Keep up with the ever-evolving AI landscape
Unlock exclusive AI content by subscribing to our newsletter!!
Exclusive Interview with Rocket Fuel CTO, Mark Torrance on the future of AI in marketing
September 28, 2015
We got the opportunity to meet Rocket Fuel CTO, Mark Torrance to get a much deeper understanding into Rocket Fuel's approach to next generation advertising, and specifically how artificial intelligence is at the core of the offering.
A business which has scaled rapidly since it's launch in 2008 with AI at the heart of the company. Here's the first instalment of our two-part special feature, enjoy!
AI Business: Could you give us an overview of Rocket Fuel how AI is evolving the digital Ad market?
Mark: Rocket Fuel has been around for almost eight years now. It’s a company that helps advertisers show their ads to the right people on Internet websites and we use machine learning and AI to do that.
Even digital marketing, traditionally has been surprisingly human-driven. You may have electronic means to buy advertising, but it’s people deciding who the ads are going to be shown to: ‘I’m going to show ads to people who have been on my website before or have shown interest in buying a luxury sedan according to this data I can buy from another company; or I’m going to show ads to people who I’ve never shown an ad to before because I want a certain frequency cap.’
All of those are interesting ideas and they work better than random, but they’re just people’s intuitions about who to show ads to. If you let machine learning systems figure out who is actually responding to the ads that we have shown so far and who are spending the most money buying products or demonstrating interest in the products that we’re advertising, we can get a feedback loop going where we can learn which online ads we exposed that person to may have influenced that purchase.
All of those things can feed machine learning models to help them get better everyday so the ad campaign automatically gets better and better and better at showing the ads to the people who are most likely to respond. So that is the approach we take at Rocket Fuel. We have evolved from doing primarily direct response advertising, where we’re measuring the sale of things online like pizza you order from a website, to brand advertising where you’re showing video ads or display ads. In this latter case, we’re not trying to motivate somebody to take an action online but rather just influence them the way TV ads do and then maybe when they’re in the grocery store or shopping for a car they will actually do something based on those ads that they were exposed to.
AI Business: What are you doing to strengthen this proposition?
Mark: We acquired a company last year called [x+1], which is in the marketing services space and has augmented our offering with a data management platform (DMP) that allows companies to take the first-party data that they have about their own customers and make that actionable online. So if we’re working with Chase or Discover or Electronic Arts, for instance, and they want to show certain ads to their best customers or to their most promising prospects then they can use the data they have about those customers in their own database and turn that into actionable data online. What offers are shown to people within a game or when they are visiting the website of Discovery or Chase – all of these are examples of where Rocket Fuel uses AI technology.
AI Business: So has AI always been part of the offering or has it evolved with the business?
Mark: AI has always been part of the offering, and in fact it was the founding premise. The team that founded Rocket Fuel started out working at Yahoo and applying AI but just in one part of advertising called ‘behavioural targeting.’ There was another group at Yahoo involved in contextual targeting and were not applying AI; contextual meaning the context of what is on the webpage at that moment of time. The founders of Rocket Fuel ended up getting together and leaving Yahoo because they really wanted this broader vision of using behaviour, context and demographics.
I would say that the history of Rocket Fuel was very focussed on AI and machine learning specifically from the beginning, and the way that it evolved was very powerful and very successful. We want to evolve to a place where people and machines are working together, collaborating to make the performance of ad campaigns even better than either one could alone.
AI Business: With regards to behavioural advertising, is it a shift to the ‘Why’ rather than the ‘What’ which is the biggest opportunity?
Mark: I think that the ‘what’ is used as the training signal, but it’s not just that ‘what’ by itself.
What we do when we learn is, ‘okay you just bought a pizza,’ if we actually look back at all of the ads in time that we showed to you and we look at what you were doing online at the time that we showed you those ads according to our data, then we say, ‘hey the last ad that we showed you, which apparently really persuaded you, was on your phone, not on your computer, that is interesting… we showed the ad to you when you were at work not when you were at home… we showed you that at 2pm and it was sunny outside at the time and the last thing that you had done on your phone right before we showed you that ad was that you had been on an app that was about the weather and the last thing you did on your other device – which we’ve joined together with the phone data – is see an ad about a pizza just five minutes earlier.’ That is the kind of data that we have about you as an individual.
We don’t know your name or your phone number or your address, but we do know all this very rich time-stamped data about what you’ve done related to advertising. So all of that information, when we say behavioural advertising, that is what we mean. We’re using that kind of data to inform this question of ‘should I show you an ad right now for Pizza Hut or for a car or for a boat or for a private island or for some other thing.’
We have about 3,000 advertisers that we work with at any point in time, and we’re looking at the consumer for all of them, not just for what might align with the last thing that you interacted with.
AI Business: That's it for part one, we'll be releasing part two shortly so be sure to check back and tweet us your comments @Business_AI
You May Also Like
Generative AI Journeys with CDW UK's Chief TechnologistFeb 28, 2024
Qantm AI CEO on AI Strategy, Governance and Avoiding PitfallsFeb 14, 2024
Deloitte AI Institute Head: 5 Steps to Prepare Enterprises for an AI FutureJan 31, 2024
Athenahealth's Data Science Architect on Benefits of AI in Health CareJan 19, 2024