The Voices of Money20/20: Finalytics.AI’s Craig McLaughlin Talks Personalization in the Digital Age
By John San Filippo
Finopotamus was onsite at Money20/20 USA in Las Vegas Oct. 24-27. Co-founder John San Filippo interviewed more than two dozen technology experts on a wide range of topics. The results of these interviews are presented in this series called “The Voices of Money20/20.”
“Every single credit union website today tells the same story to every visitor, regardless of who they are,” Craig McLaughlin, CEO of Finalytics.AI, told Finopotamus. “It's the exact opposite of what would happen with your best member service representative in the branch. The MSR can quickly judge based on the questions the member asks, what's the right thing to present to this member?”
In the COVID-19, all-digital era, the challenge according to McLaughlin is replicating that dynamic experience in a digital setting. That’s where artificial intelligence (AI) comes in.
“We use data and artificial intelligence to dynamically change the digital experience based on behavior of that particular user,” said McLaughlin. “For the member, they just see that the credit union cares about them and is listening to them. In other words, it seems to be an experience tailored to their needs.”
McLaughlin offered the example of someone looking for a mortgage loan. He said that on average, applying for a mortgage loan online requires 2.7 visits comprised of 13 different pages on the credit union’s website. By dynamically adjusting the experience according to each action made by the member, the Finalytics platform can reduce both the time and complexity of this process.
How It Works
“What we’re doing is tracking the behavior of a visitor and we're picking up on attributes like their geographic location as they're coming into the site,” said McLaughlin. “Based on those attributes, we're dynamically re-rendering the experience for that individual visitor.” He added that Finalytics looks only at behaviors, not actual account information. That means the system works equally well on both sides of a member’s login credentials.
“We're moving away from cookies,” said McLaughlin when asked how the system keeps track of a member’s behavior. “We're using the local storage of the browser. And we're also piggybacking on the web analytics that the credit union is using. A lot of credit unions use Google analytics, some are using Adobe analytics. We're piggybacking on that data set so that we can see who is a repeat visitor. When somebody comes to the site, we're looking at their attributes. And based on that, we're making predictions about what product they might be interested in, what service or support they might need.”
AI Versus Old-School Predictive Analytics
Finopotamus then asked McLaughlin the difference between an AI-driven system like Finalytics and the predictive analytics that credit unions have been using for 10 years or longer.
“If you go by the bare bone definitions of it, predictive analytics are most commonly doing an analysis and looking at historical data, then making a forecast about the future. That’s it, you’re done. You've made your forecast,” responded McLaughlin. “I think the key difference with the machine learning aspect of it is, as it's seeing more people convert or drop out of the funnel, it's changing those signals that were used for the prediction.” In other words, the system keeps getting better and better at making those predictions.
“As our platform starts to gather more data, it learns from both the funnel abandonment, but also the funnel completions,” he continued. “Also, those signals change over time. Buying behaviors for things like mortgages change month to month and year to year. So, it's keeping track of people being more competitive with their shopping, less competitive. Are they consuming more content, less content?”
Timing Is Everything
“Buying behaviors for credit union products actually change a bit depending on the time of the day and it can vary between different credit unions,” noted McLaughlin. “We did some analysis for a credit union in Southern California and we found that people on the Chrome browser after 2:30 p.m. on Mondays and Tuesdays are much more likely to be interested in checking and savings products compared to other products.” The Finalytics system can adapt to that knowledge, he added.
“At a typical credit union website, the overall conversion rate of visitors to completed applications hovers around 1%. By making the experience more relevant for the visitor and predicting, not just their products, but what type of messaging is appropriate, you can start to increase that to around 3%. Going from 1% to 3%, may not sound like much, but that’s a 200% increase in conversions,” McLaughlin said, adding that for a typical credit union, this metric can add up to millions of dollars in additional assets each year.
How It Integrates
“We don't need any new data pipes,” said McLaughlin. “It's a tag-based system, similar to an analytics platform. The credit union will deploy our tags on their public website, the funnel, and their online banking areas. And that's essentially all we need in terms of system integration.”
He added that configuration is managed through an administrative console. “You can go in and configure which banners you want to use, which photo assets and imagery. And you can either use our library which we design around the credit union’s brand guidelines, or if they want to use their own, that's fine, too.”
He continued, “So there's really no integration necessary. However, many credit unions have told us they want to manage some of these banners inside of their content management system, and that's fine with us. In that case, we're fine integrating with their CMS where our platform is essentially telling the browser, this is the product that they'll be interested in this is how to message it. Then the CMS picks that up when it's rendering the page.”
Finopotamus asked how a credit union knows when its Finalytics deployment has been a success.
“We do a fitness assessment when we're in the sales process,” said McLaughlin. “Our platform is built on top of Google analytics, grant our application access to your Google analytics. We will go through and look at their data. We'll come back and say, here's where your current benchmark conversion rates are. Here's where we think we can be if we apply AI machine learning to it. It will improve by this percent. We lay out a forecast and we say, this is what we're going to do.”
McLaughlin noted that these forecasts cover growth in both membership and assets.
“Credit unions don't have an AI problem. They have a product growth problem,” concluded McLaughlin. “They need a digital growth solution. They need a new plan. So it's not about developing an AI strategy. It's about developing a ‘how do I grow my credit union’ strategy.”