Driving Customer-Centric Transformation in Banking

Sonali Divilek at UXDX USA. Video: https://youtu.be/lQsmkvziezs

Readable transcript: edited from the recording's captions for readability (fillers and false starts removed, punctuation and section headings added). Wording is the speaker's own. Timestamps are positions in the video. Names marked [?] could not be verified against the audio.

Digital banking at Chase scale

[00:00:00] Good afternoon, everybody. Thanks to UXDX for giving me the opportunity to speak to you all today. As Joe said, I'm Sonali Divilek. I'm the head of digital at Chase. I came to the firm about three years ago and have built my career primarily working in the financial services industry, building products for consumers and small businesses. Today at Chase, my team builds the features and the experiences you see when you open the Chase mobile app or when you go to chase.com. Our job, as we see it, is to power the digital platform that gets customers to their full financial potential and helps their dreams come true. So that's our why.

[00:00:47] We are the front door to all of the interactions, or at least the daily interactions, that most customers have with the bank. Over the past five years or so, we've seen incredible adoption of digital banking. Industry research tells us that people are using their banking apps weekly, and at Chase, since 2019, we've actually seen 10% year-over-year growth in our digital active users. So we spend our time making sure that those experiences are always on, that they're performant, and that our customers are safe, engaged and satisfied.

[00:01:30] Today I'm going to walk you through how we actually ship product at a place as large as Chase. We have over 300,000 employees. How do we do that and still keep the customer at the center of everything we do? I'm going to start by describing the enormity of the scale we have at Chase. We serve over 80 million customers and more than 6 million small businesses. We have 4,800 branches in every single state, and that's 900,000 customers walking into our branches daily.

[00:02:14] On the digital side, we have 68 million customers engaging with our mobile and web experiences. That translates into 12 million unique app users every 12 hours. It means 15 billion login sessions per year. It means 26 million customers coming in each year to use Zelle, just to get paid or pay somebody else. That kind of impact is really staggering, and it also creates a tremendous responsibility for us to make sure we have our customers' backs. We are working on their finances, which are critical to them, so we take that very seriously.

[00:03:00] People are also using our channels for important life moments. They're looking to buy a new home, they're looking to plan for retirement, and they're using our savings and budgeting tools to improve their financial health. We care about what we call engaged sessions: people not just coming to the app, checking their balance and leaving, but actually engaging in some other service. Since 2019 we've seen that increase by 25%.

Investment and the quad

[00:03:33] I have to point out that delivering on that requires tremendous investment. In 2023 alone, we spent over $3 billion [?] on our product and technology agenda. A significant portion of that is focused on infrastructure: making sure we're on the most modern tech stacks and being really flexible with our platforms, which increases the resiliency and stability of our experiences. But what's exciting is that more than half of it goes into delivering new products and features that serve our customers. We're helping them redeem rewards points, book travel, start their own small business. That's a really dynamic place to be playing.

[00:04:35] That investment also powers a huge organization. We have 25,000 people across our product, experience and technology focus, and we have that across disciplines: product, design, technology, and data and analytics. At Chase we call that our quad.

[00:05:00] When we think about our quad, it's really important that each of these disciplines comes together and starts together to solve the crux of the problem. We're not working in silos and working sequentially. We're coming together and saying: what is the crux of the problem, how do we solve it, what should we build and how should we build it? It's important to do it together, because we all bring a different perspective.

[00:05:24] The first is our engineers. They are thinking platform; they're telling us how to build something. We have our designers, who are responsible for the end-to-end customer experience: how is that experience going to make the customer feel? We have our data and analytics team, and it's really important that they're there to instrument the products we build so we can actually learn from customer interactions. We have 17 terabytes of data. I never knew what a petabyte [?] was, but it's a lot of data, and we're using it to come up with insights and improve our products.

[00:06:07] And then you have product, which is me and my team. I think of us as the band leader. We're trying to get the whole group together to play the same song. All of us are really in it together, and culturally that tight relationship is really, really important. As a quad, we work together to keep the customer at the center. We are listening to customers: we do call listening, we look at customer research, we do focus groups. We have to be deeply curious about their problems and really fall in love with those problems, not with some solution that we think is going to be perfect. And then we fine-tune. We iterate. We have the humility to know that what we put out there isn't going to be perfect, and we allow for that.

A culture of experimentation

[00:07:02] Critical to doing that is this culture of experimentation. We're building a culture where we measure, we fail fast, we keep testing, and we rinse and repeat. Gone are the days when you would say, "Okay, a year from now there's going to be a big launch, and here's the product, and we know exactly what it is." We really have to change that kind of mindset, and that is infused within the teams.

