Industry Insights

Background image of user typing on a calculator with floating interface elements surrounding them

FICO CAO Scott Zoldi on Analytical Innovation in Financial Services

Posted
July 25, 2016
scott zoldi

As Chief Analytics Officer of FICO, Scott Zoldi has a unique—and informative— perspective on the role of analytics within the financial services space. With a background in analytic science and a PhD from Duke, he is an active member of the San Diego tech community. Our Burtch Works recruiting team that focuses on analytics within financial services, connected with him recently to discuss career advice, analytic innovation and challenges, and keeping up with this exciting and rapidly changing industry.

If you were a junior predictive analytics or data science professional in the current job market, what advice would you give your former self in terms of career progression and advancement opportunities?

Constantly network and look at finding the right experienced mentors in both data science and business that can help be an advocate and guide one’s career. The marriage of analytics/business equates to maximum value and achievement in one’s work and ability to contribute. Very often junior analytic/data scientists over-emphasize their technical achievement, as that is often what they spent their degrees doing and they lack solid business groundings. Being able to translate business problems into analytic problems to solve is at the key of our business. Some junior professionals harm their career progression by staying too close to their comfort zone in the technical side, and don’t focus on where their skill levels are weaker on the business side.By not having strong business understanding of the problems that they are working on, they can’t effectively translate their results into business outcome, or they may not recognize how they could pose the analytics differently to be more effective for a client. In addition to mentors within and external to your company, it’s critically important that you become the go-to resource for clients, who will deeply understand the challenges they need solved and typically will be more than happy to spend time directing strong analytic talent to help them focus on them. At FICO one of the great parts of our business is working with so many different clients on so many different analytic problems – this provides a huge resource of experiences. If you are at a single company working on one type of problem, then network and be inquisitive to expose yourself to other analytic problems/business problem areas.

As a hiring manager, what kinds of traits, abilities, or skills do you look for when hiring analytics professionals?

FICO has been hiring analytics professionals since 1956, and over the years, we’ve learnt that the best candidates have more to them than mathematical skill alone. Besides modeling skills, we need someone who can look behind the numbers to the business problems. A good data scientist is able to ask smart questions of his data, starting with: What decision are we looking to improve? How can we reach it? What is hindering us? How can we measure our improvement?The best data scientist can communicate her findings clearly to colleagues and clients who don’t have an analytical background. The best analytical talent can share ideas and insights in a way that is meaningful to any business audience. I can’t always judge this from a CV, so I make sure candidates can convey their ideas well in the interview.I put more emphasis on skills and mindset than degrees. Of course, a strong background in numerical science is a necessity, but good data scientists must also be detail-oriented, inquisitive and open-minded.

american-express-89024_640
Change and innovation in the analytics space seem to happen in the blink of an eye. Certainly this is true in fraud within financial services, where a company like FICO has to keep up with the ever-changing tactics of the fraudsters. With technology and tools constantly shifting, how do you keep up? What advice do you have for others trying to stay current in their technical knowledge and skills?

Keeping technical knowledge current is important, to help scientists understand what is out there and what should they and their companies should pay more attention to with respect to new tools and methods. At FICO we have the luxury of having a very large analytic scientist team, where different scientists naturally have technical areas of interest that differ from others’, and meet in regular scientist meetings to explore and discuss new methods. We further focus on small proof-of-concepts of technology to compare and contrast new methods and tools. Not every new technique/tool is an improvement and many are not ready for prime time in terms of stability. These kinds of ‘art of use’, proof-of-concepts allow quick investigations to take place.What one should not do is treat every new open source tool/analytic library as a new junk food to consume. Focus on the communicated differentiation/benefits, determine if those benefits are needed in the analytics employed today, research other’s experiences through your network, and if convinced give it a try. I find it helpful to maintain a list of new tool/innovation areas to research and to prioritize. This helps catalogue the areas that your company and you are most interested in. Just putting together this list provides the visibility and prioritization of what to investigate.I am also a huge advocate of creating your own new tools/algorithms to solve problems. Often doing this comes after deep knowledge of your analytic/business areas and the current state-of-art of tools/methods. This often leads to creation of new intellectual property/patents, which are very important. Often we find that data scientists become ‘tool users’ which ultimately means that their knowledge of the methods or algorithms they are using are often not deep. This can lead to large problems down the road in terms of use of these methods in analytics problems and the erosion of the analyst’s own technical strength. By creating one’s own intellectual property or deeply investigating, one can use new tools/technology for convenience (meaning it’s easier to use what’s on the shelf or in open source vs. creating one’s own versions, which come with development and maintenance costs).

