Source: Shutterstock.
Understanding a member's needs and how they see their relationship with their credit union means recognizing and exploring how they interact with the outside world. Detailed information about a portion of these interactions can be captured whenever they use their credit union's products and services to complete an interaction. Debit card, credit card and ACH transactions are gold mines of information about how members are, and are not, engaging with their institution.
Beyond engagement, a credit union leader can learn about gaps in offerings, hidden preferences of their members, and even tactical insights that can help guide product development decisions. This can be done by identifying similar services being leveraged by other financial institutions through transaction data. In addition, transaction data can tell a credit union who their true competition is – which may come as a surprise. It provides a peek into the day-to-day lives of members and helps cultivate a deeper understanding of how credit unions can best serve their membership. Transaction data can also be used to increase marketing efficacy, develop strategies for boosting interchange income and even find new community partners.
The first step of using this transaction data to gain a deeper insight into members' needs, patterns and preferences is to identify the sources of this information. Can you access this data now? How much of it can you access? Who can access it? How much detail can you get? How accurate is the data? Credit unions can develop reports and generate insights more efficiently by first gaining a clear understanding the abilities and limitations of their data environment. This understanding will also help a credit union create more precise use cases and set more meaningful priorities when deciding what data to mine when. The next step, and the most important one once the data is in hand, is to clean the data.
Data coming from ACH files, card networks or any transaction source is notoriously dirty and frequently changes. Maintaining the quality, integrity and consequently the overall usability of that data is very labor-intensive. Credit unions that are committed to using transaction data to better understand their members must positively identify and allocate resources to this data beyond the initial setup and data governance deployment effort. The only thing worse than no data is bad data – not making stringent and frequent quality control efforts over transaction data can quickly render this valuable information worthless. Fortunately, advances in artificial intelligence and machine learning (AI/ML) technology can aid credit unions in the endeavor to mine and refine their transaction data.
AI/ML offers a different approach to cleaning data. Historically, credit union IT professionals would use a signature-based approach. In this approach, commonly misspelled words or known substitutes like "WM-Super" instead of "Walmart" are programmed exactly. The problem with this approach is that it is very brittle. As card transaction data changes over time, this approach has a hard time keeping up. The use AI/ML allows a data scientist to write a model once that can automatically clean the data. This prevents the problem all credit unions leaders see routinely – broken reports with missing or bad data. With AI/ML, the data clean-up process is much more resilient and requires less human intervention, which is great for operational efficiency.
Not all data points contained within the transaction datasets are created equal. There are many redundant, superfluous and non-relevant fields. Card transaction streams can contain hundreds of records, and ACH files follow a hierarchy that contains various rabbit holes of useful and useless information. Getting support from operational experts, often found in-house, is a great way to incorporate many cross-functional teams into data transformation efforts. Transaction data is also inherently unstable, requires context to be useful and may contain errors. Card transactions, for example, contain terminal information inputted by a human, making details such as name and location either erratic or incorrect at times. ACH descriptions change purposefully to thwart stop payments. But in these volatile and messy streams of data lie valuable information that cannot be found anywhere else.
Mining transaction data goes beyond simply knowing what a member's favorite gas station is or if they are more likely to have a mortgage with a community credit union or a national bank. Transaction data can press play on a discovery process that will show what is important to a member in three key ways: Physically, financially and habitually.
- Physically: You know where members live and you may know where they work, but knowing where they spend their money is arguably more important. Do your members prefer ATMs near their home or on their regular errand-running route? Is there a time of day or day of the week or holiday when your members may make more transactions, and can this impact staffing, network traffic or ATM cash loads? Knowing the physical patterns – where and when – of your members' demands can ensure the credit union is consistently in the right place at the right time.
- Financially: Transaction data is one of if not the most telling sources of data for how a member is performing financially. It can indicate major lifestyle changes or events that may impact opportunity for a credit union to grow their relationship, or signs of deterioration or abandonment of their relationship with the institution. Transaction data, and shopping patterns in particular, are so valuable that major lending fintech disrupters are successfully basing underwriting decisions on an applicant's online purchase patterns rather than traditional debit-to-income calculations.
- Habitually: People are creatures of habit and typically dislike change. This is especially true when it comes to which card they pull out of their wallet when they make a purchase. Are you top of wallet for major online retailers? How about for big box brick and mortar? Mom-and-pop shops? What about when a member is traveling or makes a first-time purchase at a new restaurant? What about when they Zelle a friend some money or use Apple Pay? Each of these habits can tell a credit union how it is perceived and where they fit into their members' financial lives.
Those who put forth the effort to extract, clean and make usable the data from card and ACH data sources will be rewarded with fresh insights into their members' needs. As credit unions look to compete in the financial services landscape, this level of insight quickly transforms from luxury to necessity. Transaction data will also serve as a valuable tool as credit unions look to support those rebounding from the COVID recession and look to rebound themselves. Transaction data can give a credit union the knowledge and opportunity to become more than a medium for moving money – it can support trust and foster a more engaged and sincere relationship that is centered on a member's and community's financial well-being.
Ray Ragan Ray K. Ragan is the Co-Founder of Clear Core in Tucson, Ariz.
Timothy Strasser Timothy "Buck" Strasser is the Founder of Clear Core.
© Arc, All Rights Reserved. Request academic re-use from www.copyright.com. All other uses, submit a request to TMSalesOperations@arc-network.com. For more information visit Asset & Logo Licensing.