We spoke with Jake Tupa, SVP of Technology Innovation at Choice Bank about data innovation in banking.
Watch the whole conversation in the video, or read an edited transcript below:
Choice Bank, a regional bank in the North Dakota and Minnesota area, has been able to stay at the top of their game and be competitive in a market that’s not an easy one to be in. Their technology approach has been to bridge the traditional way of doing things with the modern FinTech way, and Jake began by explaining how they go about that:
Jake Tupa: Choice is a smaller bank by asset size, but one of the things that we try to make sure we do is to have a best in class technology stack, and really that we can compete and punch above our weight a little bit in terms of the resources we have and what we're doing when it comes to technology.
We've been working on how we can bridge some of the legacy tech with modern requirements and build and innovate and iterate quickly on some of those platforms that were designed for some very specific use cases a long time ago, and now we need to bridge the gap.
CloverDX: What do you think would be the biggest challenges that you see in a in a modern world of banking?
In the financial tech stack, nothing really ever gets sunset or stops being supported. If you look at the backbone of US banking infrastructure, a lot of the payment networks haven't changed drastically in 30 or 40 years. So that that backbone is there and we a lot of times find ourselves needing to bridge and support older use cases for certain things.
You think about the role that checks play. You used to get your paycheck, walk into your bank and deposit it, and now you're probably taking a picture of it on your camera, it's getting converted to a digital file and your bank may have to get that to one or more of their partners.
So there’s situations where there's networks and technology and systems built on older infrastructures, more complex infrastructure, and is not really designed to be extensible, and we have to find ways to open that up and make sure in these more modern use cases we can get data to where it needs to go.
"We have to find ways to open that up and make sure in these more modern use cases we can get data to where it needs to go."
And what approach have you taken in building your technology stack?
We’re not a top five US bank in employee size, but we want to have the technology to compete there. So part of what that means for us is we need tech platforms to help empower us, make us more productive and deliver faster and more reliably.
And that also means being able to reach the edge cases of what we need. So a good example is you may not have a clean REST API that you can hit for a data request all the time. You might have something in a mainframe format file that needs to get to a Linux box or a Windows box. You might have an interaction layer that's built all in XML and it has to get over into JSON.
"The more we can build and reuse and operate efficiently, then that empowers us to focus on projects that drive business value."
The more we can have platforms that let us take that complexity and really work with it proficiently, and not spend all of our time building syntax definitions, restructuring standard payment files - the more we can build and reuse and operate efficiently, then that empowers us to focus on projects that drive business value.We have to be able to do that quickly and at scale and hopefully with less resource than if you're building things totally from the ground up. Which is part of the space where Clover’s filled an important role for us.
Maybe you could help us understand some of the use cases that you are seeing?
One of the technical challenges we tackle quite often is where we have two sets of parallel data – we’ll have a system and a partner will have a system - and we have to ensure that data matches between those systems in an efficient way.
A great example of that is we may have an incoming payment file that is going to need to get processed by both systems and maybe a single side of the system only needs to see a version of that file. So we often we’ll have to dig into an old payment file format, maybe be able to push it to an API, filter it out for what's relevant to that other system, and have confidence that it's going to do that consistently every single time. And then on the counter edge of that we need to pull data back out of that system and make sure it replicated the way we expected.
Could you help us better understand how the technology comes to life in your organization? And what sorts of teams you have implementing these projects, and maybe how CloverDX fits into that?
We’re really sitting with a technology team that's a little bit more lean than maybe a large bank - we have a smaller team of maybe two or three core builders that we look to solve a lot of those problems with. And so the less time we can spend on detail and syntax and things that don’t drive business value, and the more we can operationalize those two or three towards loading transaction sets over to new platforms, the more efficient we can be.
So one of the ways we've done that at Choice, is we have a technology innovation team focused on those different use cases. That's a little bit different than a standard IT team that a bank might have, so we really look to focus on being as productive as possible with those resources that we have. And that involves looking at platforms and processes and support that we can get from teams like CloverDX, and things that can really get us to 100% of the completion.
"We really look to focus on being as productive as possible with those resources that we have.
And that involves looking at platforms and processes and support that we can get from teams like CloverDX... things that can really get us to 100% of the completion."
Because we’ve seen cases before where maybe you will build a process or design an integration and maybe a certain platform can get it part of the way there, but maybe you have to download a file manually first, or maybe you have to convert it from one character set to another. And it's just difficult, but it's important work that lets you automate things end to end.
Those are the types of things we really spend a lot of time and focus on so we don't have to re-solve new problems or have things resurface later on.
I think what I'm most proud of is we don't have a large, complex team of a lot of different technology resources digging deeply into a lot of different issues, we have to mobilize and really prioritize and focus on building the use cases as effectively as we can and spending our time where we know it's important.
I actually have two questions come through from the audience, which I think are quite relevant to this. So one of the questions is what are your areas of technology investments, and which resulted in the highest measurable ROI?
It can be really tempting I think in technology to find an interesting problem and spend a lot of time solving it in a really elegant way. If you work in technology, chances are you find that fun and something that's exciting to solve.
But in terms of ROI one of the interesting or one of the most significant pieces of our tech stack, has been the role that CloverDX plays for us. And the reason it's been that way is we can help empower our teams and developers to spend less time on, let's say, the nuances of syntax. So to be honest when we look at ROI from a Choice perspective, and specifically in my tech innovation role, we’re really oriented towards how can we see new opportunities and have confidence in our ability to execute technically to start building that new revenue.
So this is a really long winded way of saying we don't always know the ROI upfront of some of our investments.
"The better confidence we can have in executing a tech stack, building new things, the more opportunities that we're going to be able to see"
What's been a great story for us is how Clover’s enabled us to jump on new opportunities that maybe we didn't have defined previously. So the better confidence we can have in executing a tech stack, building new things, the more opportunities that we're going to be able to see.
