Thought Leaders in Big Data: Interview with Dr. Saed Syad, Chief Data Scientist at AdTheorent (Part 5)

Sramana Mitra: So your system sits in between the advertiser and the different publisher venues into which you are placing those ads?

Saed Syad: The publisher is on the other side. We are on the side of advertisers.

SM: But you still need to be able to access publishers. Otherwise, how are you going to publish these ads?

SS: Through a middleman – the ad exchange. The ad exchange gets the publishers’ information and we bid on the placements.

SM: You bid on placements on behalf of advertisers or on behalf of the ad exchangers?

SS: On behalf of advertisers. Ad exchanges are a middleman, like a stock market. It is a place of transaction. They get the information from the publisher about the site. If they want to show something and place it on that site, they send it to different DSPs for bidding. There is a publisher, then the ad exchange, then us, and then the advertiser.

SM: What are your most successful clients in leverage your kind of optimization?

SS: One of them is Bank of America. Based on what they told us, they have 10 times more clicks compared to their past campaigns.

SM: Where do you see this real-time bidding industry going? In each of the three vectors you talked about – scale, accuracy and speed – what is the scope for enhancements?

SS: RTB is all about speed. RTB and mobile are going to be the future of the Internet advertising. Recently Facebook’s stock went up 30% because of their focus on the mobile side. In this area, real-time bidding is going to be the future of online advertising. There is speed and scalability there instead of an approach of finding a person at the right time at the right place. We can have a real-time model with a much higher level of accuracy. The future is going to be all about RTB plus mobile.

SM: My question is specific for RTB. People understand that RTB is an important part of where this industry is going. But what are the open problems and scopes for enhancement within RTB?

SS: On the RTB side, we should go beyond click. We should go to conversion. The challenge in the future is how good we are in predicting not just clicks, but also conversions. On the predictive modeling side there are other aspects, for example, the quality of the ads, taking care of fraudulent clicks or conversions, and some other challenges. But on the predictive modeling side, we mostly want to make sure to bring more variables for our models from different sources; you mentioned transactional data before in this regard.

We also need to move from clicks to conversions and take into account the buying habits of customers. If we have RTB on mobile, we can track the movement of customers, which you don’t have on the desktop. We also have challenges with how to detect the trajectory of the customer moving from one place to another. Those are the types of challenges we will face in the near future.

This segment is part 5 in the series : Thought Leaders in Big Data: Interview with Dr. Saed Syad, Chief Data Scientist at AdTheorent
1 2 3 4 5 6