Insight. Innovation. Industry.
Business / Technology /  •

Turning data into dollars

Credit unions need to up the ante on analytics to capitalize on what members want

Elizabeth Duff has some simple, yet effective, advice for credit unions looking to tackle big data projects: bring your chequebook, and be flexible.

“If you are embarking on this path, you have to be open-minded about how you are going to use [data] and not be afraid of it,” warns Duff, CFO at Newfoundland and Labrador Credit Union (NLCU), which has 21,000 members and more than $540 million in assets.

Duff saw firsthand the power of harnessing data when her financial cooperative undertook a member segmentation project in 2012. But it didn’t start out as such. Originally, NLCU’s management discussions focused on launching a loyalty program to reward its members, but that discussion morphed into a conversation about “How do we even know a loyalty program is important to our membership?” Duff told Enterprise in an interview.

So NLCU hired a market research firm to find out more about what its members wanted and needed, how they felt about the credit union, and their level of financial literacy.

Instead of focusing on the more common “what questions” researchers asked the trickier “why questions,” measuring psychographic characteristics of individuals and things like personal and social values and lifestyles.

That initial focus group testing prompted NLCU to advance to the next stage, creating a customer segmentation survey to break members into subgroups so that the credit union could better align its services with its members’ needs.

Through its efforts, NLCU identified six distinct life stages through which its members travelled. It formed subcommittees to translate the research into action, and over two years those committees devised and implemented projects designed to enhance member experience.

NLCU then embarked on what became known as the “Big Embrace,” a cultural shift to move from being a credit union that was product focused to one that is member focused based on that customer segmentation framework. “We are in an evolution and we have evolved,” says Duff.

NLCU’s experience is indicative of the change that is underway in the credit union system, as more financial institutions embrace data projects to drive their business forward, says Linda Young, a B.C.- based management consultant who highlighted NLCU in her study: Rightsizing Big Data for Credit Unions.

Building on prior research

Young’s study, conducted for the Filene Research Institute, follows up on the organization’s earlier work: Big Data and Credit Unions: Machine Learning in Member Transactions. (See “Riding the big data wave,” Enterprise, July 2015.)

That study looked at transactional and member profile data from five anonymous North American credit unions, examining 500,000 members, and 250 million transactions using variables such as gender, credit scores, product balances, and income to search for correlations.

The researchers found patterns predicting member behaviour. For example, they were able to determine with 30 per cent accuracy the next best financial product for a cluster of members. The study also identified ways Big Data could be used to improve underwriting.

“You have to be open-minded about how you are going to use [data] and not be afraid of it” – Elizabeth Duff, NLCU

Young’s study is an extension of that work and is a deeper dive across the credit union system, looking at their readiness to use analytics and data to enhance their operations.

She identifies the weaknesses in the system but also spotlights the efforts of four organizations: NLCU, Edmonton-based Servus Credit Union (380,000 members, $14.2 billion in assets), BlueShore Financial (41,000 members, $3.5 billion in assets) in North Vancouver, and Denver, Colorado based Westerra Credit Union. All have successfully tackled projects involving everything from customer segmentation to credit card enhancement, and enterprise-wide deployment of business intelligence tools and strategies.

To better understand where credit unions are in their “data journey,” Young carried out in-depth interviews with 32 credit union leaders and analytic experts over a six-month period last year. The credit unions ranged in size from big to small.

The analytics maturity spectrum

In assessing where credit unions fall in terms of “analytical maturity,” Young says most are in the early stages of what will be a “long journey.”

She applied a matrix developed by academics and authors Thomas H. Davenport and Jeanne G. Harris in their groundbreaking 2007 book Competing on Analytics: The New Science of Winning. They examined companies looking at their readiness in three key areas: technology, people, and organizational framework. Based on those criteria they identified five stages of analytical maturity within organizations:

1. Analytically impaired: Such an organization has limited insight into customers, markets, and competitors. Its analytical process doesn’t exist and it lacks the skills and sponsorship to move forward. Its decisions are based on “gut instinct,” and it has missing or poor-quality data.

2. Localized analytics: The organization is building experience and insight using analytics; however, the process is disconnected and narrow. It has pockets of analytics in isolated areas, with limited management sponsorship. There is a desire for more objective data, but its transaction data is not integrated and it’s missing some important information.

3. Analytical aspirations: The organization has established enterprise-wide performance metrics and is building analytically based insights. However, its analytical process operates in silos, and while it has analysts in multiple areas of the business, there is limited interaction among them. They are still in the early stages of executive buy-in, but there is support, though resistance could be considerable. Its use of business intelligence tools is expanding. NLCU, for example, is in this range.