[00:07:35] And it's not just about product features. I think our engineers operate this way too. We have something we call True CD at Chase. I think in the industry it's continuous delivery, continuous integration. I don't think I have that fully right, but it's a way to automate your testing practices across your applications. I think it's really analogous to experimentation, because we're taking one technology team, and they are writing the unit tests for their code to make sure it's going to deploy correctly, but they're also doing contract testing with the other tech team, making sure that when their code connects, it's actually going to deploy the right way. Breaking it down and making sure that everything works end to end helps to de-risk what you deploy in the end.

[00:08:30] Actually, let me go back for a second. I covered this, so we'll talk a little bit now about experimentation and expectations. We know that when experiments are put out, they're supposed to fail some of the time. By industry best practice, a third of your experiments should win, a third are going to be flat and a third will fail, and you should expect counterintuitive outcomes.

Rethinking the open account button

[00:08:57] Let me give you an example of this at Chase. Like I said, we have 50 million plus customers coming into the app every day. Our mobile app is actually a great channel for customers to come in and see the full breadth of what Chase can offer. We have tons of products that are relevant to our customers. But in 2022 we saw our mobile app account production start to dip. So we took a look at this "open an account" button that you see here, which is a huge driver of that traffic.

[00:09:33] We asked why this was happening. We went to the data. We started asking customers how they would go about exploring getting a mortgage or planning for retirement. What we found is that this "open an account" button was confusing. Customers thought of it as where you get a credit card or where you open a new checking account. It didn't occur to them that there was so much more to explore.

[00:09:58] So we came up with a bunch of variants. We had "open a new account," "shop products," "explore more with Chase." We tried a bunch of stuff, and here's some of what was interesting. "Open a new account" actually performed flat, but when we dug underneath it, we found that for customers new to Chase, performance was down, because they were like, "I just opened a new account, why would I open another one?" or "I don't even know what else Chase has to offer; it doesn't occur to me." Similarly, when we used the word "products," in any of the language there, performance was down. People were like, "What's a product? I don't know what a product is." That's us talking to ourselves.

[00:10:45] In the end, we took the greatest hits of everything we learned and put up this tile that showed the different products. We call it our product lineup. It probably isn't earth-shattering, but it has led to a 13% increase in our account production, and I just told you about our scale; that's massive. And by the way, this isn't even how it looks anymore. We continue to iterate. But those are the kinds of small, incremental, game-of-inches changes that can drive huge ROI and huge results.

Score Planner

[00:11:29] I'll walk you through another example. This one is a new feature that we were able to bring to market, and it was based on one of our great digital tools called Credit Journey. We launched Credit Journey several years ago. It's a credit monitoring tool that also has identity monitoring as part of it, and you don't even have to be a Chase customer to use it. What was interesting, though, is that as we talked to customers, they said, "Cool, that's my credit score, but how do I improve it? What can you actually do to help me?"

[00:12:07] We looked across the marketplace and found that all the credit score tools were doing this very generic thing. They had one of two models. One model is "people like you": it looks at people who have the same score as you and gives generic advice about behaviors you can adopt to try to raise it. Another model is "people like them," which says, "Here are people with 750 credit scores; you should do what they do." But in all cases, it wasn't really specific and it wasn't really helping the customer.

[00:12:40] I love this quote here, so I'm going to read it: "I feel like how to increase your score, and by how much, is a true mystery." Anybody can Google how to improve, but how do I know this is going to impact me? So we had a clear why, and then the teams got to work. With a focus on design and functionality, Score Planner was born.

[00:13:08] Here's how it works. The customer starts by seeing where their credit score is, on that spectrum of fair to excellent. We then give them a prompt: here's how much we think you could move your score. The customer then tells us why they want to improve their score; maybe they want to buy a house. And we tell them how long it's going to take and by how much they can improve it.

[00:13:37] We launched this and tested it with our customers all along the way. The team gathered a whole bunch of usability concerns. People asked, "Are those action steps clear?" There were some challenges with the navigation. All of those things we were able to fix by really listening to customers and putting this out in market.

[00:13:59] Score Planner has been out for about a year. Three and a half million plans have been created. More than a million customers have actually completed their score plan, and some of them sign up for more, just racking up the credit score points. On average, when customers complete their plan, they improve their score by over 30 points. So it's a really compelling and rewarding thing to be a part of. You're improving the financial health of your customers. And we're still iterating on it. There are always tweaks; we keep going back to it to try to make it better.

