Did you know a website visitor using Edge, Microsoft’s browser, is eight times more likely to convert on a CD than someone on Firefox? Or, that Firefox visitors are 10 times more likely to apply for a base checking or savings account than for a CD?
Neither did Visions Federal Credit Union three years ago, but today it’s using a host of similar data to drive much higher applications for membership and loans online.
Beginning in 2020, leaders at Endwell, N.Y.-based Visions noticed a change happening in the organization’s interactions with members. Branches had been the dominant meeting place, but as the pandemic years went on digital channels grew to handle about 50% of member interactions. Before they had handled only 26%.
Face-to-face visits were the institution’s best means to serve members through staff listening and responding to their needs. Digital channels, on the other hand, the website in particular, needed to become more human. It could not listen to people’s needs nor respond in a way that came close to simulating the branch experience.
The Winds Are Changing:
Before 2020, only a quarter of member interactions happened in digital channels for Visions. Just three years later, digital channels see 50% of the credit union's member interactions.
Through a partnership with machine learning company Finalytics (which specializes in automating personalized ads, articles, and webpages), the $5.5 billion credit union set out to become an advisor through web content to its members using data made ubiquitous by e-commerce, and enriched by the unique data only available to financial institutions.
In March, Celent recognized Visions as a model institution for customer engagement. It had used web personalization to drive three times more click-through rates on marketing content, three times more member applications, and four times more loan applications, compared to non-personalized engagement.
Where Did Visions Start on the Personalization Puzzle?
Virtually every financial institution in the country has invested in, or now considers, some type of “digital transformation” with personalization as a primary goal. Leaders know these projects face challenges in creating results. They can be similar to finding a cocoon at the park: You want to see the butterfly emerge, but you have no idea how or when that will occur.
Visions was able to demonstrate a business case for its web personalization. But how did it get from aspiration to reality?
To start with, it was following a real need; foot traffic was becoming web traffic. They also followed a bigger shift that’s changed shopping of every kind: People now turn to websites as the intuitive first step in researching anything, including financial institutions. Members can walk through a retail location without sharing needs with staff, but they show intent when they click through information on a website.
Strong Growth Can Start With Identifying Channel Traffic:
A major component of personalizing the banking experience for customers and members is understanding where, when and how they interact with your brand. Do they start their journey online or in your branch?
“Banking traditionally has been very transactional; it’s been a ‘be there when I expect because I have this transaction or basic task’ mentality. We didn’t know people needed us until they literally spoke to us,” Tom Novak, Visions chief deposit and payments officer, tells The Financial Brand. “Now, members and prospects are first interacting as they research us digitally, before they ever consider any other type of interaction. And they may stay fully digital.
“It’s like this new untapped opportunity in serving members that’s been waiting there,” he explains. “We wanted to add value at that digital touchpoint at the moment of need.”
Novak described the personalization opportunities available from website visits as a “treasure trove” constituted of data commonplace in e-commerce today. Unlike other areas of banking where the industry has to fund unique de novo technology, web data is decades deep in SaaS investment to improve matchmaking between audiences, advertisers, and marketing content. Behavioral data from a website, as well as search engines, social media, and browser cookies – referred to as third-party data – are the most accessible to any organization today.
“I think of it like a puzzle where you set the edges first,” Novak says. “Digital interaction data can be an easier edge to set in a lot of respects. It’s inherently less of a lift, although there is work there, when you’re dealing with mainstream SaaS platforms like Google Analytics.”
Banking Personalization, A Logic of Its Own
Visions could not have doubled or quadrupled web conversions on third-party data alone, however. Data just isn’t enough; it needs to be interpreted correctly for the unique aspects of banking.
For example, the personalization used by e-commerce does not look very intelligent if copy-pasted into banking, Mark Ryan, chief analytics officer at Finalytics, tells The Financial Brand. Retail shopping engines “are built around a really basic kind of logic. If you buy your son a Lego Millennium Falcon, then they’ll recommend another Star Wars Lego product, a Star Wars non-Lego product, or another Lego product,” he says. “For selling auto loans that logic just doesn’t help. If you go get a new car loan, the institution shouldn’t then promote a used car loan to you.”
Ryan has worked for more than 24 years as chief data strategist for Extractable, a digital and data analytics consultancy that’s served clients such as Visa, Charles Schwab, PayPal, and Wells Fargo. Two years ago, he founded Finalytics with its chief executive Craig McLaughlin, who also sits at the helm of Extractable.
“If you go get a new car loan, the institution shouldn’t then promote a used car loan to you.”
— Mark Ryan, Finalytics
A unique personalization logic for banking starts by observing people as they vote with their clicks, Finalytics has found. Trends then emerge that can tie together seemingly unassociated data points. Browsers tell institutions a great deal about a shopper, for example. “If you visit most institutions’ websites, they’re promoting a certificate,” Ryan says. “But a visitor on the latest version of Firefox has a 0.1% chance to convert for a CD whereas an Edge browser has a 0.8% chance. When someone first visits, we may only know their browser, but they’re all we need to know when promoting a CD is a waste of your time.”
