Credit unions interested in advancing their data analytics efforts will find a wealth of information in a recent article in the McKinsey Quarterly. Simply entitled, “Making Data Analytics Work for You – Instead of the Other Way Around” (Mayhew, Salah, and Williams), the article provides an easy to follow list of steps for any organization to get the most out of their investment in data analytics.
The authors emphasize that improving corporate performance is the only meaningful reason for organizations to pursue data analytics. As a result, they state two important principles:
- The data used for analytics must be “purpose-driven”. That is, data points used in analytics must be chosen carefully to ensure they align with high value processes within the organization.
- The outputs of data analytics must drive action. To do so, analytics must operate across organizational silos, represent realistic scenarios, and most importantly, is actually used to drive decisions.
With these guiding principles, the authors suggest eight steps for achieving effective use of data analytics.
- Ask the Right Questions
Asking vague questions such as, “What patterns exist in the data?” are not an effective way to approach data analytics. Rather, it is better to start with specific use cases. For example, a questions like, “How can the loan application process take fewer days without sacrificing quality?” not only pertains to a real business issue but it also provides analysts an area in which to focus.
- Think Really Small… and Very Big
It is a myth that all innovations uncovered by data analysis are blockbuster “game changers” that revolutionize the organization. Of course, this could occur, but a more likely means of boosting organizational performance is to find many small areas for improvement. If a credit union has data available for analysis at the transactional level, analysts can explore processes at their smallest component parts. Finding opportunities in many small areas can deliver an overall large reward.
- Embrace Taboos
One area that is often “taboo” among a credit union’s data sources is unstructured data. This is data not stored in rows and columns such as free form text captured by Member Services Representatives during phone calls with members. While the data may have value, credit unions often have little experience making use of it because it is perceived as being “low quality”. The authors observe that there may be big insights within this data and organizations need to develop a way to understand it. One suggestion they offer is to build a “data provenance model” that evaluates all data sources and assigns a “score” for each source in terms of reliability. Having this perspective allows unstructured data to be viewed in context and thereby used accordingly rather than being ignored.
- Connect the Dots
Data silos are a major organizational issue that data analytics aims to rectify…if it is allowed to. If data is locked up by functional area, some major insights might never be discovered. For example, if a credit union’s loan origination data and loan servicing data are not integrated, lending lifecycle analyses are not possible. Data simply must be integrated across the organization for data analytics for reach its highest potential.
From Outputs to Action
- Run Loops, Not Lines
The authors recommend the OODA (observe, orient, decide, act) loop as a way to make data analytics iterative and thereby more effective. Many credit unions are early in the data analytics journey. Taking the iterative OODA approach will result in a more reality-based effort that is by its very nature customized to the credit union’s unique situation. It is an approach that emphasizes learning and continuous improvement in the quest for actionable insights.
- Make Your Output Beautiful and Usable
Data analytics without compelling visualization is like a talented actor without makeup. Unless the makeup and costuming are just right, the underlying talent is less appreciated. Analytical output achieves its highest impact if it is presented in a clear and interesting manner. Organizations maximize their analytics investments if they provide excellent data visualization.
- Build a Multiskilled Team
Credit unions, especially smaller ones, often have one person dedicated to the data analytics program. Yet, as the program grows, the organization needs to plan to build up the data analytics team with a wide range of skills to meet the ever-increasing appetite for information. One option for smaller (and sometimes larger) credit unions is to partner with experienced data analytics vendors who can provide the expertise needed to make the most of growing programs.
- Make Adoption Your Deliverable
A “build it and they will come” philosophy is not a good basis upon which to build a data analytics program. If fact, potential users of data analytics output should be involved with the program from the start. Business Unit leaders need to be very explicit about what problems need to be solved. This will drive not only what data is acquired but it will also drive choices in tools and training. Making adoption a top priority from project inception is essential for success.