Bill Simmons is a co-founder and currently the chief technology officer of DataXu, a digital marketing management firm that offers platforms and services for advertisers and agencies in order to increase their advertising efficiency and marketing ROI. Bill obtained a PhD in astronautics and aeronautics from the Massachusetts Institute of Technology (MIT). Before working for DataXu, Bill was engaged in developing and testing programs for space launch vehicles. He is known as one of the pioneers in real-time software and optimal system design. In this interview, Bill talks in detail about how DataXu helps its customers and processes billions of big data streams every day, and what kinds of opportunities there are for entrepreneurs who wish to start in this field.
Sramana Mitra: Bill, let’s start by introducing DataXu to our readers.
Bill Simmons: DataXu is a technology supplier to advertisers and agencies. We build a digital marketing management platform. What that means is we help advertisers and agencies manage their big data and put it to use to make their marketing more efficient. We have everything from data management to analytics and insight management platforms all the way down to media buying optimization technologies.
SM: What is the genesis of the company?
BS: DataXu was born out of MIT. My background was in aeronautics and astronautics. I studied combinatorial optimization and graduated in 2008. I was applying the methodologies we developed to design Mars missions, which is a combinatorial optimization problem – choosing how many crew members, what kind of technologies, what kind of trajectories, what kind of engines, and what kind of missions to apply. All of these different combinations make the design very different. It may seem like an unnatural match, but advertising on the face of it is also a kind of combinatorial optimization problem, where you have billions of users on hundreds of thousands of sites, including mobile display and social video, and you potentially have thousands of advertisers, each with hundreds of ad creators. You need to be able to decide what goes where to make the best return on your marketing investment.
SM: Interesting, because I am a computer scientist educated at MIT. Going back to your proposition: where do you plug in? Obviously, this combinatorial optimization problem is a vast problem. Whom are you serving? Are you serving advertisers or portals? Where are you applying your algorithms?
BS: It depends on how a brand manages marketing. Some manage it directly, so we work directly with those brands. In other cases they go through an ad agency to manage their marketing technology and marketing strategy. In that case we work directly with the agency, advising the client on a periodic basis.
SM: You are working on behalf of the advertisers of specific brands and their spending on media buying?
BS: That is an interesting point. Toward the end of 2008, an opportunity emerged to buy media through real-time bidding. That means that you could use an algorithm bidding system to buy media without owning or operating any media. What that enabled is companies like DataXu to emerge, which are [...] only technology companies. We don’t have any direct relationships with any publishers or any kind of media. We don’t own any media. We only represent the buyer. Our business model is about metrics and KPIs [key performance indicators] aligned with buyer success. There are companies that have emerged on the other side of the fence, too, that only represent publishers and don’t represent any buyers. What this model enables is a much cleaner market. You can think of it like buying a house. When you buy a house, there is a buyer-side broker that represents the buyer and a sales-side broker that represents the seller. It is a much cleaner market that encourages liquidity.