How To Use Analytics To Identify At-Risk Members

Three credit unions share what they’re learning at the leading edge of data diving to spot troubling trends and troubled members as the movement deals with the COVID-19 pandemic.

The COVID-19 pandemic is roiling credit union members’ finances in unprecedented fashion, and forward-leaning cooperatives are slicing and dicing data as quickly and deeply as they ever have to find out who’s at financial risk and how much risk is involved, including for the credit unions themselves.

Analytics directors from across the United States regularly talk with about the tools they’re using, how often they’re using them, and what they’re learning about how to use analytics more effectively.

Sharing their insight and best practices below are Daniel Hirschlein of Grow Financial Federal Credit Union($2.8B, Tampa FL), Mike Wiseman of CAP COM Federal Credit Union ($2.0B, Albany, NY), and John Sahagian of BCU($4.0B, Vernon Hills, IL).

Different (Key)Strokes For Different Folks

A variety of tools and teams analyze member data and present actionable insight to decision-makers.

Mike Wiseman, Manager of Business Intelligence, CAP COM FCU

We have an enterprise data warehouse that is accessible via Tableau. Throughout the different operational areas of the credit union, we have established analysts or power users’ who use Tableau to manage tactical data needs. The analytics team is set up as a Center of Excellence to support the analysts in the organization and is responsible for advanced analytics that support more strategic needs. The analytics team uses tools such as SQL, SQL Server Integration Services, and R. Daniel Hirschlein, Grow Financial

We use Trellance M360 data warehouse as our primary repository for member activity and account status. My team and I retrieve data within M360 and present the information via Power BI to our business units. This data visualization enables those teams to track, trend, and understand member activity. We have some other data systems we have not integrated into M360 yet, so we pass along static data snapshots to support member activity analysis. Mike Wiseman, CAP COM

We have a hybrid analytics structure, with both a data office and people with analytic skills embedded in functional business units. The analytic tools we use vary by function and use case, such as R Studio for model building, Axiom for profitability, Tethr for voice analytics, etc. SQL Studio and Excel are still workhorses for basic data access and manipulation, but Microsoft Power BI is quickly growing at BCU as a common tool for information self-service and insight sharing. John Sahagian, BCU

Frequent Data Dives To Find Members At Risk

CAP COM looks at weekly trends, Grow Financial imports external data, and BCU updates a daily dashboard.

Early in the pandemic, we were examining key data, such as emergency relief loan volumes, advances of unused lines of credit, and cash/liquidity needs of our members on a daily basis. As we have seen a stabilization of the trends, we now view most information on a weekly basis. Mike Wiseman, CAP COM

We track a lot of data daily and understand that changes in member transactional behavior, new loans, deposits, and assistance programs can potentially impact liquidity. We’re also incorporating external data such as how unemployment and claim numbers are changing in the markets where we operate into our analysis. Daniel Hirschlein, Grow Financial

Starting in March, as the speed and magnitude of COVID’s impact started to take shape, we knew it would be critical to always use the most up-to-date data. The situation was evolving quickly. So, we developed an interactive rapid response dashboard in Power BI that our leadership team has used daily to stay close to trends across BCU and the country. John Sahagian, BCU

BCU keeps skips, extensions, and the mortgage pipeline up to date on its Rapid Response Overview dashboard.

Seeing Financial Distress In The Data And Responding With Member Service

Parsing the data in new ways helps to uncover potential risk to the loan portfolio as well as predict changes in coming deposit activity.

Our initial focus was on members seeking payment or fee relief as well as members applying for emergency relief loan products. After a few weeks, we added data regarding members receiving unemployment and stimulus deposits. A key focus is understanding the relationship between unemployed members and their ability to meet their lending obligations as well as their use or saving of stimulus funds. We’re also starting to explore member behavior with regard to branch interactions now that our branch network is open to walk-in traffic. Mike Wiseman, CAP COM

John Sahagian, Chief Data Officer, BCU

For members who have requested payment deferments, we’re trying to understand if there are common attributes at the loan level [e.g., FICO, payment amount, LTV, employment industry, geography]. By doing so, we can gain insight into the potential risk of our loan portfolio that has not been modified. Also, based on the characteristics of members who initially received a government stimulus payment, we were able to build a profile of a member whom we expected to receive a subsequent stimulus payment. This enables us to understand changes in future deposit activity. Daniel Hirschlein, Grow Financial

We’re focused on areas such as credit and debit card activity by merchant categories, requests for loan payment extensions, origination pipelines, and unemployment claims deposits, among other things. We are also following voice-of-the-member metrics daily, as we know our members will let us know how they feel we’re handling things. And as branches started closing, we had to closely monitor deflection of traffic into other channels such as the call center, chat, and secure messaging within online banking. Voice analytics has also helped us quickly understand and adjust to the key themes and concerns embedded in service conversations. John Sahagian, BCU

Best Practices And Lessons Learned

Good data is essential for good implementation, advises Daniel Hirschlein at Grow Financial.

Daniel Hirschlein, AVP of Analytic Services, Grow Financial FCU

That means:

  • Do not inter-mingle products that are fundamentally different, such as emergency loans versus normal unsecured loans.
  • Incorporate tracking mechanisms to understand the state of the loan at the time of deferment.
  • Centralize pandemic reporting to come directly from the analytics team.
  • Work cross-functionally to understand the inner workings of new products and assistance programs.
  • Identify what other areas could benefit from an analytical approach (e.g., provision expense).

Keep calm, and let your data carry you on, says Mike Wiseman, CAP COM.

He also advises credit unions to:

  • Adapt. Chances are, you’re seeing patterns in your data you’ve never seen before. Don’t overreact but understand what your data is telling you and adjust your processes, products, and plans accordingly.
  • Focus. It can be tempting to try to report on everything, but identify a few important metrics and targeted KPIs that will make the biggest impact.
  • Perspective. Your member behavior will most likely change and has changed already. If you spotted trends early on, you might need to return to your assumptions surrounding activity. Determine if those trends are still valid, and do your best to maintain perspective.

Join The Credit Union Analytics Conversation

With monthly meetings discussing topics ranging from technology to how to build your team, there’s no limit to what you could learn when you connect with others who want to improve how they use analytics to better serve credit union members.

Join The Analytics Community


July 6, 2020

Keep Reading

View all posts in:
More on:
Scroll to Top
Verified by MonsterInsights