4 minutes
To unlock the full potential of data, the industry must demystify the discipline and establish strong data cultures.
Data science is often confused with other disciplines, like data engineering. Knowing the difference is important because it helps credit union leaders more confidently take the first steps on a journey to data transformation.
So, what is data science? To be fair, there’s a reason it’s so often misunderstood. Data science is a broad, interdisciplinary field, drawing from many other scientific disciplines to inform all kinds of strategic pursuits. However, the field becomes easier to wrap our minds around when we focus on the singular thing all data science projects have in common. That commonality is the strategic objective to extract actionable insights from large collections of data. It’s not so convoluted when we look at it through that lens.
Data engineering, on the other hand, is a discipline focused on the tactical elements of data science, such as building and maintaining databases.
Once credit unions have base knowledge of data science as a discipline, they can begin to explore the strategic use cases that make the most sense for their overall business objectives. Of course, it’s not all that easy. There are several common stumbling blocks credit unions find themselves confronting as they pursue data science projects. Here are a few we’ve encountered in our work.
More than Just Models
A common confusion we often encounter is the idea that data can only bring value in the form of predictive analytics (i.e., machine learning models that forecast future trends or behaviors). Yes, using data to predict the future is a possible and valuable goal, but it might be quite a way down the road before a credit union reaches this level of data maturity. There are foundational building blocks that must be set to ensure data quality, completeness and compliance within any application.
For credit unions that are early in their data journeys, it is possible to start with data-driven actionable intelligence to solve use cases for here-and-now problems like hiring, marketing or helping members in financial distress. Credit unions can create a culture of making decisions to improve processes based on data evidence. For example, identifying and sorting ACH data during the pandemic helped credit unions find and reach lower-income members who had lost their jobs or were struggling to meet their debt obligations.
It’s a Team Sport
Another common fallacy among young data organizations is that data science is accomplished with a single team that defines projects and design solutions in isolation. When this misconception is executed, an organization misses out on a critical element of success—every data application must be tied to business needs. Having a full view of organizational needs is really difficult for a team functioning only within its particular silo.
Regardless of how well-trained or highly qualified a team of data scientists is, that team can only implement solutions for different departments if the data scientists interact and learn from other functions. They must understand their needs and what they wish to achieve with data. It is paramount that data science operates as a “humble practice” that walks alongside other functions towards success.
Recent research and anecdotal evidence point to huge potential for data science within the financial services space. Data-driven solutions are increasing revenue through improved fraud detection, increased product sales and the identification of new operational efficiencies, particularly in lending, insurance and wealth management. It stands to reason that these applications would be improved by data science. That’s because until now, the financial industry has largely performed them with great manual effort and processes.
Culture Is Key
Lastly, and perhaps most importantly, the biggest challenge of successfully implementing data science within a young data organization is the lack of a data culture. Certain leaders understand the potential and want their organizations to be data-driven, yet they are not quite clear on how to get there. Typically, this is because the full weight of the organization is not behind data transformation. All parts of the credit union need to be exposed to the potential of data. This is where establishing a data culture comes in. In organizations with a strong data culture, all team members are jointly responsible for generating, caring for, improving and utilizing data.
When a credit union has fully unlocked the potential of data, multiple departments are powered by data-driven, actionable intelligence, and leaders begin to see a beautiful process unfold. Employees own their metrics and use data for improvement; they give feedback on how to better utilize data; they request the collection of new variables that are key to the mission; they understand and participate in the creation of solutions that will improve their ability to serve members well. At this stage, humans are informing the data machine as much as the data machine is informing humans.
Demystifying data science patches communication gaps that can exist between data teams and the rest of the credit union. It opens new possibilities for the business to make the most of its data, which in the end, is all about improving the financial lives of people who are counting on their credit unions for personalized plans and smart daily decisions.
Vera Aiken, Ph.D., is product manager for data strategy at CUESolutions Platinum provider AdvantEdge Digital.
