Reinvention of Spending

The pandemic is forcing finance to innovate faster than ever before

How Northwestern Mutual CIO Neal Sample changed his playbook in 2020.

The pandemic is forcing finance to innovate faster than ever before

CIO Neal Sample adjusted in this unprecedented year.

Photo: Northwestern Mutual

In times of crisis, people naturally tend to worry about their financial plans. Northwestern Mutual's CIO Neal Sample had to make sure the technology powering those plans was one thing customers didn't need to sweat.

Sample, who joined the firm in fall 2019, oversees IT strategy and performance for its network of more than 6,000 financial advisers and for the customers who interact with the company's site. When the pandemic hit, Sample says it wasn't a signal to change course; instead, it was a chance to double down on implementing new technology into both sides of the adviser-client relationship.

"We have this conventional wisdom that says change needs to happen slowly, but we've proven to ourselves that we can change very quickly when we need to," he said. "And that's something that I think, coming out of this crisis, will stick."

In an interview with Protocol, Sample discussed how rapid tech adoption during the pandemic has changed the way the company will approach future tech rollouts, how the company aims to use as much data as possible without being creepy and why sometimes legacy systems are the key factor in getting digital transformation right.

This interview has been edited and condensed for clarity.

How does your product roadmap look different today than it had before 2020 unfolded?

It's not super different, and that's something that I think is really good for us. We've been on a journey to move our experiences to digital. We think of ourselves as primarily providing financial security, and there are different products associated with that, whether you're thinking about annuities or insurance or investment products. But it all falls under the umbrella of financial planning, and so our push has been to move those things into a digital framework so you can get your illustrations — where we talk about the value of our products — for planning. You can see them online; you can engage with them and interact with them when you want to. For us, that doesn't change.

It's been more of an adoption question. We always have a build curve and then an adoption curve, and we're actually ahead of our adoption curve. What that means though, is that our expectations are rising faster and that we also need to build faster to meet those expectations. Our client's expectations aren't really set by our experiences; they're set by the other digital experiences in their life. We find that they expect their online experience to match what they get out of Netflix or Amazon or Apple. There is so much exposure to digital in other channels that our bar is very high, not because folks have a comparison in mind in the financial services industry, but because they compare it to everything else they use every day. And that's the experience of what we're trying to achieve as well.

Externally, have you seen a behavior change that's altered your thinking in the ways you're deploying any of the tech you use?

I think there's been a significant shift to digital not only for our advisers but for our clients as well. Their willingness to engage in digital channels, whether it's a website or a mobile app, has gone up. The readiness is there, and some of that's driven by us and our clients, but a lot of it is driven by the environment. We've generally seen more folks paying attention in times of need, so I think that has an effect on driving digital adoption. At the end of the day, we provide financial security to 4.6 million clients, and [we responded to] their need to check in on that security, to look at balances in a whole life insurance policy or to look at their investment portfolio.

We have a system of financial advisers to help you directly, but sometimes it's not a deep question. It's just to check in or see what their balances are. That's pretty easy on a digital platform — you don't have to make a phone call. You don't have to talk to your adviser. But you can still get a little bit of comfort by checking in.

In a non-COVID world, what did your adoption curve look like? And are there pieces of those normal processes that you'll look to retire post-COVID?

Internally, our corporate headquarters team was primarily in person, so our need for digital meeting tools wasn't as high. Skype was good enough. We made a choice to go to a different platform, and in large part, that choice was driven by the fact that our workforce was distributed heavily overnight. Normally we would have gone through a fairly long change management curve when we rolled out new software. There would be training, there would be a yearlong adoption curve and subsequent retirement curve. In a 2020 world without COVID, we would've had early adopters, folks who are comfortable with what they've already got and a laggard class. We would have had stage gates and analysis on when to push. This time, we opened it up to an opt-in model, and basically everybody opted in.

I think there's some fundamental changes that have now happened, and some of those changes are not just process and tools, but appetites. On a typical day from a home office, we might've seen a few hundred folks working remotely on VPN. Overnight, we went to 5,000, 6,000 or so, and then another 15,000 or 20,000 from the field. Those were dramatic differences, and we didn't know how that kind of shift would work. We have this conventional wisdom that says change needs to happen slowly, but we've proven to ourselves that we can change very quickly when we need to. And that's something that I think, coming out of this crisis, will stick.

How does the cloud play into that digital transition?

That's something we've been working on for a long time. We have some applications that are cloud native, [but] on the other hand, we also have some of the oldest legacy technologies you could imagine. We have some mainframes in our data center. We have some really great modern development practices in the cloud using DevOps, but we've still got big, legacy investments in more traditional technologies. We're living in a tale of two cities right now.

The migration is the hardest part. Writing new native applications is fairly straightforward, and there are a lot of good tools and good practices. But moving the legacy to the cloud is a bit more complicated. And this is definitely a journey that we're on. I would not call it a solved problem yet.

