FICO's chief analytics officer discusses mobility, machine learning, consumer empowerment and why big data is no savior.
Andrew Jennings Senior VP, Chief Analytics Officer, FICO
Best known for its credit scoring and fraud-detection services, FICO has been working with and interpreting huge volumes of data for much of its 57 years.
More recently, the company has broadened beyond financial services, offering its big data decision-support skills to brands like Samsung, Capital One, Best Buy, Conagra and the National Football League.
Helping lead this charge is Andrew Jennings, who took on the new title of senior VP, chief analytics officer, 18 months ago. Jennings helps FICO bring predictive analytics to thousands of companies in more than 90 countries through a variety of platforms, including applications, tools, wireless and cloud.
"What we've learned is ... machine learning tends to be naive," Jennings told InformationWeek. "Understand the domain, understand the data and then understand where and how this thing you've created is going to be used. If you don't understand those three things, it doesn't matter what algorithm you use, it will fail."
InformationWeek spoke to Jennings in March via phone.
Name: Andrew Jennings, senior VP, chief analytics officer, for FICO.
Tenure at job: I've been at FICO since November 1, 1994. The chief analytics officer job, about 18 months. Before that, I was head of research for three years.
[The new title of chief analytics officer] was a recognition of the growing importance of analytics in the business world. FICO has been in analytics since day it was founded, over 50 years ago. But we never actually had a chief analytics officer. Now that analytics is at the forefront of the way that businesses can and should be operated, it made sense to have somebody designated as the chief analytics officer to be responsible for both a large number of the analytic products but also, more importantly, research and strategic direction.
Career accomplishment of which I'm most proud: Given my background of being a risk manager, writing checks and having to meet sales numbers, I've always been very focused on doing the research side of the job [that's] focused on business problems. I like to say, "It's not the FICO University. It's FICO R&D, and we are a for-profit company."
Decision I wish I could do over: Taken more time after graduating with my PhD and going to work.
My most important career influencer: Not someone I worked with. I'm an economist by training ... the person who influenced me most, and his ability to think and describe still does, is John Maynard Keynes. What I like about Keynes is while he was mathematically brilliant, his books are extremely well-written and they're not full of formulas.
Current top initiatives:
-- Mobility. That's a huge area of growth, and it's a focus for FICO as a company. It's a channel which establishes a conversation with a consumer. Since at the heart of what we do as a company is forecasting and attempting to predict consumer behavior -- will you pay your bills or not? is that really you at the end of that credit card transaction? -- the mobility communication takes that to a whole other level. Using messages and interpreting responses to more intelligently communicate to both parties' advantage, analytically, is quite a complicated challenge. The company as a whole has an initiative around that because it impacts the work we do in fraud, collections, marketing and so on.
-- Harnessing the power of big data algorithms and machine learning [but] translating that into an analytic form that is transparent, and from which you can calculate what we call "reason codes." [These codes] are explanations for consumers, management and regulators why certain decisions have been taken.
Most disruptive force in my industry: Consumer empowerment. I don't really have a prediction for how it'll play out, but the ability of consumers to take more control over the data they generate.
Biggest misconception about big data: That it's a savior, that it'll solve all problems. There's a lot of debate about causation and correlation. They are not the same thing. There are circumstances where correlation is sufficient and there are circumstances where it is absolutely not sufficient.
The reasons big data projects go wrong: Just because it's big data doesn't mean there's no ROI or business-case required. You need something concrete you are trying to solve, that has a meaningful impact on the business.
A promising technology: The ability to understand and adapt models in close to real time. We call it "self-calibration." [This is valuable] especially in situations where I don't have time to collect data or in situations where I don't have a lot of historical information. For example, we're very big in fraud detection. We're working with a couple of banks that have historically poor data, [and] we've been able to take that kind of technology and create models that begin to learn and adapt.
Andrew Jennings At A Glance
Education: BA and PhD in economics, and an MSc in agricultural economics from the University of Wales.
Person I'd most like to have lunch with: Admiral Lord Nelson.
First job: Selling ice cream in Weymouth, on the south coast of England.
If I weren't involved in IT: I'd start a landscape maintenance company.
Attend Interop Las Vegas May 6-10 and learn the emerging trends in information risk management and security. Use Priority Code MPIWK by March 22 to save an additional $200 off the early bird discount on All Access and Conference Passes. Join us in Las Vegas for access to 125+ workshops and conference classes, 300+ exhibiting companies, and the latest technology. Register today!
6 Tools to Protect Big DataMost IT teams have their conventional databases covered in terms of security and business continuity. But as we enter the era of big data, Hadoop, and NoSQL, protection schemes need to evolve. In fact, big data could drive the next big security strategy shift.
Big Data Brings Big Security ProblemsWhy should big data be more difficult to secure? In a word, variety. But the business won’t wait to use it to predict customer behavior, find correlations across disparate data sources, predict fraud or financial risk, and more.