Digital Banking Gets Personal

In the 1990s, a friend of mine banked with a local bank. All of the tellers knew her by name, and when she went to get her first car loan, the loan officer patiently walked her through the entire process—in person, sitting in his office. 

Three decades later, one of the big three banks now owns that community bank. There isn't a branch in my friend's town, so she does her banking on their digital app, which makes her feel like she's doing business with a machine. They occasionally send her offers that she says demonstrates the bank knows absolutely nothing about her or her financial needs. 

My friend's experience helps to illustrate two major changes in banking: the deterioration of personal relationships and the move to digital, which COVID significantly accelerated. Before the pandemic, digital banking was considered a nice feature to have but not critical. The pandemic changed all that, of course. When people couldn't physically go to their bank or credit union, they needed to be able to conduct their banking online. 

The relationship between the customer and financial institution went from personal to transactional, which put local banks and credit unions at a disadvantage for two important reasons. First, those institutions have long differentiated themselves with personal relationships, which are more challenging to cultivate in the digital realm. Second, consumers have come to expect the same type of personalized digital experiences they get from companies like Amazon and Netflix from their financial institutions—and any company they do business with, for that matter.

In a recent Q2-sponsored survey conducted by The Financial Brand, roughly 7 out of 10 financial executives surveyed said providing personalized digital experiences to consumer and business customers is very important. Yet, only about 2 in 10 are very confident they're providing personalized digital experiences. If that 20% only knew about Q2 SMART.

For years, Q2 SMART has enabled banks and credit unions to have direct, targeted conversations with their end users from within the digital banking application. Using millions of data points from online banking transactions, Q2 SMART identifies behaviors that suggest interest in particular products. The suggestions can be used to build out and manage targeted marketing campaigns—all within a single, easy-to-use platform. 

But Q2 SMART is far more than a marketing tool. And now, with our new patented systems, Q2 SMART is even more capable of distilling vast quantities of user data down into valuable and useful insights.

Making sense of the data

Community banking is about relationships—relationships between the bank or credit union and the individuals or businesses who rely on them for all their financial needs. The financial institution wants to serve its account holders and members in as personalized a manner as possible. 

That's one half of the landscape for our customers: Their account holders and members. 

On the other hand is data. Financial institutions have a vast amount of data about their account holders. That data spans everything from how a person manages their financial life to how many savings and checking accounts they have to whether they have children and whether those children have savings accounts. Does the account holder or member have college funds for the children? What about credit cards, home loans and retirement accounts? The list goes on. 

So there's this tornado of data swirling around, and it keeps increasing. As your account holder base and the respective data grow, providing personalized financial services to your users becomes increasingly difficult. 

Q2 SMART specifically addresses this challenge. It takes that enormous mass of data—much of it unstructured and much of it coming from the far reaches of the financial services landscape—and condenses it down into clean, intuitive abstractions.  

We primarily use traits, which are simply aspects of behavior associable to account holders. Traits span a complexity spectrum ranging from very simple to very complex, involve a multitude of data and represent things like recommended financial services products.

For example, an account holder could have a trait that would indicate it's appropriate for the financial institution to recommend a 16-month CD to him or her. The set of data that would inform a trait on that end of the spectrum is very complex and not as simple as something like average account balance. 

It's important to note that a trait isn't only data on the end user. It's also an analysis of the data. Traits can represent vastly complex processes that involve more data than just what's associable directly to an end user, though the trait itself will be associable to the end user. We leverage the user's data plus all of that individual's nearest neighbors' data (more on this later).

And as noted earlier, traits are not solely about marketing financial products. For instance, a trait may capture the cumulative risk that's associable to an account holder when he or she is performing a certain kind of transaction, like an international wire transaction. That's the kind of trait that matters to an employee in fraud prevention, for example, but not so much to a member of the auto loan marketing team. 

New patented systems 

Our new patented systems give banks and credit unions even better personalization tools at their disposal through Q2 SMART. 

Let's start with the traits that recommend specific financial products tailored to meet the needs of an account holder. How does that happen with hundreds of thousands or millions of account holders and myriad products? We can look to Netflix, which has a similar challenge. 

Netflix has a vast catalog of content as well as a massive number of users, each unique and distinct in their own way. Netflix figured out a long time ago that it's overwhelming to users if it puts its gargantuan catalog of movies and television shows in front of them. Netflix needed a way to pair content precisely and show each person content he or she will find compelling.

But how do you do that when all the content is different and all the users are different? Netflix and other companies (such as Amazon and TikTok) ultimately began using a set of mathematical techniques called user- and item-based collaborative filtering. Using data, we can make guesses about how alike two people are, and then once we understand how alike they are, we can understand more effectively how to recommend products to them. Netflix figured out that if you can do this with two people, you can do it with 2 million people.

Now let's come back to banking. One of the patents we've taken out is for a new system that recommends targeted financial products to end users in much the same way Netflix and Amazon recommend novel products and content to their customers. 

Generally, two users are more likely to be "near" to each other if they engage with similar products in a similar manner. So, we couple our vast data assets with various machine learning techniques to produce a well-defined mathematical measure of similarity between different users. 

We take everything we know about an individual or business and then view that knowledge in the broader context of everything we know about everyone else.

Based upon our deep knowledge of the individual in question, along with similar historical knowledge of thousands of similar individuals in similar scenarios, we're able to do something very simple, yet powerful: Recommend the single most important financial product for their life or business at a particular point in time.

With Q2 SMART, consumers like my friend never need to receive product offers completely unrelated to their needs again.