we continue to see an uptick in self-serve clients, and we expect that to continue in 2013. I think the reason our technology gets adopted is because our service is good.
SM: A couple of days ago I was talking to a big data analytics services provider operating out of India. They a couple of hundred people who do very sophisticated analytics work and provide analytics services. He was saying that there is a real shortage of analysts who can cross that bridge between technology and the business metrics and analytics. I think there is quite a deep need for these kinds of services that are domain specific and for people who have deep domain knowledge in terms of both the businesses and the technology needed to make it work.
BS: There is a merging need for people called data scientists. Data science involves three skills: It is understanding business, it is understanding mathematics and statistics, and the third one, which is actually the hardest one, is understanding the results in a compact way that is effectively communicated. This is called data visualization. We can often find candidates who have two of these skills, but having all three is extremely rare. As we build up teams here, we generally try to hire people who have two of the three skills. If you have all three, that means you are a superstar. I had a lot of conversations with local universities, trying to put together programs to train people to be data scientists. A lot of those programs are centered on those three themes – business acumen, math and statistics, and data visualization and communication – but the third part is really the toughest.
SM: Which universities are you working with to build these kinds of skills?
BS: I have had conversations with Northeastern, MIT, Harvard, and Northwestern in Chicago. They picked up on this trend that this is needed and they are trying to create programs to fill it.
SM: That is very interesting. Is your team entirely in Boston?
BS: We have 13 offices around the world now, but our technical team is primarily in Boston. Our engineers and data scientists are all here.
SM: The last segment I would like to discuss is the following: Given where you sit – you are obviously highly in tune with what is happening in the application of data sciences and advertising optimization – what are some of the white spaces and opportunities you would point entrepreneurs toward?
BS: I think some of the major challenges in digital marketing are measurement and unifying this measurement across channels. A typical professional person typically has three devices he or she uses on a regular basis: a mobile phone, a tablet, and a laptop. It is becoming increasingly challenging to measure precisely what is going on. There is a lot of activity in startups trying to crack that problem, including in DataXu. That is an issue. There is also the opportunity to make more traditional media more programmatic. For example, some very high-end websites still sell over the phone or scanned PDFs. There is an opportunity to service those kinds of clients and give them a way to sell their media programmatically. Also, TV and radio are still sold in a very traditional way. Bringing all those things together allows companies like DataXu to help large clients understand the value and buy more of it.