When it comes to data, two things are certain. One is the old adage of “garbage in, garbage out” — if the data quality isn’t there, it’s useless. And the second is that you can have all the data in the world, but all of your efforts are wasted if you don’t do anything with the information.
Many products currently available to community banks now come with their own built-in analytics. Alternately, analytics platforms can exist independently and combine data from multiple sources.
Analytics are powerful, but how can you draw meaning from the data? We talked to four different experts in the areas of marketing, product, security and AI to get their takes on why analytics matter and what the future holds for financial institutions looking to rely more heavily on analytics.
A bank’s digital products can encompass everything from digital banking, to online account opening and loan origination, to mobile apps. Whether your bank develops these products internally or partners with a fintech, improving the overall customer experience has become a top priority.
“Product analytics helps companies answer the question, ‘How do our digital products drive our business?’” says Shadi Rostami, senior vice president at Amplitude, a product analytics platform. “With behavioral data down to the individual user level, financial services companies can better understand how their customers are engaging — or not engaging — with their products.”
Rostami said that as a result of the pandemic, more companies are understanding the value of behavioral data. In the past, analytics may have been as simple as “How many customers opened accounts online?” But to truly understand the customer journey, an event-based model provides more granular detail. Product analytics instead would identify the exact steps and clicks during the account opening process. If the account opening process is started, but not completed, it gives the bank insights into where customers get stuck.
But product analytics do not exist in a vacuum. “In order to create holistic change across the customer journey, all teams — from product to marketing to customer success — need to be able to leverage these insights to maximize their teams’ strategies,” Rostami said. If a digital account is opened, but few features within online banking are used after the fact, that becomes an initiative for marketing or frontline staff — to better educate potential customers about the bank’s offerings.
According to Amplitude’s own data, daily active usage of digital products increased 54 percent from 2020 to 2021. That should come as no surprise, as community banks quickly rolled out new digital offerings to meet customer demands.
Along with this increased usage comes the need to hone in on exactly what customers are doing within the products, and why. In the past, a bank’s sales or marketing efforts might have been a primary source for new business. “Today, digital products have become the revenue center of an organization,” Rostami said. “More companies are understanding the value of behavioral data in bridging the gap between product experiences and customer needs and wants.”
We’ve come a long way from “advertise everywhere to increase visibility.” Today’s digital marketers are smart with their budgets: They target the right channels and search terms.
“As more users go online for their banking needs, we’ve seen higher click costs for many advertisers using Google and social media platforms for pay-per-click ads,” says Brian Reilly, managing director at BankBound, a digital marketing agency. Reilly also notes that search engine visibility for banks is more competitive than ever. “With all the increased activity and competition, data is critical to understanding what works, what doesn’t and where to go next.”
Reilly emphasized that it’s critical for banks to establish goals prior to investing resources in any channel. “Establishing trackable goals to analyze and measure marketing performance is absolutely key to effectively allocating limited resources in an increasingly competitive marketplace,” he said.
Limited resources can be enhanced by combining marketing analytics with product analytics: Targeting customers based on behavioral data. But as of August 2022, this has become an area where banks and digital marketers need to tread carefully. The CFPB issued an interpretive rule that digital marketers must comply with consumer financial protection law. The CFPB and other agencies can sue digital marketers for unfair, deceptive, or abusive practices under the Consumer Financial Protection Act.
Digital marketing today relies heavily on analytics to maximize the audience’s interaction with ads. Couple that with behavioral data (versus demographic, geographic or other more static data points from the past), and marketers can create a more personalized ad experience by targeting segments of people who are more likely to take that next step with a product or service.
While the data is readily available, banks will now need to contend with the CFPB’s ruling and what it means for digital marketing in the future.
Security and IT analytics
While marketing and product analytics focus on customer behavior, security and IT analytics take a much wider view. They seek to identify potential bad actors attempting to breach the bank’s sensitive data or exploit weaknesses within employees that could lead to such a breach.
“There are specific analytics that can often act as the canary in a coal mine,” said Mark Sangster, cybersecurity expert and chief of strategy at Adlumin. “[They] offer a chance to detect threats or eliminate exposures before they lead to a massive operational disruption in the form of a ransomware attack, data breach, and fraudulent transfers.”
Sangster said that criminals use known vulnerabilities to gain access. These could be stale accounts (from ex-employees), passwords that don’t have an expiration date, or employees’ improper access to systems/data not within their role. “The number of open vulnerabilities, the percentage of critical [vulnerabilities], and the percentage patched provide a proxy for baseline exposure and help align resources to reducing business risk,” Sangster advises.
Additionally, several security-related analytics have emerged to help identify evolving threats. One of these is user and entity behavioral analytics, which reviews employee activity and how specific devices (such as laptops) are used. Activities that seem outside the norm can be early indicators of unauthorized access. Threat intelligence is also on the rise, collecting data from numerous sources such as the government, law enforcement feeds, private feeds and internal data. Combined, patterns and signals of suspicious activity can emerge and help teams stop attacks before the damage is done.
Sangster believes that security analytics will continue to evolve. “It’s more about regaining first move advantage, thinking like the adversary, and countering their moves before they make them,” he said. “Today, almost all security analytics are reactive. Tomorrow, they must become predictive.”
If there’s one thing that Sean Kamkar, head of data science at Zest AI wants you to know, it’s that AI will not lead to a robot takeover. “AI’s become almost a buzzword in pop culture that equates itself to something out of ‘The Terminator,’” Kamkar said. “In reality, it’s not scary. Artificial intelligence ‘learns’ generally in a controlled environment, with human oversight.”
For applications, such as Zest AI’s credit underwriting platform, AI-powered insights are heavily scrutinized and heavily regulated. Much of the advancements over the past decade have been from a subset of AI called machine learning, the predictive potential that can underpin many types of analytics.
“AI-powered analytics, fundamentally, is the benefit one gets from using better math and more data,” Kamkar explains. “ML models are more flexible in terms of the kinds of patterns they can identify.” By relying on more data, nuanced patterns can be identified, which ultimately enables better decision-making. “AI-powered insights offer better recommendations regarding how and where to market new products or target new customers,” Kamkar said.
For banks that are looking to leverage AI-powered analytics, Kamkar provided a few suggestions. The results of AI should be “real” for example — based on prior business outcomes — instead of conceptual. And the AI platform should be proven to improve the fairness of decision-making, alongside accuracy. Banks should also ask for model documentation to ensure that an AI platform meets compliance requirements.
“With greater adoption, AI won’t be a topic to dispute in the boardroom,” predicts Kamkar. “AI will be leveraged to address key business problems.”
Closing the gaps in data
While these different types of analytics are powerful on their own, banks should also consider the next level: Combining data from more unified insights to provide even greater insights. If product and security don’t co-exist, for example, the analytics from each platform may not paint a complete picture. Creating a single source of truth and closing the loop on data should be a top priority.
And the reality is that making connections is an art as much as a science. The data is there, and the power comes from interpretation. While AI and predictive analytics can certainly help with that interpretation, the real results will come from asking, “What comes next?”