Sramana Mitra: Could you tell me a bit more about how this works from a technological perspective? “Tying an event down to the millisecond level in a video” – how do you do that?
Derek Rodner: That is actually our special [approach]. In addition to being the only player in the retail operations big data space, we started our business doing something called point of sales video auditing, and we have a proprietary [technology] that enables us to take real-time data on the network and attach that with video in real time. In fact, 10 percent of the toll routes in the U.S. carry our technology. Admittedly, it’s an older technology, but it’s the same basic underlying technology.
SM: How big is Agilence?
DR: We have around 50 employees and about $8.5 million in revenue this year. We were founded in 2006, and we are venture backed – granted ventures from our primary venture firm out of San Francisco. Then we have several employees [near] each capital [partner] located in Pennsylvania, and other capital partners also based on the East Coast. The other company just changed its name; basically, it is a venture fund based out of Schneider Electric.
SM: And where are you based?
DR: We are based out of Mount Laurel, New Jersey.
SM: Let’s shift to a different topic, since we put these pieces of context together. From you vantage point, what do you see in the industry? What are the trends in big data as it pertains to video analysis?
DR: As far as big data and video analysis are concerned, there is a significant drive to a point where the unstructured data – that is, video – can be analyzed, whether it is in a camera itself or in an application, to get meta data out of that unstructured data and use it to count people. How many people are coming in and out of a store? You can also use that data for dwell time analysis – how long does someone linger in a particular area, by a rack of sweaters or a particular promotion?
There are also advanced analytics that look at demographics. Did a male walk in the door, and what is his age range? Gender, race, and age are some of the things we are looking at. In addition, there is analytics for facial recognition. Taking the face of somebody who walks in the door and comparing it with the database of the best shoppers or with the database of known criminals. Those are some of the video analytics that will be coming out of this space. More and more, the goal is to take all that information and marry it into a single application. The video analysis space is creating a lot of the “big” in big data.