Building relationships on analytics

It’s hard not to wonder why more bankers wouldn’t leave their big-bank jobs after hearing Ben Wallace’s story. Wallace first worked at First USA, which was acquired by Bank One, and then ultimately became part of JPMorgan Chase. He ended up running the giant bank’s Consumer Technology Group out of Wilmington, Del.—a good job, but requiring a lot of travel—much of it around the world; tough with a young family. He and his wife had been mulling over a different lifestyle when a recruiter reached out to him about an opportunity at $1.2 billion-assets Orrstown Bank, based in Shippensburg, Pa., near where his wife’s and his parents lived.

Wallace joined the bank in 2013 as part of a management rebuilding following some tough years. Not only did Wallace eliminate most travel, as executive vice-president of operations and technology, his responsibilities became much broader, more quickly than would have been likely at Chase. He now oversees traditional operations, IT, and the bank’s digital and virtual channels. At the time he joined, the board took the opportunity to not only remediate the balance sheet and invest in a risk management upgrade, he says, but also to ask, “What’s the right technology for a $3 billion to $5 billion bank?” Wallace has been a key part of helping in that transformation.

The timing has been fortuitous. With six consolidations occurring around it, including BB&T’s acquisition of Susquehanna Bancshares, Orrstown Bank has been able to capitalize on the disruption.

In the following dialog, edited from an interview transcript, Wallace talks about how a community bank can enhance its traditional strength—customer relationships—by becoming data-driven.

Q1. Community banks pride themselves on knowing their customers. At a recent convention, you indicated that when you came to Orrstown Bank, you 
realized its customer knowledge 
wasn’t as good as it could be. Why?

A. There’s this perception that as community banks, we win and lose all day on our relationships—we know people by their names, go to church with them, see them in the branch a couple times a week. That may have been the model, but when you begin to look at the stats, you realize you have fewer and fewer visits into these branches. In our case, maybe 40% of the people in a given year will not visit a branch.

So then you begin to ask yourself, “If our differentiator historically has been that high-touch personal relationship, but we don’t have the physical manifestations of that anymore, how does our ability to win change? And how do we respond to that?

“When they call our call center or visit us online, how do we capitalize on it to make people still feel and believe that they’re talking to a little bit different organization than a money center bank?”

That’s the journey we’ve been on the past three years: creating a system using customer data and analytics to make money and grow relationships.

We want our agent on that telephone that greets the customer not only to potentially know them from the community, we want them to know what they’ve been doing with the bank, how many relationships they have. Many front-facing folks don’t have 20 years of banking under their belt, so it’s hard for them just to say to somebody, “Hey, I’d like to earn more of your business.”

So we built a customer engagement score. We’ve given it to our customer service center people and said, “When someone calls, you can at least begin to use these four or five data points along with how much the engagement score has changed over time.” And also, “Here are some of the things you may want to talk to them about because of what we understand about them.” We’re finding that giving staff data and the tools to use it is extremely helpful because they feel a little bit more empowered.

Q2. What does the engagement score consist of? How do you use it?

A. We score deposit relationships, lending relationships, and wealth relationships. Let’s say a customer has four accounts with us. On those four accounts, we measure “recency” and frequency, and take monetary measurements. Recency is how many days since the last customer-initiated transaction occurred. Frequency is average transactions per month based on a window of six months. And then monetary is what’s the average transaction amount of those debits and credidts over the same period.

So all this tells you that maybe you have a fairly modest borrower, not a whole lot of means, but uses the card every day at McDonald’s. Or you may have someone who uses the card to buy gas for their boat—so that customer gets a different score than the other.

At the end of the day, we arrive at a score of zero to nine for the three buckets—deposit, lending, wealth—nine being the most wealthy. We aggregate that across three buckets so the score could be between zero and 27. And that score is used to engage the customer. It’s fairly basic, but it’s so much more powerful than trying to have someone look at a screen and think, “Well, I see they have three deposit accounts and a loan, but that’s all I really know about the customer.”

For a call-center employee, we present the scores and the trend on their screen. So if the customer was a “four” six months ago, the rep sees how they’ve changed over time. They also see a bar chart of where they are in deposits, loans, and wealth. So at one quick glance, they may see, “This person’s a 20—he’s a really engaged customer across more than one line of business.” Not only do they see he is a VIP customer, but they may decide to not try to cross sell him a debit card. All that goes into the scripting we prepare for them and the training. We try to keep it very basic—score and trend for each business unit.

We want it to be a discussion and to make sure that we preserve that community bank feel. Frankly, I suspect customers think, “You have all this data about me, at least act like you know who I am.” I certainly do when I call a business.

