Personalization in Banking: A Conversation with Zack Anderson

Full Transcript

Part 1: Why is personalization in banking broken?

EDUARDO ROMA: Hi, I'm Eduardo Roma. I'm a partner in our financial services practice, and at Bain I lead our customer experience practice globally. I have the pleasure of welcoming today a friend, Zack Anderson, who is a customer-centric champion with experience working across multiple geographies and sectors. We're here to talk about personalization. Welcome, Zack.

ZACK ANDERSON: Thank you. Great to be here.

EDUARDO ROMA: So, we've traveled around the world, spoken with many banks, and arrived at this conclusion: that banks have been training customers to ignore them. How did that happen?

ZACK ANDERSON: I think it comes from the fact that banks are so product focused—and regulators reinforce that—as opposed to being customer- and client-focused. As a result, they have, over time, just spammed people with lots and lots of messages related solely to their products, especially in their owned and earned channels, which should be their most valuable channels. And even when we talk to banks about personalization, they're thinking about personalizing the product offer, not really personalizing the communication and the helpful messages they can give to their customers.

EDUARDO ROMA: We’ve talked about engaging customers in the moments that matter. That requires understanding customers at an individual level—not only who they are, but their intentions, their preferences, their individual trajectories. From your experience as both a customer-centric champion and a technologist, what are some of the ways banks can develop that capability?

ZACK ANDERSON: You start with the demographics and what you know about the customer from onboarding and KYC [know your customer]; that's a rich place to start. But the second really valuable source is transaction data, especially for retail customers but also SMEs. Understanding how people are spending is really informative. And then the third piece, which is underutilized, is telemetry—how people are using their apps, where they're using them, what they're doing. That's a rich set of information, but you have to be really thoughtful about how you're using it. It's an insight into a person or a company's almost entire existence—you can see where they're buying things, where they're getting their coffee, all of that.

I think it's helpful to frame it around helpfulness: If you're using that information to help somebody reduce their anxiety, gain control over their finances, save money, or get a better financial deal, your customers will go right along with you. If you're just using it to personalize a product pitch, that's where it starts to feel a little creepy.

I remember running a UX session with a small business customer who was really frustrated. She said, "Look, you have all of my data. I am struggling to make my business profitable. You're experts in finance. Help me figure out how to run my business better." She was just trying to help her own customers, and she wanted the expertise of the bank to help her succeed. That really motivated me. Because of the knowledge banks have about their customers, about data, about the economy, they have a responsibility to be genuinely helpful. To do that at scale, especially on the retail and small business side, you need to do it digitally and with the data you have.

This was just coming out of Covid, so you can imagine the stress on that business owner. Her message always stuck with me: "I give you all my data, you have all my money, you have all this knowledge about something I don't know--help me succeed." If you come at it with that kind of curiosity and alignment with what customers want and need, you can do some great things. And of course, you have to be compliant—if people opt out, if they don't want you using their data, you turn it off.

But we found that the better we did our job at being helpful and personalizing, the fewer complaints we got, the fewer opt-outs we got, and the more engagement we got, every time. From a mechanical point of view, that increases the audience for your messages. If you're poor at it, it becomes a downward flywheel. If you're good at it, it becomes an upward flywheel.

Part 2: How to align on helpful messaging?

EDUARDO ROMA: Helpful messages — it's really a big shift from an organizational perspective, from a leadership perspective. We have product teams with commercial objectives, and we need to get aligned on delivering helpful messages to customers. How do you get leadership teams over the line on that shift?

ZACK ANDERSON: People are used to thinking about marketing and cross-sell/upsell as a megaphone: You blast it out at the customer and hope some of them respond. And the reality is that just doesn't work, especially in banking. Teams are very used to thinking about audiences—they’re used to buying media, so they buy audiences. They want the messages they've worked on and believe in to go out to as broad an audience as possible, and they do care about conversion rate. But when we've talked to a number of organizations, the conversion rate in people's mobile apps in banking is often 1% or less. Which means 99% of the people who see the message, you just burned credibility with them.

So, one of the things we did was simply start reducing the audience size. We worked hard on improving the targeting packages and reducing down to who really responded to the messages. We were driving to get to 5 to 10% engagement rates. And we actually changed the metric to number of engagements—where are we having engagement outside of the regular flow?

It did take a big management push. But one of the things that was effective was actually demonstrating the results: As we took those messages, reworked the targeting packages, reworked the copy, we drove up engagement rates and actually increased sales. That proved you could have smaller audiences and still see cross-sell go up.