When you're making that migration, where do you see the opportunity to do it wrong?

I think I've only made up one term in my career, and that is "Chernobylizing." You take a legacy system that's radioactive, and you encase it in concrete. You leave it there until it's run its useful life and then you shut it off. That's where I've seen the greatest misstep: There are certain transactional systems that do well on the mainframe that are using legacy data. They have really high reliability and availability assumptions. They're primarily back office and batch processing. They're not designed for real-time or the flexibility or elasticity of the cloud. Those are the types of applications where I've seen people stumble, where instead of being thoughtful and having a lot of fine detail about what goes to the cloud, they say, "we're going into the cloud." They try to wrap their arms around all of the legacy and decide that everything can go. And I don't think that's true.

There are some purpose-built applications that run on mainframe, for example, that are tuned for it and are ideal for it. And they're going to run poorly in the cloud. Just because it's a more modern stack doesn't mean that your application architecture is ready for it. People should ask themselves: How is it that we have this technology that's been around for decades and decades and won't go away? Some of it is the switching costs, but in other cases, you wouldn't move off it because it is the right tool for the right job.

What are some examples of the applications that don't migrate well?

At American Express and Express Scripts [Note: Sample previously worked at both], we also had significant mainframe investments. And the one thing didn't move were the real-time transaction processing systems, the ones that required ultra-reliability. For example, in a pharmacy transaction, once a script comes into a pharmacy, you have 800 milliseconds to figure out the billing information, the safety protocols and the drug utilization review, and then return that information back to the pharmacist. At American Express, it was as simple as being at the till and wanting to swipe your card. [Both had] very tight timeframes and ultra-reliability requirements. Those are the kinds of things that are kind of perfect for the mainframe.

The alternative is an elastic cloud that [only] maybe achieves high reliability because components can fail out. But you can't fail out in the middle of a financial transaction or in the middle of a pharmacy transaction that has a safety component to it. It's not like loading a web page. It requires fixed transactional integrity. And those were kinds of problems for which a mainframe is the perfect tool, not just a tool that is good enough.

Shifting gears, financial services is a data-rich environment. What's next for you in terms of data accessibility and strategy?

For us, the idea is to know everything about our clients so they are well-known and well-serviced, regardless of how they show up. They may show up to their own digital experience on the web, they may be talking to their financial advisor, they may be calling into a service center with a question. The customers should be fully known in all of those domains.

When we think about data, at the heart of it is that golden customer record. We understand all of their transactions, all of their products, all of their demographics, all of their needs. And it's always available so that we don't feel the customer is a prism and we're only seeing one facet of them depending on their activity at the time. The next step is, what do you do with that?

Our real focus is on next best action: how we can service the customer in a way that they want to be serviced before they even come. In a lot of cases, we have data feeds coming in that show something as simple as direct deposits. And when we see that it now comes from a different source and it's a different amount, that indicates a job change. Maybe that job change is a good trigger to check in with their financial plan. That way, we don't even need the client to tell us; we can say we've already done the work for you. That's the kind of thing where we can do predictive needs analysis, and that is a really powerful outcome from a data program.

In developing the predictive analytics framework for clients, what's the bellwether of success?

I think we have examples from other industries of both success and failure. Look at Netflix. They've spent a lot of time predicting what their viewers will like and have managed to keep them very happy. Or look at Amazon and their recommendation engine. Their ability to anticipate is very high. On the other hand, if you look at other online retailers or you look at advertising, sometimes they can go into the realm that I would call creepy. When you go to an auto manufacturer's website and price out a car, that stuff will start following you around the web, and it feels a little bit invasive.

We've got some idea of what it looks like to meet customer expectations in a relatively friction-free way that adds value, and we've got some good examples of where it goes a little bit far. We have to be sensitive to our right to play. When someone has a financial adviser, the expectation is that the financial adviser is on top of it, so they should anticipate needs and be able to make those suggestions. And that's one of the things that data allows us to do in a way that meets the customer expectations — which is different than some of those examples of things following people around the web and showing up in places where we're not expected.

Are people generally open to having that data enhance the adviser relationship, or do they shy away from it because of the sensitivity?

I think it varies actually quite a bit. Not to over-generalize, but, for example, older clients tend to be less digitally savvy and have fewer digital expectations. They may not use the website or an app. They may want to exclusively talk to a financial adviser face-to-face or over the phone. On the other hand, our younger clients are ready to jump into the portal. They're ready to do things like link bank accounts or credit cards and mortgages and see all of their financial life in a digital dashboard. And they want all of their information to flow into and be relevant to their plans because they're digitally native. The art for us, which hasn't changed with COVID, is to meet our clients where they want to be — to not go too far for a certain group, but also to not fall short when there are very high expectations. We have to be very digitally savvy, but not digitally pushy. And there's a difference.

More from Reinvention of Spending