Much of banking personalization has logic like this CD example. An institution might know that a visitor wants a CD because they’ve visited CD content before. But otherwise, especially for anonymous visitors, demographics, time of day and location – data gathered from web visits and third-party data – are tested and analyzed to correlate it with marketing content and conversions.
A Digital Frontier in a Familiar Marketing Channel
Defining segments traditionally has been a horse-shoes-and-hand-grenades affair. Visions began looking to discover, rather than guess, visitors’ needs. “We’ve started acting like scientists,” Novak says. “We formed a hypothesis, maybe from a belief or an intuition, then we looked to see if our findings corroborated it.”
The credit union serves communities in New York, New Jersey, and Pennsylvania with 60 locations. It thought people were visiting the brand from all over its states. “What surprised us was data showed that was not the case at the time,” Novak says. “We expected more prospects than members and that we were increasing our brand awareness and coverage. Instead, we were reinforcing our existing members’ loyalty.”
Discovering that most visitors are members, behavioral insights from Finalytics could open doors within the member base. “Bank and credit union websites, just by coincidence, are structured like a survey where the question is: Are you a retail member or a small business?” Ryan observes. “Visitors get segmented nearly immediately in their research process.”
“What surprised us was data showed that was not the case at the time. We expected more prospects than members and that we were increasing our brand awareness and coverage. Instead, we were reinforcing our existing members’ loyalty.”
— Tom Novak, Visions Federal Credit Union
Imagine a member who’s had a small share account since they were a child. Now as an adult, they’re a real estate developer, restaurant owner, or independent consultant, and they want to compare their long-time credit union against a different institution they’ve used for their business. Visions can pick up on that activity.
Small business visitors behave very differently than retail members; someone looking for a personal loan tends to view three to seven pages whereas small business loan researchers tend to view around 30 pages.
“Business visitors are also much more likely to do an onsite search,” Ryan reports. “People are many times more deliberate when they’re looking for a product for their job or business versus for their personal finances.”
Learn More:
- A Few Simple Tweaks to Banking Websites Can Increase Sales
- What Bankers Should Know About Google’s Helpful Content Update
- Need a website development firm? We have a list of them here.
Personalization At Scale
As an institution develops its personalization, segments multiply quickly because it can become extremely prescriptive.
Visions found, for example, that web visitors between ages 35 to 65 are most likely to click through to an application for an RV loan when they are using Microsoft Edge in Pennsylvania.
In the case of RV loans, pictures of people smiling next to their RV work well, but that just one type of visitor. Alter any element of the segment and you may have another new and valuable segment with entirely different image or message needs. Visions has about 252,000 members.
Even if it had only one segment for its eight retail deposit account types, and its 12 retail loan options, it’s staff would need to manage a gargantuan library of images, ads, and webpages to serve those segments on the web. And, all of it would need to load to the website during visitors’ browsing sessions. It’s just not possible for human hands. Only automation makes scaled personalization possible.
Software made to learn valuable segments and delivery the right content is required, Ryan says. Otherwise, it’s impossible to deliver high-multiple ROIs. “Every credit union and bank we’ve talked to is already at capacity in terms of coming up with content and creative, and the new development required to keep up with personalization,” he says. “They need an engine using machine learning.”
Worth a Thousand Words
When Celent studied Visions, it noted the impact of seemingly small personal touches, such as the age and demographic of people shown in ads and images.
Photos and content about the visitor’s local community has an emotional impact, says Bob Meara, senior analyst with Celent’s banking practice. “If you sponsored a non-profit in Poughkeepsie, N.Y., and your website visitor sees content about your commitment to their community, they care about what’s happening in their backyard,” he says. “They don’t care about a static announcement of something the institution did in Syracuse. Over time, that kind of interaction can be acutely relevant and generate positive feelings.”
Location data available through IP address allows web art to create an identity connection with a site visitor, sometime in ways they may not even notice. “For example, if someone’s IP comes from an army base, people in the military don’t use car loans to buy Mercedes Benz,” Ryan explains. “You want to show them ad imagery and messaging consistent with a military member; that’s usually imagery of a soldier leaning on a truck.”
When Visions implemented Finalytics, it rolled personalization out to about 12% of its members and showed double to quadruple the conversions compared to members not receiving personalized content. Now, it’s rolling website personalization out to the full membership this year. The credit union is still on its journey to provide humanized digital personalization to members, but it has found early wins for members and for the organization.
“If you drive on a trip from New York to Florida, there are towns and cities that come first as you reach mile markers in your journey – you can see the progress and benefit from reaching each mini-destination,” Novak says. “We haven’t arrived by any means, but serving members on our website was a very logical place to achieve our first milestones.”