Lagging Contenders: How Credit Unions Can Catch Up in Data and Analytics

By Peter Keers
July 26, 2017

The message has been ringing loud and clear throughout the credit union industry for years: make better use of data and analytics or lose “member share” to more progressive CU peers or (horrors!) banks.

Despite the warning cries, the proportion of credit unions embracing this trend is (horrifyingly!) low. A recent McKinsey & Company report emphasizes the fact that in many industries are only achieving only a fraction of their “digital potential.” However, the report observes, “In the United States, the information and communications technology sector, media, financial services, and professional services are surging ahead…”. Other players in the marketplace served by credit union have a big head start.

Credit unions who have been sitting on the sidelines waiting can wait no longer. To get off the bench, these organizations need to ask:
• What are the basic questions about the organization’s strategic direction that cannot be answered today?
• How can existing data be better “generated, collected, and organized”?
• What would data outside the organization be useful?
• What skillsets are missing internally, and to what degree can they (or should they) be outsourced?
• Once “insights” are uncovered from analytics, what are the practical steps to leveraging them to create value?

Maddening Silos

Most credit unions know what they need to do: grow members, retain members, grow products per member, etc. They also know the data needed to support strategies and tactics in achieving their goals is locked up in their operational systems. These “silos” trap data so that even simple questions about a subject area cannot be asked. Thus trapped, the data cannot be linked to other valuable data so that questions bridging subject areas are impossible.

Nevertheless, a necessary exercise must be done to inventory the trapped data by silo and make a least a rough blueprint of how the data would be used. Each silo should be analyzed for what questions could be answered simply from that source. Once all relevant silos have been analyzed, then the connections between the silos need to be identified. The resulting blueprint will serve as the starting point for how the credit union could use the data it already possesses.

Break It On Down

The McKinsey report suggests exploring improvements in generating, collecting, and organizing existing data. In terms of generating data, thought needs to be given to the source data itself. What existing systems are simply past their prime? For example, underlying architecture of some legacy core systems on the market for 20+ years has never changed. Extracting data for analysis from their underlying databases is often a more complex (and expensive) effort than up-to-date systems.

Improvements in collecting data can be summed up in the phrase, “data quality.” Too often, poor data quality complicates an otherwise straightforward analytics task. The frequent culprits are the antiquated source systems as noted above. Their input edit logic is often inadequate, so Member Service Reps and other employees can all too easily introduce errors into the data. Even before replacing old systems, credit unions can:
• Tighten up input editing logic in the current software where possible
• Train employees on how to avoid input errors
• Undertake other data cleansing steps on existing data like checking for incomplete SSNs, etc.

Organizing data in a better way is achieved primarily via a data warehouse. Data from multiple source systems within the credit union is extracted, transformed if necessary, and loaded into a separate database specifically built to support analytics. The new data organization also allows additional data sources to be integrated over time. The data warehouse is the ultimate data silo breaker.

Outside Chance

Pioneering credit unions data and analytics that adopted data and analytics innovation years ago had a big advantage over credit unions that are just starting now. The early adopters only had to focus on their internal data. Over the years, however, large volumes of valuable data outside the credit union have become more accessible. “Big Data” has exploded onto the scene.

For Credit Unions, Big Data Takes Two Forms.

• Structured - External data that is neatly arrayed in rows and columns but comes in very large volumes. Examples are census data and real estate price databases like Zillow. Associating this data with individual members sometimes simple but in other cases more demanding. After linking the data, however, members can be viewed in an increasingly broad perspective that could highlight otherwise unknown opportunities. For example, Zillow real estate values can uncover mortgage refinancing or HELOC opportunities when linked to internal member loan data.

Also, in the structured category are “pools” of data to which credit unions voluntarily contribute. For example, loan data, scrubbed of personally identifying information, is added to the pool by many credit unions and can be used for such purposes as meeting the forthcoming CECL regulations.

• Unstructured – Social media sites like Facebook, Instagram, and Reddit generate an immense volume of data that, when properly tapped, can reveal amazing insights very quickly. While census data has a lag time from collection to publication, social media data is immediate. Credit unions successfully using such data could quickly spot attitudinal or behavior movements among members and respond before positive opportunity fades away or negative trends get out of hand.

“Beginner” credit union need not jump into Big Data scene right away. However, they should prepare to be “fast followers” because there will be competitive pressure to do so.

Power from the People

Credit unions about to embark on the data and analytics journey are aware of a tough reality: there is very little if any expertise internally in this subject area. Even an otherwise top-notch credit union IT staff rarely has data and analytics specialists. These skillsets must be acquired by existing personnel (a long process), hiring these skillsets (high demand, short supply, expensive), or outsourcing.

The outsourcing option has a lot of advantages for credit unions. There are two major ones. First, the number of firms offering data and analytics services has blossomed over the past few years. There is a wide array of products and services from which a credit union can implement its data and analytics strategy.

Second, many of these vendors have gained valuable experience in the credit union space over the past few years. Their ability to deliver valuable services at a reasonable price point improves every day driven by the forces of increasing efficiency and competitive pressure.

Choosing a data and analytics vendor will be greatly facilitated by the data inventory done in the previously mentioned silo analysis. Credit unions with a clear idea of what they are trying to accomplish will be much better prepared to assess vendor delivery capabilities.

A final note. Getting involved with a vendor is that it does need to be forever. A long-range plan to develop internal capabilities with teaching and mentoring provided by a vendor is a great way to gain greater control over the credit union’s data and analytics program.

Getting to the Value

After all this work, there better be a payoff. Unfortunately, some pioneering credit unions found out that an attitude of “build it, and they will come” does not work. Just as there needs to be an inventory of internal data, there also needs to be a people inventory. This second inventory extends not just to IT but to the entire organization. A data and analytics culture should be developed at all levels of the organization. In many ways, this is the most difficult task. Becoming a data-driven organization is a longterm proposition. However, there are practical steps to make this a reality.
• Work with a Data and Analytics Strategy Consultant – A great way to get a solid start is to engage an experienced consultant to help craft the comprehensive plan. This ensures no major area is missed and important details are addressed.
• Look for Small but Valuable Wins – Cherrypick some easy wins for the first projects. Although these may be modest in terms of overall impact, results come in faster. Hearts and minds will be won over by “wins,” and any hiccups with the program can be quickly uncovered and remediated.
• Enlist Tech-Savvy Workers as Evangelists - Look for employees who “get it” and are more enthusiastic than the average person about the potential for technology to accomplish great things. Younger workers aren’t the only ones who fit this description. Look also for more experienced workers who can be big influencers among their peers.

…

The McKinsey & Company report also includes a lot of very “bleeding edge” ideas about artificial intelligence and robotics impacting the digital potential of organizations. However, while credit unions starting their data and analytics efforts must be aware of these important trends, sticking to the basics and getting started now is critical for long-term viability.