We don’t have any 27’s by the way. If a customer is a 24, that’s a decent proxy for a profitable customer in most cases. At present, we don’t do things like transfer pricing, and we don’t shoot out a dollar number.

We’re currently working on applying the score data to mobile and online. It will be much more of an inline offer concept, along these lines: “We know you’re very engaged; here are a couple of offers that may be of interest.”

Q3. Even community banks have silos. How do you get around that in building this relationship database?

A. Any bank that has some retail business will really struggle without some type of relationship system. Once you embrace the concept of understanding customer engagements and monetizing them, there has to be a framework. In the age of Salesforce.com, this technology has been somewhat commoditized. We’re using a fintech solution provided by Race Data.

We aggregate every relationship we have with a customer in one system, so if they call, I can tell what relationships they hold with the bank at any given time. That’s where this all has to start, and you can use it to drive marketing or specific offers. Also, you’re going to need some framework to capture engagements and to trend it as the regulatory environment matures. The Consumer Financial Protection Bureau, for example, is expecting us to track complaints and correspondence with customers.

Often, I hear from other banks that the wealth team uses one system, the lending team uses another system, and never the two shall meet. But we’ve been pretty prescriptive that any referrals go in this system. It’s not perfect, but the power of getting one uniform system is really impactful.

To make sure it happens, we drive compensation and branch scorecards out of it. We tell people, “If you want to get paid for a sales referral goal, the only referral system we have is this one.” So the more you can tie the system to normal business processes, it takes care of itself.

Q4. You’re using a fintech solution. Generally speaking, do you view fintech companies more as potential partners or as competitors?

A. It depends on which one, but we’re pretty bullish on the fact that fintech firms can be good partners and not really a competitive threat.

A lot of our friends are now working at these companies, and in speaking with them, we find that lending and payment start-ups need banks more than maybe they recognized at the outset, right? They may have technology, but they don’t have customers, trust, or brand loyalty.

For the right opportunity, they can be really good partners. However, it’s still early and many of them have solutions that are only 70% there—they’re just not quite operationally ready for a bank yet. But without a doubt, some of them have better technology than most banks because they’ve had so much money to pour into it.

Q5. Speaking of technology, what 
did you have to do in terms of IT infrastructure to get your relationship program up and running?

A. Much of the heavy lifting was in the extraction and management of the data from our various core systems of record—working to redact and cleanse the data and send it over to our fintech partner. Race Data aggregates the data on its platforms and runs some of the algorithms to develop the scores.

We think many banks can do the same—they’re probably already taking extracts to their financial systems. Then it’s just a matter of what do you do with that data.

The work isn’t insignificant, but because of the challenges we had in 2013, we invested in talent that helped us do it. Back then we asked ourselves, “What business are we in? Are we going to run data centers and servers?” We concluded that we would never be able to run a data center as efficiently as Chase can—certainly not as efficiently as Amazon or Microsoft. So today, our bank has an insignificant number of physical servers—we’re very, very light. Jack Henry [& Co.] hosts our core platform. We have systems with Amazon and at other vendors.

I’d much rather use our technology horsepower on the stuff we talked about before—how you do algorithms around data and customer information versus keeping a system up and refreshing it every three years.

Q6. Where are you now in terms of becoming a “data-driven bank” and where do you want to be in five years?

A. The first step was to gain a more complete view of the customer. We got that done, and over the last 18 months, we’ve been rolling out the customer engagement score. Phase two will be, “How do I use the basic data around engagement to actually anticipate customer behavior?”

I think it’s clear we’re going to have to continue to invest in this space as people engage more and more with our online channels, and fewer and fewer people walk into the branch. If we don’t, we’ll find we have a sea of people using our online systems and won’t know much about them. If all we are is a commodity banking provider, we’re going to lose as switching banks becomes easier.

Q7. Apropos of that, you have 22 branches. How do you see the role of the branch evolving?

A. It’s funny because when we speak with millennials, they always say, “While we may never go into a branch, we still like to know we can.” They’ve seen their parents go through the Great Recession, and they want advice on saving and planning. So I think the local branch still has a tremendous impact on customer acquisition in certain markets.

Like other banks, we’re retrofitting our branches to be less reliant on teller-driven transactions and more on engagement, which gets back to the whole idea of the engagement score—making sure that when people do walk in, it’s a much different dialog.

Today, most people come into branches—at least ours—to transact, engage for a specific purpose. Very rarely are they coming in just to explore. Yet people think nothing of going to an Apple Store and playing around with an Apple device for 15 minutes. Wouldn’t it be wonderful if people approached their financial lives like that—come into a bank and say, “Are we doing everything right?” Maybe in time we’ll get closer to that.