Part two of our interview with Rocket Fuel CTO, Mark TorrancePart two of our interview with Rocket Fuel CTO, Mark Torrance
Part two of our interview with Rocket Fuel CTO, Mark Torrance
September 28, 2015
Here's the second part of our exclusive interview with Rocket Fuel CTO, Mark Torrance where we dig a little deeper into enterprise adoption and global market opportunities for AI and new age ad tech!
AI Business: Is it true to say enterprise adoption of advanced ad tech is growing rapidly?
Mark: It certainly is, in terms of the depth of interest from enterprises in this kind of data, because the data can inform, not just the performance of the ad campaign, but really enterprises can now understand, with the same data, who their customers are, and what they really want. For instance, the company can gain insights that inform future product features or other things about how to better serve their customers when they use the digital data that we have access to. AI Business: Which verticals are you currently getting the most up take with? Mark: We have vertical specialists who work in certain markets – automotive, retail, travel, packaged goods, gaming and finance are up and coming rapidly. These are the major ones for brand advertising. The trend over time is advertising dollars are shifting from TV toward online. Partly that’s because TV is getting more complicated and there is now OTT (over the top), and many different channels, and partly it’s because the consumers are spending more of their time and attention online even for television content.
AI Business: Beyond the current verticals, which do you believe have the biggest potential longer term for advanced online advertising?
Mark: I think that anywhere brand advertising is important is going to be a later adopter and the spend will be huge. Again this would be more like consumer packaged goods and automotive. Even as big as those are already, I think that their digital spend is probably 1 or 2% of their total marketing spend in those verticals today. Whereas in retail I’m guessing it’s more like 30 to 50%. The fastest adopters have been in e-commerce because they can measure the sales directly online. If you’re trying to get somebody to go into a store and do something that is a little more abstract and so it’s going to come later. Companies like Macy’s, for example, would still spend a vast majority of their dollars offline.
AI Business: How does the geographical split fall across the earlier adopters we’ve discussed?
Mark: Europe and emerging markets are growing most rapidly. In Europe we started in the UK and France and then Italy and Spain and other countries. I think that the fastest emerging markets for digital advertising include the parts of Eastern Europe that were a little slower to adopt this, and also Latin America; they’re all growing really fast but are still a lot smaller than our core markets today. Growth in the US is still robust and still represents a large focus for us.
AI Business: What is it that really differentiates Rocket Fuel in the market?
Mark: I think that there are other companies that use machine learning to build models. Mostly those companies are doing that to put people into segments or groups of people.
Rocket Fuel is differentiated in that we are now talking about ‘Moment Scoring,’ which is differentiating every moment of influence where we could put a message in front of a person. So we might say you are a better prospect for this product now, but a worse prospect tonight at midnight when you’re just online browsing around for other kinds of stuff. So we are not putting you in a segment where you’re in or you’re out where we just blanket advertise to you. And that’s using a model but in a very different way. With our approach, we make decisions on whether to show you an ad and which ad to show you based on how we score the moment against a very wide range of attributes. The segment approach would simply advertise to you every time it saw you online without regard to any sense of your receptivity at that moment.
We actually rebuild our models automatically every day, and the models have hundreds of thousands of features in them, and they’re trained on all these different behavioural and contextual and demographic aspects that we know. All of these elements are things that, when other companies are not using AI, a person has to guess, or fail to and then you just have this missed opportunity, this sub-optimal performance. But if you’re applying AI in the right way it can work so much better.
AI Business: How do you see it evolving beyond where we are today, looking forward over the next five years?
Mark: I think that we’re going to see a few areas. One is a thing called premium optimisation. For example, if you know that you want to show your ads on the New York Times website, we could say we can also show your ads on a bunch of different sites you’ve probably never heard of and get better performance than if you just showed your ads on the New York Times. But if you have decided that you’re going to show your ads on the New York Times, then rather than just show your ads to whatever random subset of traffic New York Times gives you, we would like to be sitting in the middle of that, optimising that. So we say, ‘Hey, New York Times, give us 4x as much traffic as the advertiser is really going to buy, and we will use our automation to cherry pick the right subset of that, that will drive the advertiser’s objective’. So that’s one of our offerings that is new.
Another exciting area I’ve kind of alluded to is the fact that it is easier for direct-response advertisers to use digital than for brand advertisers to use it. Brand advertising might have the goal of getting you to buy the product later or it might have the goal of getting you to use the product more – like a credit card that you already own, just to remind you that you could be using these Chase reward points if you were to go and spend and put things on the credit card.
So I think that understanding brand advertisers’ needs better and having the digital channel be more like TV in the way that people can buy it, like using gross rating points and things that speak the language of brand advertisers, all of that is a direction we’re headed in.