4 Data Science Strategies for the Biden Era
Data can be a source of comfort and confidence as the pandemic rages on and the Trump Administration comes to a close.
Now that the dust is beginning to settle on the 2020 election, the path forward for our nation is becoming a little clearer. Although the election results fell short of the surging blue tide promised by Democrats, the culture in Washington will nonetheless shift back to a progressive-leaning agenda, led by President-elect Biden and the House of Representatives. Regardless of what happens in Georgia, the Senate tradition is to take a conservative position when it comes to the decision to vote at all. The upper chamber takes its role as the ultimate step in the legislative process seriously and approaches legislation thoughtfully, not vigorously.
That doesn’t mean change won’t happen. In fact, change could be tumultuous in the short term as a bitter lame duck Congress is forced to negotiate coronavirus relief, economic stimulus packages and the federal budget. Meanwhile, businesses and families continue to suffer from the coronavirus as the infection rate remains stubbornly high.
But here’s the good news: The long-term outlook is generally positive, and that’s not just good ol’ fashioned American optimism talking. During the Great Recession, our economy imploded from within due to systemic failure; this year, it was sideswiped by an outside enemy that has a limited and somewhat predictable lifespan. While we must build a better foundation to withstand future pandemics, our economic and financial foundations are in relatively good shape.
So how do credit unions move forward now – right now – knowing things will get better but not knowing exactly how we will get there?
Let’s step back from examining Election Day weeds and take a look at the big picture: The mood and the philosophy of the Biden Administration. President Trump’s approach to political negotiation was so vastly different from his predecessors that it changed the tone of America. The Biden campaign promised a return to comparatively congenial times, and the voters spoke.
For credit unions, it means a call for evolved member service. That’s no short order, because it comes at a time when Americans are still digging out from beneath the COVID-19 economic rubble and require modifications, second chances and extra assistance. And, new leadership at the CFPB that is emboldened by a supportive president could increase regulatory scrutiny on service, particularly as it relates to collections.
How can credit unions respond to meet these challenges? Employees are exhausted after nine months of rising to meet unprecedented challenges and have nothing more to give. Meanwhile, credit unions are bracing for losses and don’t have room in their budget to hire more people or make large capital investments in technology.
The answer is data science – a more affordable and better targeted application of technology. Despite being somewhat new technology, tools like machine learning, AI, advanced analytics, natural language processing, modeling and smart business logic are constantly improving, and financial institutions are finding new ways to apply them. The opportunities to use data science to reduce losses, improve efficiencies, grow members and take service to a new level are almost overwhelming.
If there’s one thing we don’t need more of right now, it’s being overwhelmed. So, here are four achievable and exciting ways your credit union can implement data strategies for life under the Biden Administration.
Ramp Up Collections With Advanced Analytics
We shared in a recent CU Times article that data science, and advanced analytics in particular, have drastically changed collections strategies. Here’s a new example. Most credit unions can easily identify members who are delinquent or consistently late with payments. Most are also proactive in their outreach efforts to offer solutions. But are those solutions likely to be successful? By combining what you know (late payments, transaction trends, unemployment, etc.) with advanced analytics like sentiment analysis, predelinquency modeling and recovery scoring, you can offer risk-based solutions that are more likely to perform.
For example, imagine a member who is not unemployed but has fallen behind on their last two payments. They have a source of income, but there are also signs of a diminishing financial status. Running this member through smart modeling may unearth that they are likely to withstand economic stressors and produce insights that explain why. Armed with this information, you can promote more effective treatments that reduce losses and build goodwill that will pay off when the economy recovers and members are ready to borrow again.
Diversify Attributes
But remember, modeling is only as good as the elements it considers. You must consider a number of macro and micro datasets. One unique approach to achieving this diversification is to use granular demographic attributes like a ZIP+4 sample, for example, which includes on average four to six households. This granular level incorporated into your big data science efforts will allow for trustworthy results and further improve your ability to offer a repayment path best suited for each member’s individual situation.
Update Occupational Data
This is an occupational-specific recession. The coronavirus has adversely affected industries dependent upon in-person attendance such as travel, tourism, entertainment, restaurants and bars, and retail shopping. Because President-elect Biden has pledged to take a more aggressive response to COVID-19, we can expect a slower reopening in affected regions.
For that reason, forward thinking credit unions are defining risk profiles by attributes that align with occupation industry, such as unemployment, income estimation, pre-pandemic payment trends, ACH patterns and anticipated return to work. Some credit union leaders say they are performing outreach efforts, including call campaigns and digital channels, to gather that updated employment data and sentiment on future employment. Internal or outsourced business development teams use this information to predict overall portfolio risk, as well as target which specific members are struggling and determine the best individual loan treatments.
Match Collectors to Members
Data insights are mute if you’re not infusing them into processes that optimize operations and improve performance. Here’s a fascinating application for business logic: Use fact-based tracking to monitor collectors in real time, assign a skill level and skill sets, and use that information to determine which members are assigned to each collector. The result produces favorable outcomes for repayment, the member experience and employee engagement and satisfaction. Tracking performance-based metrics like right-party contacts, payments and promises, instead of just attempted calls, can drive skill level assignments. A collector who is exceeding right-party contact goals deserves to be served high priority members, and objectively matching collectors to members will also improve recoveries.
Data can be a source of comfort and confidence in a time when the pandemic rages on and the Trump Administration comes to a close, but the path forward is still unclear. Simple assumptions about how the mood and agenda will change in Washington provide plenty of opportunities to build for the future. President-elect Biden will return to the same approach as his former boss, which means credit unions must ramp up their member service efforts, especially for those who are struggling to recover from coronavirus-related income losses. Because budgets are tight and time is valuable, a thoughtful, targeted data science strategy is a clear winner.