Banks have applied analytic models and decisioning technology to virtually every part of their business—from evaluating a consumer’s credit risk to determining the best treatment for a delinquent account. One of the few areas that lack this level of sophistication at most banks is pricing. The decision on what price to set has huge consequences for the overall profitability of the institution, yet often these decisions are made without the analytic and technological sophistication needed to ensure the optimal outcome.
How Prices are Set Today
At most banks, prices are set by a pricing committee. This committee is comprised of individuals from a variety of departments across the institution including risk, marketing, compliance, and finance. This committee meets regularly; with every member bringing their own goals, opinions, and spreadsheets filled with numbers and formulas to the table. They then proceed to determine the institution’s new prices by engaging in a long and arduous debate that Frank Bria (Banking Professional Services Director at Earnix) calls “the battle of the spreadsheets”.
The outcome of this “battle of the spreadsheets” is an agreed upon list of different prices for each product and customer segment, which is then distributed to the entire bank.
A Suboptimal Approach
The problem with making pricing decisions based on this approach is that it relies, almost exclusively, on the decision-making abilities of the pricing committee. As we have learned from countless other examples in the financial industry, human beings’ decision making processes are prone to error for many reasons. We’re biased, we’re inconsistent, and we don’t always make decisions based on the right information. Credit risk decisioning is a great example. To a human being, the way a person dresses or the amount of money they make seem like perfectly valid indicators of credit worthiness. It took analytic models like the original FICO score to show us that, statistically speaking, factors like income were not that predictive of an individual’s credit worthiness.
Given that the vast majority of bank decisions are now based on analytics and statistical models, why are banks still entrusting pricing decisions to the fallibility of human decision making?
Looking at it the Wrong Way
The following pricing formula (commonly used by pricing committees) is a great example of why this approach to pricing is suboptimal.
Price = cost of funds + cost of default + operational expenses + acquisition costs + profit margin
This formula seems logical. It takes well understood cost variables (sometimes determined using sophisticated analytics) and adds in the main outcome of the “battle of the spreadsheets”—the margin—to set the final price. The problem with this approach is that it is asking the wrong question. It’s asking what price the bank needs to charge in order to cover their costs and make an acceptable profit margin. It’s focused on the wrong variables…just like those manual credit risk decisions based on income and the color of the applicant’s tie.
A Better Way: Price Optimization
The question that banks’ pricing decisions should be based on isn’t how much they need to charge in order to cover their costs, but rather how much they can charge based on each customer’s price sensitivity.
This is a fundamentally different way of looking at pricing. It relies on analytic models to predict what each customer is likely to be willing to pay. It enables banks to continue using all of their cost variables while optimizing their prices based on those price sensitivity models, thus improving their overall profitability.
This approach is called price optimization and it is starting to be adopted by many large financial institutions— the same institutions that once took a chance replacing manual credit risk evaluations and spreadsheets with credit scores and decision engines.