How Banks Are Using Big Data to Set Deposit Rates

Worried that the Federal Reserve will start raising interest rates next year, many banks are reconsidering their strategies for attracting and keeping deposits — and turning to analytics software to help.

The software can be fed reams of information about economic conditions, competitors' rates, and individual customer behavior to help calculate rates that institutions should pay for deposits. Bank of the West, Fifth Third and Bank of Montreal are among those that have jumped on the trend, and more are likely to follow.

"It's not the type of thing where you flip a switch from one way of doing your deposit pricing and switch over 100% to relying on a tool," said Mark Erhardt, senior vice president for retail product management at Fifth Third, which is based in Cincinnati. "It's more of an evolution of how you use the tool and increasingly build it into your decisionmaking process. At every turn of the crank, we get a little better and more confident."

In December, Fifth Third began using analytics software from Nomis to analyze millions of records per month, looking at demand curves, interest rates, and deposit flows within households, aiming to find the ideal price based on the volume of deposits it wants to gather and the maximum amount of interest expense it's willing to incur.

Some consumers are very price sensitive and will move large amounts of money for a small increase in interest rate. Other customers, even offered a large increase, don't move their money. The software provides a predictive score of customers' price sensitivity, based on factors like past transaction activity, credit bureau score, and household income.

In the second quarter, Fifth Third started using the software to test various scenarios, including what the effect would be on various products if it raises rates.

"Weekly we're looking at the short-term trends, but once a month we have a more extensive deep dive where we look at the tests we're running versus the control markets," said Erhardt. "We're trying to understand, if we aren't getting the results we expected, what might be causing that?"

Bank of the West, meanwhile, has been working on what it calls "relationship pricing" for a little over two years. The bank gathers intelligence on all services a customer uses across the organization and generates customized prices using miRevenue software from Zafin in Palo Alto, Calif.

But Bank of Montreal has turned to analytics software from Novantas to aid its rate-setting process.

Martin Nel, vice president of personal products for the Canadian bank, said that it first turns to internal data sources.

"We would want to know, is the customer depositing money for a couple of days and pulling it out again?" he said. "Is it money they're saving up with the aim of building up a nest egg for retirement?"

Going forward, the bank will obtain regional data about competitors' prices (it already has some of this data). In its calculations, it considers the liquidity of the deposit, the customer's profile, competitors' rates, and other relationships the customer has with the bank.

"All these things produce an enormous amount of complexity," Nel said.

The pricing software is meant to give the front-line staff the right information to make pricing decisions. In Canada, deposit pricing is a negotiation, a sometimes painful one. Nel hopes the software will make it less painful.

"People sometimes think this is to take advantage of customers," said Darryl Demos, head of Novantas Solutions. "That's not the case here. This is about giving customers the best value proposition. Sometimes generic policies miss the subtleties of the financial relationship."

The advantage of using analytics software to help with decisions is that the software can pore through a lot more data than would have otherwise been possible with humans looking at spreadsheets or mainframe data.

"It's very hard to look at trends over time, or trends in one market versus another, without an analysis tool," Fifth Third's Erhardt said.

When he goes into meetings, he said he has more in-depth analysis, with more trending and granularity, than in the past.

But "we'll never get to 100% of pricing run by the decision engine," he said. "It's really better decision making based on better information."

In the past it might have been hard to determine whether a change was happening because of something the bank did or something related to a macro environment change.

"Now we're able to isolate a specific change and compare it to a test market versus control market and also compare it to macro trends," Erhardt said. "Once the competitive market heats up, the speed and accuracy of the data will help us make better, more confident decisions in the future."

In embracing analytical solutions now, the banks are trying to get ahead of the Fed's next move. Fed Chair Janet Yellen said in March that the central bank is likely to begin raising rates "about six months" after it winds down its bond buying programs. That would indicate that rates could start going up in March of next year.

"Obviously we want to be ready," said Bank of Montreal's Nel.

Many Canadians keep their money in liquid accounts, and are likely to migrate to "term investments," a Canadian product similar to certificates of deposit in the U.S.

Erhardt acknowledged that right now, most customers "are not yet shopping around."

But the overall low rate environment is a good time to test the new analytics software while it's still relatively quiet.

"We are seeing more banks do what we're doing, which is test more aggressive offers on a limited basis and see how that goes, then continue to roll with it or pull it back and rework it and roll it out again," he said.

For reprint and licensing requests for this article, click here.
Bank technology
MORE FROM AMERICAN BANKER