Why and How KeyBank Has Become Big Data-Driven
The Utah bank learns about the behaviors and preferences of the Spanish and Asian populations in its footprint by mining a plethora of data sources.October 1
The San Francisco bank's CIO shares an update on an enterprise-wide customer analytics project and why CCAR is a "mixed blessing."September 30
KeyBank over the past year has become very deliberately data-driven: anyone who wants to make a decision based on business experience or gut instinct needs to provide numbers to back it up or risk having that decision overturned.
Big Data and analytics have become popular, and some believe overused, buzzwords in banking and many other industries. In some cases, this is more rhetoric than practice.
At KeyBank, it's real. A year ago, the Cleveland bank hired David Bonalle away from American Express to drive the use of analytics throughout the bank and form an analytics center of excellence. He sits on the executive committee. The marketing department of the $89 billion-asset bank reports to him. His analytics team makes business recommendations and in some cases vetoes decisions, such as a bid to keep open a non-performing branch, where the numbers show a comeback is impossible.
The impetus for making Big Data a focus for the KeyBank came from CEO Beth Mooney, Bonalle notes. A few years ago, she reassessed the bank's strategy and decided it needed better analytics and insights.
"The insight Beth Mooney had was, if we're going to serve our clients well, we need to understand what their needs are and who they are," says Bonalle, executive vice president and director of marketing and insights, who spoke at American Banker's Banking Analytics Symposium in Boston on Friday.
He was tasked with forming an analytics center of excellence to support the entire bank, a centralized team that would feed data and reports to all business lines and the marketing department.
This was a huge change for the bank. Previously, each line of business had its own analytics group and marketing had two, for a total of 13 analytics teams.
This led to multiple versions of the truth and problems. "Two months after I got there, the head of retail asked me how many accounts we were acquiring online," Bonalle says. "When I told him, he said, "Eric from your team just gave me a different answer. "That was a little embarrassing."
In addition, there was no real career path in the analytics groups and they reported to product and marketing people.
"The people in the analytics group were order takers," Bonalle says. "The business side didn't understand what they were working on or who had asked them to do it; the analytics team didn't give proactive advice. "Often, the business line leaders would ask for data, then not like the data and say it was wrong. The analytics group would re-pull data and both sides would become frustrated.
To help support the new group, the IT department consolidated the bank's data infrastructure, taking it from 13 to two data warehouses.
A goal for the entire bank became: all decisions will be based on analytics. "This is easier said than done," Bonalle observes.
Analytics projects are now prioritized. The business line executives have to clearly state why they want data and what their goals are for it. Those that align most closely with business goals go to the top of the list. Ad hoc requests that are considered mission critical are accepted. "We make sure they're mission critical," Bonalle says.
The last priority is requests for "nice to know" data. "When a business leader says, "You know what would be nice to know?" I say, that's great and we're not doing it," Bonalle says. "If it's going to drive a business decision and has a high return, we'll do it. If it's nice to know, we're not doing it."
With the analytics they produce, Bonalle's team always makes a business recommendation. "We feel we owe them an answer as to how to proceed," he says.
Instead of consensus-driven decisions, "the new assumption is whatever we do next has to have highest net present value," Bonalle says. "It's all about NPV."
For example, if a branch is unprofitable, the analytics group will spell out what it would take to make it profitable. "It used to be the head of a district would say they were going to turn that branch," he says. "We'll tell them you'd have to quadruple productivity for the next six months. As long as that's what you're signing up for, let's go."
Bonalle says the value of new projects his group has delivered in the past 12 months is $50 million, through closing branches, changing staffing, and improving marketing.
"Our marketing budget was flat from 2012 to 2013, and we are on track to produce 71% more accounts in 2013 than last year," he says. "This is a combination of changing a major campaign, a lot of test and control and getting rigorous about where we invest our dollars based on what analytics are telling us."
Bonalle offers a few "lessons learned" from KeyBank:
1. Have the analytics teams sit as high in the organization as possible.
2. Spend time aligning executives on analytics, or better yet, hire executives who love analytics. "Our head of retail is such a person, that's changed our lives," he says.
3. Focus on the "client interaction model," in other words on serving internal clients from the business lines. Be there to serve business objectives.
4. Embed yourself in the business. "Analytics people should know, live and breathe that business," he says.
5. Make developing a center of excellence a priority. Invest in training and provide career opportunities.
6. Be willing to say "no" a lot. "We were constantly committing to things we couldn't deliver," Bonalle says. "We'd have to say, we can't do that with the resources we have."
7. Never say "no." "We don't ever say we can't do something," Bonalle says. "We say, here's what we're already working on, here are the resources we have, what do you want us to do? Stop a project, delay a project, add more resources? We leave the decision up to them. We don't let them say, "do more with less." That's not sustainable."
8. Be objective. "We have to be the impartial representative of what the data and analytics are telling us," Bonalle says. "Once you lose that, it's difficult to convince executives to leverage what you have."