Leadership and takeaways

[00:14:38] This is a little bit of a Jamie Dimon plug, but what I will tell you is that when you want to do this experimentation, you really need that culture shift to start from the top. This is our chairman and CEO, and he reminds us: it's about the data, it's about the facts, you've got to do the analysis, rinse and repeat. What I love about what he says is that you have to appreciate that failure is a normal outcome and something you're going to learn from.

[00:15:11] Before I end, I'm going to foot-stomp a couple of things that I talked about today. One: make sure you know your why. What is the problem you're trying to solve? Keep reminding yourself what that is. Two: measure and celebrate the outcome, not the output. Don't fall in love with your solution and just get it to market; you really want to make sure it's moving the needle. Three: you need to constantly experiment. Your ROI will improve, and your customers will be happier. And finally, know that you're never done. Product development is a living, breathing thing, and you just need to keep iterating on it and bring the best to your customers. With that, thank you so much.

Q&A

[00:16:06] Host: Thank you. There we go. That was fantastic, as promised. I in particular found the quad really interesting, and I think everyone in the audience took away quite a bit when it comes to the measurement of those outcomes and the focus on them. I'm curious: the quad as a concept is going to be familiar to many of us, because we typically share the framing of the three-legged stool. How does the quad help to accelerate outcomes in an environment like a bank, which, no offense to anybody, is not really known for moving that fast?

[00:16:52] Sonali: It's a great question, and actually when I came to Chase it was something I was really surprised by: how do you ship so much and get so much output? I think at a place as large as this, being really disciplined about your OKRs, and what metrics you're trying to move, is what allows people to row the boat in the same direction. We actually revisit our OKRs on a very regular basis. We do it as a quad. And you can imagine it's not one quad of 25,000 people; we have a bunch. But we make sure we're lining up each product area against those OKRs so we know we're moving the needle. Listen, it's not a 15-person startup, so we do have a lot of coordination. But I think holding yourself to the metrics you need to move helps everybody do it together.

[00:17:49] Host: Absolutely. For those of us who have worked in the large-scale financial and legal space, there are different tactics to employ when driving outcomes that are somewhat constrained by compliance and regulatory bodies. Can you give us a little advice on how you've dealt with that in the past?

[00:18:11] Sonali: It's a struggle.

[00:18:14] Host: I know. It's not like I threw you a softball.

[00:18:19] Sonali: I think there are two parts to it. One: we're dealing with people's finances, so it's just not an option not to have a strong approach from a risk management perspective, to make sure you're keeping customers safe, making sure they're not getting scammed, and to hold that responsibility. Emotionally, with the team, we talk about that quite a bit, so that people actually feel those rules are there for the right reasons. Similarly to how I said to get the quad together at the beginning, I do bring my control partners, my compliance and my legal folks in on anything we start, even this Score Planner stuff, because you're looking at people's bureau information, to make sure we're thinking about those implications up front.

[00:19:07] Sonali: The other thing I will tell you is that we look at experiences that are not just in the financial services industry, because we should be pushing ourselves. We should be looking at those experiences and saying, "Yeah, but they got to do that. What made that fintech able to do it, and is there a way to do that in a way that's safe for the customer?" So it's definitely a thing.

[00:19:30] Host: So, correct me if I'm wrong, it sounds like you extend the quad and bring them in.

[00:19:36] Sonali: Sure, definitely. And I would say your lawyer probably wants to be in it with you. We're doing the fun stuff. They want to be there and they want to help, and when you bring them in as part of the problem solving, I think you get a better outcome.

[00:19:50] Host: It's a lot more fun than doing discovery, for instance, right? I want to double-click a little on the notion of team structure. You talked quite a bit about acceleration and shipping. Is there anything you've found to be unique about team structure that can drive some of these smaller-footprint experiments and iterations to market?

[00:20:21] Sonali: The way we structure it is that nothing is autonomous, but we have carved out, I think, about 90 different products and product areas, each with a head of product, a head of tech, a head of design and a head of data and analytics. So they are somewhat that quad. People in those areas don't think of themselves as "I'm on the design team." Their allegiance is to the product they're on. I think those smaller groups that can work somewhat autonomously and run their own experiments actually help get the outcomes.

[00:20:57] Host: Gotcha. So it sounds like it's the same quad model, but shaped a little differently depending on what the business need is.

[00:21:07] Sonali: Yes. And honestly, your allegiance being to your area, not your functional discipline, helps, because it really brings collaboration across all of the expertise that each of those members has.

[00:21:23] Host: Love that. It seems like we're about at time, because I'd hate to throw another hardball at you and say you have 45 seconds. So, Sonali, thank you so much. It's really been fantastic to have you here.

[00:21:38] Sonali: Awesome, thank you, and thanks, everybody.