T3 Talks: T3 CIO Jack Lynch Talks Data Science And Hot Tubs At The South Pole
Jack Lynch, T3’s first Chief Intelligence Officer, sits down with Director of Marketing Intelligence Christopher Macauley to talk about machine learning, data science, the future of the practice at T3, and, yes, hot tubs at the South Pole.
Christopher: What’s the most interesting fact about you that would not be on your LinkedIn profile, your resume, or your Facebook profile?
Jack: I helped build the first outdoor hot tub in the South Pole. In grad school, we were doing research and building a neutrino detector using the ice in Antarctica as the detection medium. First you have to bury the detectors deep under the ice—the best way to drill in ice is with hot water and long hoses—so when we weren’t drilling, we had these tubs of hot water that we set to 108 degrees. I have a photo somewhere of us sitting in the hot tub with frost growing on our noses and the barbershop-looking pole in the background marking the South Pole.
C: What do you appreciate most about data science in regard to problem solving? Is it different than what clients value the most?
J: I think clients value the ability to drive return in whatever version is relevant for their business, whether that’s loyalty, higher value transactions, an increase in volume, etc. But what I find most exciting, and the reason I studied physics, is having interesting problems and challenges that can be solved via the use of reason, observation, and the scientific method. Data science is a perfect match for that. It allows me to tackle a lot of problems I wouldn’t normally be able to tackle. It’s exciting to use data, scientific methodologies, and emerging capabilities within data technology and analysis, including things like machine learning and artificial intelligence to solve problems.
C: What advice do you have for clients struggling with data science, and where are some areas they can find quick wins to move forward?
J: There are a couple of pieces of advice for clients. One is you can’t just jump in. There are a lot of tools that let you do things but it’s easy to make mistakes when looking at data and only trusting the tools. There are a number of cases where someone does an analysis and they miss something fundamental like not properly handling a missing data element, which leads to seriously erroneous results. So the first piece of advice is to make sure you have someone with a background in data be part of the analysis.
In terms of small wins, it really depends what their business is, but a quick win could come from better data integration to get a more consistent view of their customers across multiple touch points. This might be tracking behavior of their customers when visiting their site and then changing an app experience based on that behavior. Or observing their app behavior and changing the message they put in an email. Those small pieces of data integration are excellent opportunities to show some return on investment. I also think there’s some opportunity within some of the small test and learn programs, whether it’s tweaking messages through A/B or multi-variable testing. You get 50-70% of your returns just doing the basic things we’ve been doing for a while in data-driven marketing. The last 30% comes from some of the more advanced processes with more involved segmentation and machine learning.
C: Machine learning—more than “black boxes”?
J: Machine learning is interesting. The challenge with the practice of data science is we’ve become awash in larger and more diverse data sources. The advancements in machine learning over the last 5 years have been very significant in large part due to the big technology providers like Google, Netflix or Amazon trying to improve their products. Coupled with ever-increasing data storage and processing power, machine learning is coming of age.
The opportunity for machine learning within digital services and marketing is to ingest all the data sources we have around our customers—their transactions and interactions across all brand touchpoints—and use that to predict what the next best experience is for them.
It can feel very “black box” to marketers and again, that’s why you need some folks who are trained in the right areas to make sure they’re building the models correctly.
C: Any advice from your previous employer?
J: Marrying the arts and sciences is where the magic lies in building a data science practice. To marry the business value to the ability to process and take advantage of the data is probably the best piece of advice.
The second piece is in not setting the sights too far ahead. Have a long roadmap of where you want to go, but don’t miss out on opportunities to quickly improve your value. You can tell a story of magical, automated marketing machines making perfect decisions and delivering experiences to customers, but you would be over promising. We feed that story a lot as marketers, and it’s possible to get there, but don’t try to build that right out of the gate. Instead, think of the small wins that will help validate the capabilities and get funding for future endeavors.
C: Through what leading technologies do you see T3 applying our data science chops in 2018?
J: I spent most of my career in data-driven marketing technologies—going back to the old days of deploying ERP and CRM systems and trying to take advantage of the data to deliver the promise of one-to-one marketing. For a long time, that was only possible in limited channels. It might be in the email channel alone or it might be the website but then the website didn’t integrate with email or any of your other marketing endeavors. Mostly, we’ve been trying to invent the technologies to make that possible.
But now we have systems—like marketing cloud services—that provide the tools that will let you connect data about your customers coming from multiple channels and deliver messages back inside of those channels. That promise of one-to-one marketing is now possible. Our current attention is now shifting to messaging and experience. Now that I can deliver a personalized message to you across almost any channel and do it at the right time, what’s that message supposed to be? What’s the right experience to influence your behavior to build that relationship and add value? What will be the more engaging experience?
That’s where T3 is uniquely suited. T3 has that combination of technologists and engineers that make it possible and we also have creative and innovative thinkers that can design experiences and understand customers. I think more than any specific technology, that combination of people who understand customers and their experiences married with the technologists who help deliver it—that’s where the real magic is going to happen.
C: What are your thoughts on the notion that Big Data is killing innovation?
J: Absolutely wrong. Data doesn’t kill anything. Data only enables. As you gain access to more information, you better understand the challenges that your customers are facing. It opens up the field for more areas to innovate and find innovative solutions to address challenges that you may not have seen without that wealth of data.
C: In 2017, what have been the most inspiring and/or creative applications of data science?
J: Waymo just launched their driverless service in Phoenix without a human in the driver’s seat. I don’t think most people (including policy makers) appreciate how much autonomous cars are going to revolutionize our way of life. There have been really promising advancements in cancer treatments where data science has been able to recommend targeted therapies based on DNA, and even beyond that to the RNA and protein-expression levels. Imagine that: recommending treatment based on not only the type of cancer but on how it is specifically interacting with you.
Oh, and Alexa tells funny jokes.