How to Mine Data for Gold

Cross-selling and targeted marketing are just two of the benefits of using your community bank’s data effectively

By Howard Schneider

Exciting times are ahead of us,” predicts Boris Sestan, data analytics supervisor at the $1.2 billion-asset Colonial Savings in Fort Worth, Texas. He says it was “a natural progression” for Colonial to get into data analytics as it became aware of the sheer amount of customer information contained in its systems.

Analyzing all the data a community bank holds can help automate certain proven banking practices, such as when an agricultural lender includes crop price forecasts when underwriting loans.

But data analytics is about more than putting together spreadsheets for individual loans. In its most inclusive sense, it means using information from both inside and outside the bank to boost sales, streamline services and create more targeted marketing. That said, developing expertise in data analytics can be tricky for institutions. Customer information is often spread across a number of different databases within a bank, and the formats may be incompatible with one another. This is because many banks have several legacy databases created in different formats, along with reports generated from that information using different types of “bolt-on” software.

Take it off your plate
Data analytics service providers do much of their work in this area. “The bulk of our efforts is getting data into shape to do analysis,” says Michael Le Lion, global client and partner director at Datawatch Corp., headquartered in Bedford, Mass. But once data are accessible, clients can perform their own analyses. “You can start small and grow,” he says.

Sestan agrees that accessing bank data is challenging. “It’s very cumbersome,” he says. “There are bits and pieces all over the enterprise system.” But he agrees with Le Lion that it’s not necessary to use third-party software to practice analytics. The key, he notes, is to first formulate a business question and then find the data that answer it.

Say a community bank wants to increase personal lending. How can it most effectively identify customers who are in a position to take out a loan? Sestan says existing mortgage data can be invaluable here. “There’s lots of information on a mortgage application,” he says. Knowing a borrower’s income, credit score and debt-to-income ratio lets Colonial, which is primarily a residential mortgage lender, target qualifying homeowners with personal loan solicitations.

Data analytics can also help to streamline marketing. At Colonial, mailings are produced automatically for recent property buyers who may need funds for home furnishings or improvements. Rather than budgeting for mass advertising, community banks can look at current borrowers to determine who has the need and the capacity to take on more debt.

Make it meaningful
While searching for meaning in all the unstructured data held in core banking systems can sometimes seem like a fruitless enterprise, keep the faith. It can provide what you need to retain and grow current business. For example, a routine debit on a customer’s account gains significance once it’s identified as a loan payment to a competitor.

A red flag like this should be of particular interest to community bankers. A study from Javelin Strategy & Research found that community banks are keeping customers who have checking accounts but are losing “secondary relationships”—such as loans—to other institutions.

Knowing more about their checking account customers can therefore help community bankers develop strategies to keep them from leaving and expand share of wallet. Analytics can spot trends revealing both opportunities and risks, says Ryan Caldwell, founder and CEO of data-analysis firm MX Technologies Inc. in Lehi, Utah. It can alert managers if a customer’s income rises or a borrower’s financial activity shows signs of distress.

Using data to reach conclusions gives bankers the opportunity “to expand their horizons,” says Mary Jones, community banking lead at SAS, based in Cary, N.C.

She notes that analytics let banks aggregate all the accounts in a household to get a complete picture of a customer relationship from a household perspective and spending patterns. Bankers can then forecast when customers are ready to buy a car or take out a mortgage, and marketers can send automated loan solicitations to these prospects.

There are compliance benefits to data analytics, too. Jones notes that examiners are asking “for more sophisticated analysis” of bank portfolios and lending activity. Rather than asking the IT department to pull information from different sources, bankers themselves can use analytics to generate reports featuring elements they’ve chosen, saving time.

Proceed with caution

Core banking data, combined with data from online and social media findings, can produce a more complete picture of customers and their financial lives, leading to next-generation marketing, such as sending a targeted text message as someone walks past a bank branch. Yet bankers “have to be careful about how they interact” with customers, Jones adds. No one wants to feel as if his or her private data are being exploited.

That said, MacDonald believes using data to make credit offers is efficient and that a data-driven approach to lending mitigates fair lending concerns by targeting only appropriate prospects.

Today, financial institutions “are absolutely overwhelmed with data,” asserts Le Lion. “It’s coming from every part of the bank.”

Seeing that information as a valuable asset, and using it responsibly and strategically, can help community bankers run their business more profitably.

The four types of dirty data

The first step in using data analysis to boost revenue and generate leads is to clean up your data. Here are four common kinds of bad data to seek out and destroy.

  1. Duplicate data. This might occur if your mortgage and personal banking teams aren’t using the same database, for example, or if there is a legacy database in a different format hanging around in your virtual attic. “To accurately determine how many unique customers your company actually has, you’ll need to de-dupe, or identify and eliminate the duplicate entries, to create a new master list,” says Michael Le Lion of Datawatch.
  2. Missing information. “Combining data from multiple sources is critical for accurate forecasting and analytics,” Le Lion says. “In addition to completing blank fields, legacy accounting and finance data must be combined with information from multiple sources for cash reconciliation, payroll reporting, labor distribution, budget planning and analysis.”
  3. Inaccurate information. “Manual hand-keying data into spreadsheets can plague analytics and have a dramatic impact on revenue or costs,” says Le Lion.
  4. Outdated data. “With billions of gigabytes of information being generated daily, data is almost instantly outdated,” Le Lion advises. “Today’s data sources are often generated dynamically, in streaming, multi-structured or unstructured formats. It is essential to access and blend this data with the existing enterprise information for rich, accurate analytics.”

Howard Schneider is a financial writer in California.

comments powered by Disqus
Top