Consolidation and increasing competition-some say hyper-competition-in the banking industry expose an absence of product differentiation. No one can tell the difference between one bank's products and another's.
Among the few points of differentiation banks have left are their existing customer bases. Banks should explore how to maximize the value of their customer relationships. This is best achieved by effectively managing customers across their life-cycle to maximize retention, stimulating more transactions and revolving balances, cross-selling more bank products, and minimizing risk in the process.
More and more, banks are grabbing for every possible source of information to better understand and predict their customers' behavior. But there are several data sources that banks typically do not use to their potential: detailed transaction data, merchant names, customer service comments, and collections comments.
Even though this kind of information is right under their noses, banks are just starting to make use of it. They will find it most helpful in predicting fundamental changes in the customer life-cycle.
Banks can use these profiles to understand their customers better and drive business growth.
For example, a large U.S. credit card issuer used detailed transaction data to build a suite of models to predict credit risk and profitability. The bank implemented this in a system that ran nightly on all the authorizations of the day. It then used transaction risk and profitability scores to approve or deny authorizations, change credit lines, and take early collections action.
The issuer's bottom line improved by $5 per card per year.
Other issuers have used transaction data to predict attrition, finding that customer runoffs can be foretold by shifts in transaction behavior. When you no longer are your customer's card provider of choice, your balances are at risk.
Combining an attrition score with another transaction-level score for profitability, an issuer can construct highly focused offers to keep the accounts most worth saving.
Other data sources that banks are starting to use are text descriptions of merchants and payee names on checks. Where your customers spend money is perhaps the best indicator of their lifestyles and life stages. New text analysis and text processing technologies let you base predictions and segmentation on free text and other unstructured data.
One issuer used the text description of merchant names to match merchants and cardholders. Then it segmented the portfolio according to the merchants that the cardholders did business with. This allowed for accurately targeted coupon offers for the merchant partners.
Merchant names carry information about cardholders' unique affinities, such as:
Travel and entertainment patterns (specific hotels, airlines, rental cars, and restaurants).
Hobbies (model trains, off-road activity, musical instruments, etc.)
Service and prestige sensitivity.
Lifestyle and life-stage identification.
Another issuer uses a transaction-based attrition model and plans to use an affinity program as a retention device. When a cardholder is at high risk of bolting, as indicated by his or her transaction-based attrition score, the issuer intends to call that person and attempt to cross-sell an affinity card.
The issuer would pick an affinity group appropriate to the cardholder's demonstrated merchant preferences. Text-analysis technology would help it pick the affinity-group offer.
These new data sources go well beyond a pure repricing decision aimed at retaining a cardholder. Instead, a unique offer can be tailored.
For example, a loyalty card up-sell program can focus on cardholders likely to leave, and offer a platinum loyalty card with a reward that matches transaction behavior.
In card cross-selling, existing bank customers can be offered loyalty cards that match their check-writing behavior, gleaned from payee names on the checks.
Deals can be struck with merchants on the basis of behavior patterns discerned from a bank's existing card base. Rewards can be offered to loyal cardholders and the merchants.
Text data in the customer-service comments and notes fields can carry further useful information about customers. These comments can help you assess customer satisfaction and thus predict attrition.
They can indicate the cardholder's price sensitivity and thus improve targeting in repricing and retention programs. They can identify fee- sensitive and dispute-prone segments. They can even tell you what competitors are doing-your customers will tell you which card they are switching to and why.
Text data relating to collections carry information about the cardholder's integrity and life situation - "The check is in the mail" versus "I'm in deeper than I planned and I want to work out a payment plan."
A cardholder saying "I was on vacation in Europe for the last two months and I forgot to pay my bill" is very different from one who says "My wife usually pays the bills, but we just separated."
The emerging sources of customer information and text analysis can be valuable in risk management, marketing collections, and customer service. They are easily accessible and are already showing how they can provide a distinct competitive edge. Banks that can understand and predict their customers' behavior know what to do to keep their business.