As we started to add in content that was helpful, the real estate became even more effective. That created another kicker on top of the better-targeted sales messages, because people looked at it and you could see the gains happening over time in real sales. At NatWest, their absolute cross-sell/upsell numbers went way up, as did the depth of engagement with products and overall digital engagement.

Part 3: How do you scale the content engine?

EDUARDO ROMA: Thinking about helpful engagement in the relevant moments that matter to customers—the sheer level of scale and velocity of learning required to achieve that is massive. How do you scale the content engine?

ZACK ANDERSON: I think the problem is that banks are not really content engines. That is the big difference between [the gaming industry, where Zack worked previously] and banking — game companies are content engines. Banks don't really live in that world. They have marketing departments that run marketing campaigns; they don't create helpful content by nature.

As your inventory expands—as you create more ways to engage customers through the mobile app, the website, and even in branch—you need a lot more content. We estimated literally a 1,000% increase in content. And the cycles of big teams with very long timelines that were designed to put a TV ad into market just don't work when you're trying to put a copy-based or simple message into a mobile app.

We really had to start rethinking the whole cycle of content generation. The scale of it and the learning speed become the big impediment to driving up engagement and then driving real cross-sell/upsell.

Our big learning—and I think this was a surprise to a lot of people—was that nobody really knew how long the full cycle took. The marketing team knew how long the creative piece took. The compliance team knew their SLA for approvals. Everybody knew their piece. But when you mapped the whole thing out, every executive in the room said, "I can't believe it takes that long." And then you say, on top of that, that's how long it takes to create one. And when it doesn't work, your next learning cycle is just as long. Everybody walks out of the room depressed.

So, we started figuring out radically how to decrease time-to-market and decrease team size at the same time. This is where AI comes in: There are so many great ways in the marketing process to use AI to speed up the process, to empower people, to decrease the headcount required, and therefore to increase the number of learning cycles.

Part 4: When does technology slow things down?

EDUARDO ROMA: You talked about technology and where AI in particular can be helpful in rethinking workflows and the skills required to run them. But there's a lot of technology deployment that isn't helpful. What did you see when talking to banks?

ZACK ANDERSON: That was my surprise as we were talking to a lot of the banks — almost everybody was doing some large new technology deployment, and the technology was really viewed as the way they were going to speed up the work. But in our experience, what we saw was that if you use the technology really, really smartly, it can speed things up, but it's almost about the org and operating model first, and then fitting the technology to that, as opposed to the other way around.

The instinct is to go buy a solution that's going to speed things up. And the purchased solutions will speed up certain aspects of the program. If you want to do copywriting faster, LLMs are great at that, and you can do lots of versions of copy. But that doesn't impact your compliance side, doesn't impact these other pieces. So, you really need to sit down, reimagine the jobs to be done, and then apply the technology to each of those spaces. That can really unlock a transformation, changing it from a many-months-long process to getting it down to a fraction of that time.

Part 5: Advice for banks trying to reset the personalization journey

ZACK ANDERSON: In all the journeys I've gone on through this, and the many I've talked to people about, I think it's really helpful to reset what you think a successful customer looks like, and come at that with curiosity.

A lot of segmentation in retail banking is done essentially by income, age, and eligibility for products. The tool I've always used to reframe that is customer lifetime value. If you can start to think about what's the likely lifetime value of a customer and what's their journey to realize that value, then a couple of things happen. As a management team, you extend the horizon you're looking at—not just the next transaction but the whole relationship. And second, you start thinking about the steps that aren't necessarily about selling a product.

Mortgages are a great example: by the time it's time for the customer to take the last step—purchasing the financing to get the home—you’ve already taught them what a down payment needs to be, what they can afford, how they might get money to augment their own savings. And you're in a great place to sell that product.

Then I think the other piece is to think about moments. There are definitely moments in people's lives where banks matter: when you get married or move in with someone, when you buy your first home, when you buy a car, when you retire, when you get divorced. All these life moments are moments when financial entanglements matter. If banks can be helpful in those moments, you're really viewed as a trusted partner.

And it creates this lovely positive cycle. The colleagues like doing it, they like helping customers, they like thinking about how to build messages that help people. The customers like it, they build trust, they buy more products, the bank becomes more profitable, and the clients get what they need. They feel in control with their bank. That upward flywheel, if you can create it, just doesn't stop. It keeps going and gets better and better.

EDUARDO ROMA: Thank you, Zack.

ZACK ANDERSON: Thank you. Good to see you.