Making Strategic Decisions With The Help of Data Analytics

Banks capture a variety of data about their customers, loans and deposits that they can harness in visually effective ways to support strategic decision-making. But to do this successfully, they must have leadership commit to provide the funding and human resources to improve data collection and management.

Bad data or poor data quality costs U.S. businesses about $3 trillion annually, and breeds bad decisions made from having data that is just incorrect, unclean, and ungoverned,” said Ollie East, consulting director of advanced analytics and data engineering at Baker Tilly.

Companies generally have two types of data: structured and unstructured. Structured data is information that can be organized in tables or a database: customer names, age, loan balances and interest rates. Unstructured data is information that exists in written reports, online customer reviews or notes from sales people. It does not fit into a standard database and is not easily relatable to other data.

If data analytics is the engine, then data is the gasoline that powers it,” East said. “Everything starts with data management: getting and cleaning data and putting it into a format where it can be used, governed, controlled and treated as an asset.”

A maturity model for data analytics progresses from descriptive to prescriptive uses for the information. The descriptive level answers questions like, “What happened?” The diagnostic level answers, “Why did it happen?” The predictive level looks at “What will happen?” Finally, at the prescriptive level, a company can apply artificial intelligence, machine learning or robotics on large sets of structured and unstructured data to answer “How can I make it happen?”

Existing cloud-based computing technology is inexpensive. Companies can import basic data and overlay a Tableau or similar dashboard that creates a compelling visual representation of data easily understood by different management teams. Sean Statz, senior manager of financial services, noted that data visualization tools like Tableau allows banks to create practical visual insights into their loan and deposit portfolios, which in turn will support specific strategic initiatives.

To do a loan portfolio analysis, a simple extraction of a bank’s data at a point in time can generate a variety of visual displays that demonstrate the credit and concentration risks. Repetitive reporting allows the bank to analyze trends like the distribution of credit risk among different time periods and identify new pricing strategies that may be appropriate. Tableau can create a heat map of loans by balance, so bankers can quickly observe the interest rates on different loans. Another view could display loss rates by risk rating, which can help a bank determine the real return or actual yield it is earning on its loans.

Statz said sophisticated analytics of deposit characteristics will help banks understand customer demographics, and adjust their strategies to grow and retain different types of customers. Bank can use this information in their branch opening and closing decisions, or prepare for CD maturities with questions like, “When CDs roll over, what products will we offer? If we retain all or only half of CD customers, but at higher interest rates, how does that affect cost of funds and budget planning?”

Data analytics can help banks undergo more sophisticated key performance indicator comparisons with their peers, not just at an aggregate national or statewide level, but even a more narrow comparison into specific asset sizes.

Banks face many challenges in effective data analytics, including tracking the right data, storing and extracting it, validating it and assigning resources to it correctly. But the biggest challenge banks need to tackle is determining if they have the necessary data to tackle specific problems. For example, the Financial Accounting Standards Board’s new current expected credit loss (CECL) standards require banks to report lifetime credit losses. If banks do not already track the credit quality characteristics they will need for CECL, they need to start capturing that data now.

Banks often store data on different systems: residential real estate loans on one system, commercial loans on another. This makes extracting the data in a way that supports data visualization like Tableau difficult. They must also validate the data for accuracy and identify any gaps in either data collection or inputting through the system. They also need to ensure they have the human resources and tools to extract, scrub and manipulate essential data to build out a meaningful analytic based on each data type.

The key to any successful data analytics undertaking is a leadership team that is committed to developing this data maturity mindset, whether internally or with help from a third party.