What are the biggest challenges facing the financial services industry today that you see analytics playing a major role in solving?

All of them! There’s really no limit to what analytics can do. But narrowing the field to the challenges that FICO addresses, I would call out three: financial inclusion, regulatory compliance and security.Financial inclusion means making credit available to millions of people around the world who currently aren’t part of the financial mainstream. They often don’t have bank accounts or credit history, making it difficult to assess their credit risk. FICO has developed new approaches to assessing these people so that lenders can say yes to more of them.Lending is one of the most highly regulated industries, and that has increased with regulations such as IFRS 9. This changes the provisions banks must make for credit risk, and will require complex new analyses to be executed very swiftly every month for reporting. In addition, there are regulations in areas such as financial crime that call on banks to adopt more analytics.Finally, there’s security. Cybersecurity is a threat not only to financial services but to every business on the planet that uses computers. We must deploy more advanced analytics to detect and prevent hackers from stealing information — the current level of industry safeguarding is clearly not good enough.

pexels-photo
In one of your recent blog posts, you talk about a concept you call the “Intelligence of Things,” or the idea that self-learning devices can become even more valuable ifthey can detect changes in themselves and their environments and integrate such information into a centralized system. What are some ways in which this technology could be applied in the financial services industry?

Great question! Today much of the self-learning analytics technology is already employed in the financial services space, but in the back office. For example, in our fraud product we have analytics where ATMs, merchants, and point-of-sale devices are constantly self-inspecting themselves in terms of normality or abnormality and feeding signals into decisions on whether the card presented at the terminal is likely fraudulent. Merchants and ATMs are routinely compared to one another and the riskiest ones are identified using peer groups based on collaborative filtering technologies.The same self-learning analytics that run in back-end analytic systems can run on the devices themselves. For example, I am actively engaged in research around mobile banking. The mobile devices that we utilize in our daily lives can help tremendously to differentiate payment transactions that are merely abnormal from those that are actually fraudulent. I see huge changes coming in the mobile devices that will help protect our accounts, secure our personal data, and help us achieve our personal/financial goals through better understanding our actions and daily movements.

As a CAO, do you see any evidence that analytical teams are playing an increasingly larger role in high level executive decision making?

Absolutely. First of all, I am meeting more Chief Analytics Officers — it’s a relatively new addition to the C suite, but it’s definitely catching on. This is a recognition that businesses require deep, sound analytic thinking at the highest levels of the organization. Second, more universities are adding analytics coursework to business degree courses. It’s too easy to drown in data, particularly today, and tomorrow’s managers will be trained on how to analyze that data, or query someone presenting analysis to them. These trends will continue, and the enterprise of the future will be much more analytics-driven (not just data-driven) than it is today.In short, the potential problem-solving capabilities of analytics in the financial services industry are almost endless. Fresh ideas from emerging analytics and data science professionals, coupled with increasingly sophisticated self-learning technology, will turn the challenges of today into the triumphs of tomorrow. Many thanks to Scott for sharing his insights with us at Burtch Works! If you’d like to hear more from him, check out his contributions to the FICO Blog and follow him on Twitter @ScottZoldi. We hope you enjoyed this Q&A and we would like to hear what you think, so feel free to share your thoughts and comments below.