On December 9th, 2014 I attended an afternoon presentation at the Credit Union Big Data/Analytics Conference hosted by OnApproach. This particular session, “The Denali Story: What Management Needs to Know About Analytics” was presented by Greg Nolder, VP of Applied Analytics at Deep Future Analytics (a CUSO formed by Denali Alaskan FCU) and discussed common reporting and analytics practices at credit unions. More specifically, he discussed his vision for the future of credit unions, and how they can remain competitive in an ever-changing industry by adopting predictive analytics strategies.
Before discussing predictive analytics, I want to first discuss five main types of analytics, and how they can be implemented differently at credit unions:
- Descriptive Analytics – Descriptive analytics is the most simplistic form of analytics a credit union (or any organization) can utilize. Descriptive analytics takes large data sets, commonly referred to as big data, and looks at what has already happened. Rather than trying to learn from the data and make predictions about how strategy can be altered, it aims to summarize the data. For example, a credit union can look at the average yield of their loan portfolio. Descriptive analytics can be also referred to as reporting, a practice already carried out by most credit unions today.
- Exploratory Analytics – Exploratory analytics, much like it sounds, refers to the process of exploring the data collected and attempting to formulate hypotheses that could lead to new data collection and experiments. Whereas descriptive analytics looks at what happened, exploratory analytics takes it one step further and asks why it happened. In the same example, a credit union can look at their average yield and ask why it turned out to be higher or lower than expected.
- Inferential Analytics – Inferential analytics takes it one step further and begins to test the hypotheses and experiments formulated in exploratory analytics.
- Predictive Analytics – Predictive analytics harnesses patterns found in historical and transactional data to identify risks and opportunities. Through utilization of sophisticated statistical modeling techniques, machine learning, and data mining, predictive analytics looks at past and present facts to make predictions about future events. Predictive analytics allows credit unions to look at the same loan portfolio discussed earlier and apply statistical models to affect the outcome of their future yield.
“One of the most well-known applications is credit scoring, which is used throughout financial services. Scoring models process a customer’s credit history, loan application, customer data, etc., in order to rank-order individuals by their likelihood of making future credit payments on time.” – Nyce, Charles (2007), Predictive Analytics White Paper, American Institute for Chartered Property Casualty Underwriters/Insurance Institute of America, p. 1
Even with predictive analytics is important to remember the commonly quoted phrase, “correlation does not imply causation.”
- Causal Analytics – Causal Analytics is one of the most desirable and one of the most difficult forms of analytics. The vision of causal analytics is that the cause of an event can be directly correlated to specific input variables. This takes predictive analytics one step further. Rather than just predicting the future, causal analytics tries to understand the mechanisms that directly influence future outcomes so that forecasts can become even more accurate. This type of analysis opens the possibility for management to make decisions that improve those outcomes and organizational results.
Analytics presents an incredible opportunity for credit unions but the key is understanding what type of analytics is most beneficial. While all analytics are beneficial, descriptive analytics (reporting) provide the least amount of value. Predictive analytics provide the most obtainable value for credit unions. When considering their Big Data/Analytics journey, credit unions need to think beyond reporting and start thinking about the future, predictive analytics.
This first step in any analytics journey is getting to the right data to perform the analytics. For credit unions, with many disparate data sources, this is often the hardest task. Once data integration is obtained, predictive analytics can be easily executed by credit unions, allowing them to stay competitive with even the most goliath competitors.
This article was originally posted on CUinsight.com on December 18th, 2014 by Austin Wentzlaff