Sramana Mitra: So you have to basically manually speed the algorithms with the parameters. You’re almost providing certain constraints within which the machine is learning, as opposed to a completely unconstrained learning environment.
Patrick Shea: That’s exactly right. Similar to guided learning, which is very popular in artificial intelligence, we’ve always said that our process is a human-guided algorithm. That really helps describe what AdDaptive Intelligence is versus what artificial intelligence brings to the market.
Kevin O’Malley: If there’s one other trend that I would add that I think is important to the space, it would be the blurring of the line and the merging of the online world and the offline world. Traditionally, offline marketers have collected tremendous amounts of information from the number of times you get an oil change to what you buy in a retail store. Online data has always been based on your digital footprint. In many cases, a bit more anonymous than the offline data.
What we’ve seen over the last few years is the convergence of those two. In order to make sense of it, if you look at 20 to 40 years ago, you had to collect the mailing list, analyze a few attributes, and do it in a spreadsheet and then send out some mail for those users. That was a very effective tactic. Now when you’re talking about the hundreds and thousands of data points that you can collect from an individual that encompasses both their digital footprint and their offline footprint, you really need some powerful and intelligent machinery to go through all that, to make decisions in real time, to incorporate new data in as fast and efficient a way as possible in order to make those decisions in real time.
Sramana Mitra: Given that discussion, what do you think are some open problems where you can do something? If you were starting a company today – not what you’re doing already – what kinds of problems do you see out there that can use some interesting artificial intelligence applications to deliver interesting value?
Patrick Shea: We try and stay as nimble of a company as we possibly can in order to address those up and coming elements in the industry. One thing about the digital media ecosystem is that it changes incredibly fast. The challenges that a marketer looks at a year ago are different from what we’re looking at today, and they’re going to be different a year from now. I would definitely say that one of the common themes is there is the adaptability and the idea going into it that you know you’re going to have to continue to evolve with the industry.
With that said, I think some of the specific challenges that the market faces today is tremendous fragmentation. There is a very famous analysis of the ecosystem called the lumus gate. When you look at it, it’s an incredibly complex snapshot of the digital ecosystem. All of the partners that essentially a marketing dollar flows through from advertiser to publisher. I think that solutions that can simplify that are going to be incredibly important.
The other half of that is having people who are experts in this new complex ecosystem that we find ourselves in. Doing the simple repetitive tasks that you would have a roomful of traffickers doing five to eight years ago is no longer an option. You need to build systems that can automate as much of that as possible. In doing so, you need to train, recruit, and bring on people who are able to work very well with those systems and think about what the problems are going to be over the next few years.
Sramana Mitra: Great. Thank you very much.