John Thuma | Data Solutions Group, FIS
May 04, 2020
As a banking executive, did you ever wake up and say, “I have to have data science?” It’s possible, but I bet you really wanted to learn how data and analytics can solve an attrition problem or help with next best action. What we really want is evidence-based prescriptive guidance that seamlessly works with our institution. Then there is the cost chasm to contend with. Many smaller financial institutions thought they couldn’t afford data science or didn’t know where to start. At FIS, we understand the need to democratize analytics and make it easy for anyone to use. We help our clients use data to optimize business process and understand that it takes more than performance indicators and business intelligence to get the job done. Let’s ‘drill’ into some of these topics.
Data science and analytics are expensive projects. According to Glassdoor, the annual average US salary of a data scientist is close to $120,000. These projects require more than just data scientists. They require storage, hardware, software, governance and a variety of other non-technical human talents to manage and operationalize. These costs are expected to rise as the appetite for solutions is outpacing resource availability. In an EY study on the future of finance, 57% of respondents noted that building skills in predictive and prescriptive analytics is critical for the future. Large financial institutions have a huge advantage over small to midsized enterprises based solely on these costs alone. How can the smaller institution compete? This is where FIS Ethos becomes a powerful ally for its clients. FIS works with thousands of financial institutions and has the ability to negotiate solutions at scale. This means that data science projects and capabilities are more affordable to the institutions that use Ethos.
Most data and analytic solutions deal with facts. A phrase commonly used for these facts is called a “Key Performance Indicator” or KPI. KPI’s measure things like “Net New Cardholders,” “Gross Profit,” or “Customer Acquisition Costs” (CAC). These facts are important for measuring the business performance of a financial institution. But they’re not enough. What we need are different ways to measure intent and activities of our account holders. To do this we need to get emotional. Our data includes measurements that enable FIs to measure qualities – not necessarily quantities - of people. Things like behaviors, emotions, and lifestyle indicators, or KBI’s, KEI’s, and KLI’s respectively.
These types of measures l lead to better individualized marketing campaigns and audiences, bank product development, fine-tuned segmentation, and optimal customer acquisition/growth and attrition management. Don’t get me wrong, “Net New Accounts” are still important, but getting emotional is the real key indicator.
Descriptive analytics, or business intelligence (BI), has been around for more than 25 years. These solutions are valuable, but new capabilities are a must. A valued partner of FIS Ethos has identified some significant trends in the Business Intelligence market.
First, integrating machine learning is must. Our customers want to know more than what happened, they want to know why something happened and what is going to happen next. Data science techniques, like attribution and predictive analytics seamlessly, must be integrated into the modern business intelligence systems The interface needs to be humanlike and real-time. Natural language processing (NLP), or human language query, is a critical component; NLP is not just a feature but the future. Real-time solutions are changing the data integration landscape and business intelligence must follow that lead. People don’t want to wait a day for transactions to flow into their reporting. Roy Schulte, vice president and distinguished analyst at Gartner states: “End users can harness increasingly sophisticated analytic capabilities through packaged real-time analytics embedded into data discovery tools and applications without prohibitive processing wait times or the need for developers to intervene.”
Data science and advanced analytics provide predictions and other statistics that are often used to optimize business. But are statistics what people really want? People want evidence-based determinations of what to do next, not given a statistic. Data science projects don’t succeed for many reasons: we can’t get access to the data, the subject matter experts are not available, or leadership just doesn’t care. How do we overcome such roadblocks? Yes, you need data, people, and leadership support to get that fixed. Once you do, the biggest mistake you will make is a platform decision looking for a problem to solve. Instead look for a problem to solve and determine who or what it hurts and how best to communicate with them. The next step is to figure out a seamless way to integrate the solution into the organization. Since, for the most part, humans are still in charge, you must communicate with language and syntax that a human can understand.