Achieving the new analytical tools’ full potential

Do banks and credit unions need to do a better job of training employees to realize the full potential of new data analysis tools?

Can they even compete with other industries for data scientists and all the other professions emerging within this field? And how do these specialists best communicate with nontechnical leaders at institutions to make these tools most effective?

We spoke with several industry consultants—as well as an expert with decades of data analytics experience—to get their takes on how banks and credit unions can get the most out of data analytics.

Alex Kwiatkowski, SAS’s Director of Global Financial Services, based in London.

This sort of technological capability was once reserved for the deeper pockets of bigger banks because it wasn’t affordable. But the size barrier is now a thing of the past.

These days, financial institutions (FIs) don’t have to spend millions of dollars installing data analytics tools within their own data center—they can access more powerful advanced analytics tools through the cloud, with pay-as-you-go options.

Digitization has been at the heart of this progress, and almost every organization has embraced digital because customers expect it. Whether an institution is large or small, all customers want to know—what can you do for me? Answering this all-important question requires insights gleaned from analytics.

The training of in-house data scientists should be continuous. They need to keep finding ways to do things differently to get better results, and then they need to refresh their analytics models. When models aren’t continuously reinvigorated, their effectiveness degrades over time.

Do all FIs need to do better? Yes, but not because they’re currently doing a terrible job. The more the digital world progresses, the more data we create, and we all need to do a better job of using existing data to make better decisions. Regardless of size, banks could always be improving this.

Do they need to hire an entirely new or different set of employees to master these technologies, such as data analysts, data visualization specialists or data engineers? Institutions don’t need an entirely new type of data science professional— they just need someone with a broad skill set and familiarity working with data science tools; someone who not only understands data tables, but also has the ability to visualize data so they can develop algorithms to determine correct answers.

It’s a lifelong learning journey, adding new skills as new technologies evolve. For those who have stuck around since mainframe computers and punch cards, and decided they didn’t need to change much through each generational evolution—that’s where institutions will have problems.

Joseph Cady, Managing Partner at CS Consulting Group in Lake Arrowhead, California.

Do banks and credit unions need to do a better job of training employees to realize the full potential of these new data analysis tools? The short answer is yes. Whether it is data analytics tools or leveraging AI with ChatGPT or AutoGPT, all will require a new skill set.

Using ChatGPT as an example, it is important to learn how to ask the right questions or phrase the task appropriately to improve the output quality. This becomes even more important as the output learns from itself and the user adds new tasks.

Do institutions need to hire an entirely new or different set of employees to master these technologies, such as data analysts, data visualization specialists or data engineers? Not necessarily. But skills will need to be developed to sort out the useful data and results from the misses.

Banking executives and users also need to be mindful of the risks these tools create, both for biases and compliance issues. They should not be mindlessly applied. There is much trial and error when using these tools. Moreover, because of their broad application, the tools will be used throughout the institution, requiring these skills to be developed within all business units or departments.

Do FIs need someone in IT to translate the potential of the new technologies to the business units and vice versa? McKinsey has observed that organizations need to shift from digital users to “becoming digital.” This includes creating a bridge between IT and the rest of the organization.

The beauty of a tool such as ChatGPT is that it doesn’t require an IT background. Tasks can be written in plain English. Thus, IT has a role in ensuring that digital skills are developed throughout the institution, at a level that everyone understands. IT, in this case, shifts from being a department to becoming an organization-wide process.

Steve Pierce, Vice President of Innovation for the $2.1 billion-asset Black Hills Federal Credit Union based in Rapid City South Dakota.

Do banks and credit unions need to do a better job of training employees to realize the full potential of these new data analysis tools?

For nontechnical bank leaders, there is often an initial misunderstanding when purchasing commercial software for data analytics and business intelligence reporting that it “contains all the magic,” including a preload of existing finished product templates. And that after IT connects it to the core banking database, it will just start spitting out charts of branch performance metrics and the next best product for the customer.

The reality is that the magic comes not from in the box but from the minds of skilled employees. That’s why companies that offer these solutions also provide the training to get the most out of them. The more mature programs range from foundational data literacy to advanced certifications.

Contemporary data analysis environments are surely more intuitive, efficient and powerful than legacy offerings. But that doesn’t necessarily mean these advancements replace the business acumen, wisdom and curiosity of a savvy data analyst. These new systems are better connected to massive data sets—though they lack the statistical skills of a data scientist who can predict what’s to come for the financial institution.

As such, these institutions must begin if they haven’t already, and perpetually invest in staff development in the disciplines of each relevant specialty to coincide with the software stack and the goals for each.

Starting is easy. All key roles would benefit from taking a well-designed data literacy course. This course sets the operational expectations, common language and foundational understanding of data as an organizational asset. Then, data analysts and power business users can begin with simple reporting as they take the coursework offered by providers.

Do institutions need to hire an entirely new or different set of employees to master these technologies, such as data analysts, data visualization specialists or data engineers?

This mostly depends on what their goals are and the structure and complexity of the data sources. Minimum viable delivery of an operational reporting environment, ad-hoc lists and the like—an FI could begin with a couple of data analysts. They don’t necessarily need to hire them.

Often, I’ve seen the best data analysts cross-train from a business analyst or other role with a business and analytical mindset. Organizations with mature goals and data sources, which include sophisticated data models, AI and predictive analytics that drive hyper-personalization and business growth—those FIs typically have a team of specialists, as I’ve described.

The rule of data analysis outflow is that you get out what you put in. And it’s almost always multi[1]plied. There’s always more to go after. As cited in a 2020 Accenture report, a Forrester study found that between 60% and 73% of an organization’s data is never analyzed!

Katie Kuehner-Hebert  is a BAI contributing writer.  

We offer actionable insights on marketing strategies in the BAI Executive Report, “Marketing strategies for the digital age.”