It’s no secret that the need for low-cost, sticky deposits is one of the most pressing issues in the industry today, particularly as interest rates are set to keep rising. The use of deposit scoring can help banks identify which customers they want to keep — and the best ways to keep them.
Scoring, which uses an algorithm to summarize customer behavior, has long been a mainstay for credit-card portfolio management. Novantas believes these techniques also can help banks target their most valuable deposit customers based on shopping frequency or propensity, rate sensitivity and balance persistence or retention.
Banks can analyze the score to identify which customers are most prone to rate shop, along with how much money is likely to move in line with rate changes. The scores can also help identify balance retention or persistence, as well as balance run-off.
The elasticity of deposit pricing traditionally has been modeled within segments that reflect products, markets, balance tiers and potentially other macro factors. A bank’s ability to optimally price and maximize spread, while continuing to meet deposit goals, was constrained only by the granularity of segmented solution.
But those traditional methods are now being turned upside down due to the growth of online and mobile channels, and non-bank competition. Customers who have drifted away from brick-and-mortar branches can easily shop and compare prices across banks and geographic regions, choosing to deposit funds in a bank that has the best value proposition for their needs.
Banks will find themselves at a distinct pricing disadvantage unless they can capture the individuality of these decisions (see Figure 1).
IDENTIFYING RATE-SENSITIVE CUSTOMERS IS KEY
The good news is that banks can manage the back book by determining which customers are sensitive to price or those who are swayed by non-price treatments and communications. Customer-level information that is built on a library of treatments and is honed through a rigorous and continuous test-and-learn process will define best-in-class deposit book management. As more banks develop these capabilities, the ability to harness customer data for modeling and managing deposits at the customer level will only grow in importance.
For example, some of the least valuable customers are chronic rate shoppers — customers who respond to a rate offer, deposit money in the bank and demand a better rate as soon as the rate promotion ends, withdrawing their money if they aren’t accommodated. These customers can only be propped up by rate.
In contrast to chronic shoppers, episodic rate shoppers are driven by events and attitudes. (Individual personality plays a large role, too.) A tax refund, for example, triggers rate awareness and new shopping behavior from a consumer who seldom if ever, follows price trends. Once an acceptable rate is found and the money is deposited, the episodic customer’s rate awareness often disappears; the money is out of sight and out of mind. The life or persistence of these deposits reflects the behavioral tendencies of these customers and doesn’t require on-going rate support. The lack of rate sensitivity and persistence of episodic shoppers makes them much more valuable to the bank than their chronic shopping counterparts.
WHAT DOES THE CUSTOMER DO WITH THOSE DEPOSITS?
Novantas has studied customer tendencies, developing a library of behavioral metrics to capture the underlying dynamics of customer-level deposit management. Key to these scores is understanding how customers use their deposits — how much is dedicated to monthly cash flow management (funds that aren’t in play because they are used to pay bills, etc.) and how much excess cash flow is held in liquid products such as checking, savings and money market accounts. Novantas refers to the liquid funds that are in excess of cash flow as “discretionary liquid balance (DLB)”, meaning they are deposits that aren’t earmarked or committed for a specific purpose and can be easily accessed and moved. The specific accounts where these deposits are held are less important than the amount retained by the customer and its variance over time. These DLB metrics serve as the basis for defining modeling criteria for scoring.
Novantas’ score development program follows best practices from credit scoring and governance as outlined by the OCC. Like all behavior scores, historical customer level data (three-years minimum) is required to define points of observation (time periods used for score development) and performance (time period used to define outcomes to be modeled). While all the scores rely upon DLB, there are different views of the metric required for each.
MAKING THE RIGHT CHOICES
A bank needs to figure out how to identify and target rate-sensitive shoppers with the right approach. Is it a give-to-get proposition that acknowledges the customer’s banking relationship, while shifting the value proposition from rate to other non-rate rewards such as credit card-related merchant offers? Other options include launching a targeted acquisition program, providing more support to employees who negotiate rate with customers, or improving the pricing of the back book.
While deposit scores can provide a considerable amount of value, Novantas doesn’t believe a bank should just abandon its existing deposit management framework. The ideal score implementation takes place in conjunction with a bank’s existing deposit-management dimensions. This provides continuity with past management efforts and allows a bank to test and learn its way into customer-level pricing treatments.
Director, New York