4. Analytical companies: They are developing integrated analytical processes and have strong skill sets among employees, but these still might not be properly aligned to the right level or role. There is broad support for initiatives at the senior management level and the company is adopting a fact-based culture. It has high-quality data with a business intelligence strategy, IT processes, and governance principles in place. Servus Credit Union is identified as one organization approaching this stage of analytical maturity.

5. Analytical competitors: At these organizations, data helps generate deep strategic insights. There is an embedded analytical process, a skilled analytics workforce exists that is supported at the highest levels of management, and the effort is broadly supported and integrated across the organization.

One such example Young identified is BlueShore Financial, which has grown assets to $3.5 billion from $800 million in 2000, with roughly the same number of members: 41,000.

BlueShore’s data analytics journey started 15 years ago, when it became one of the first Canadian credit unions to adopt a customer relationship management system to better understand its members, and then built out its investment and data strategy from there. The company has perfected wallet share expansion, tapping into Young’s observation that between 10 and 20 per cent of members create all the profit. Not surprisingly, though, most of the credit unions Young interviewed fell into the analytically impaired or localized analytics stages. The report notes “very few were firmly situated as analytical companies or analytical competitors.”

“Credit unions need to step back and ascertain what is going to be [their] game plan for long-term survival. They will lose ground if they don’t do something.” – Elizabeth Duff, NLCU

“It’s not a lack of desire to leverage the data that plagues these credit unions, but rather the lack of internal skill sets, namely, a shortage of people who understand how data is structured and organized, as well as people who are well versed in analyzing and applying business needs to that data to draw insight and outcomes,” Young writes. Other problems included poor quality-data, and fragmented technologies.

Stephen Kaiser, director of member and market insights at Servus, stresses the need to develop data analytic skills within the credit union space. “You have to have the proper people in place in order to manage relationships,” says Kaiser (an economist with an MBA), who works in business intelligence.

The report lauds Servus for growing its membership after investing in “technology, people, and a knowledge-based culture.” After expanding through a series of mergers, Servus undertook a major integration project to build a core banking system — which includes an enterprise-wide data warehouse — and develop a business intelligence strategy, spearheaded by Kaiser, who joined the credit union in 2012.

By connecting and consulting with various internal stakeholders, Kaiser has fostered a cultural shift towards fact- and knowledge-based decision-making. He says technical people and those mining data can only do so much. It’s the business lines — whether marketing, finance, risk, or operations — that need to own the data and drive efforts forward. “They are the ones who know what the data needs to be in order to be successful.” He adds that buy-in at senior management levels is critical. “Without that, you are not going to be successful,” says Kaiser.

In some cases, it’s not a question of building something new, observes Young, it’s a matter of leveraging what’s already there. “I found a lot of times that credit unions were not taking advantage of the data readily available to them within their own credit union.”

What concerns her most about her study is the sense of inertia she experienced when speaking to some credit unions. “I think there is some stagnant thinking in terms of seeing their members in a certain way.” It’s a stumbling block to moving forward, says Young, noting that “credit unions need to step back and ascertain what is going to be [their] game plan for long-term survival. They will lose ground if they don’t do something.”

She refers to NLCU’s efforts. By developing a more sophisticated customer segmentation that focuses on members and not products, NLCU has been better able to target and identify opportunities. For example, it launched a playful RRSP campaign targeted at Generation Y, which yielded 11 per cent more in deposits. It also created a successful new cash-back mortgage to respond to concerns raised by members about saving for a down payment. For those who don’t qualify, rather than ending the experience with a declined loan, it coaches them on what they can do to improve their financial situation and qualify.

NLCU’s Duff says that turning to data analytics and outsiders to help better understand your credit union memberships needs is not easy to do, and it can be expensive. “Bring your cheque book,” she jokes. It can also be a humbling and soul-searching experience, and some executives might not like what they hear. “If you don’t like what you find out, that’s too bad. You got to find out a way to make it work for your credit union. It has taken us in a new direction. We found out a lot about our membership and what is important for them.”


Don’t fight the data

Linda Young says studying data sets, big or small, can “shed light on essential topics,” everything from managing credit and operational risk to customer segmentation, predicting member behaviour, and increasing wallet share.

She notes that the financial services industry invested $6.4 billion in big data projects in 2015. That is expected to grow by 22 per cent annually through to 2020. Companies that use data to drive decisions are five per cent more productive and six per cent more profitable than those that don’t leverage data analytics, according to Harvard Business Review.

It’s the analysis of data sets and how companies use information that is expected to transform industries, improving efficiencies and enhancing operations. Take health care, for example. Young cites a McKinsey study that predicts big data will slash U.S. health care costs by as much as $300 billion annually by eliminating wasteful spending, duplicate records, and identifying patients who are most at risk of needing further procedures.

“Credit unions may not be savings lives, but they do play an active role in financial wellness,” Young writes in her report. “Big data (or small data) done right can help deliver personalized, tangible help.” ◊