One of the viewers is asking ‘Are we mainly talking about back end systems here? Or are we talking about front end?’. So for those of you who don’t know anything about CloverDX, we’re a data management platform, and we've been working with Jake for quite a few years on providing technology, and also we worked with Jake and his team on that data layer. So mostly I would say back end systems right?
Yes, exactly. And in banking I don't know that most people know when you open a banking app, how your balance and transactions load. Is your app opening and loading it at the time in a real time call? Is it taking a version of data that was snapshot in from the day before and then adding instant pending authorizations to get to your version of your existing balance?
So we're not maybe focused directly on UI/UX, but one of the heavy focuses is timeliness and availability of data when it comes to banking systems. A lot of older banking systems were designed to be really closed in and built on mainframes where you might have business logic built into what was the UI. So we're doing a lot of work around building export files and taking payment files and making them available as quickly as we can, so that information can be used in world-class products and experiences wherever we can.
What sort of infrastructure are we talking? Are you guys in cloud or do you have something private?
We have an on-premise data stack that we work with quite a bit and Clover’s a part of that, and we are also building in the cloud and Clover is a key component of that area as well.
So fundamentally with banking what we find is a heavy leaning towards, and in some cases necessity to, host some technology on-premise and on-site. But we still need to take advantage of the agility of the cloud and the accessibility and scalability of the cloud. So we're spending a lot of time focusing on how we can bridge those two efficiently and effectively to get from those things we’re hosting on site to make things available, maybe in a more extensible and scalable environment.
Maybe help us understand a bit more, even like on the technical side, some of the actual use cases
Most banking core systems are built or have wrappers around some really old infrastructure. We still need to get that data out to modern CRM's, modern data reporting platforms, regulatory bodies, into flat files, and so the places that we can interpret and get to that data is really important.
So we've done a lot of work around getting to those specific payment networks, working with those files, interpreting them in different ways, and maybe not just letting it layer into how a core banking system would typically operate with it.
"Most banking core systems are built or have wrappers around some really old infrastructure. We still need to get that data out to modern CRM's, modern data reporting platforms, regulatory bodies, into flat files.."
Quite often one of the data challenges is the matter of ‘golden’ or single source of truth. How do you kind of gather the data from the different systems and storage and make sure everything is up to date and usable?
Let's say you know where your most real time data is, or maybe where your single source is, and what if that wasn't built to be accessible? Or like, ‘I need to get it to this system, or this platform’ Or ‘I want to push it to this API?’
"We've seen a lot of importance in not just having the single source of truth designed, but being able to reach it reliably and consistenly and effectively."
Or maybe you're trying to make it available through even just a simple flat file and you just can't get to it. That's where we've seen a lot of importance in not just having the single source of truth designed, but being able to reach it reliably and consistently and effectively.
What was the evolution of CloverDX within Choice?
CloverDX had its start in our tech stack in a single very specific instance and we ended up not even using that instance long term.
But Interestingly enough, what it's enabled us to do over time is once we had that bridge built to our older data, we started seeing the possibilities for different ways that we could reuse those pieces that we had built with CloverDX to push it to other new systems.Automate your data workloads with CloverDX
And we started seeing these use cases where we had to solve problems like, well, how can I get this transaction information to this other third party? Or how can I get this customer information into a flat file for this team?
And once we had those connections built we just started building different use cases together. And it started to earn its way over time with reliability and our ability to get to data. It just became the fastest way for us to execute, and the best way for us to get value out of the data that we have at the bank.
So what do you think is maybe the secret sauce that gives you the competitive advantage that you feel you have in the some of the segments?
I think we feel our secret sauce is, or what we hope that it is, is that we're finding some spaces where if you can extend products and services that have been done in a really specific way for a long time, if you can deliver those in new and innovative ways, and actually execute on it quickly and with a smaller team, and find the right platforms and partners to support that and kind of give you the super powers to realize that vision, that's what we kind of hope our secret sauce is, right? I think banking has been widely commoditized over time and it started to become about how you can deliver products and services.
So we're seeing that happen in FinTech when you're at a checkout and you want to spend some cash - you go to the easiest method. You might go to a card, you might go to an application, you might go to your watch - that's taking the same account structures, customer structures, the same data structures, payment networks that were used a long time ago, but it requires more of the bank and the partners involved to be faster with data, more accurate with data and move it more quickly.
So that's what we hope our superpower to be. If we can provide those services faster, more consistently, more accurately, and see that vision and execute on it. Then we've done our job.
What's your feeling about either building in-house or working with someone like CloverDX where we provide the technology and some of the manpower to deliver some of these things?
I think build or buy is an interesting decision that lot of folks encounter, and sometimes you decide without actually realizing what you're deciding. If you were building a bank from scratch today, you would never go into it and be like ‘I have to hire 5 developers to build presentation software and word processing’, right? That's the easy one, you're going to buy Microsoft Office or Google Docs to fill that role.
So then you go a layer deeper and you say, well, what about transaction processing, or something more core to my business? And that's where it can start to get really tempting to want things a very specific or certain way and really try to build that yourself. And what I think you have to be really discerning about is: are you empowering your technical minds to really focus on the things that are adding value for the core business? Are you really hitting your true value function? Or are you working on things that are - maybe if you did it best in the world, would it make a big difference to the bottom line?
"What I think you have to be really discerning about is: are you empowering your technical minds to really focus on the things that are adding value for the core business?"
And that's where we've seen a lot of success with CloverDX as a product and a partner. And there's a couple different ways that that has worked for us. One is the re-usability and extensibility. CloverDX lets us do things that are as powerful as code from the bottom up, but spend less time having to build it and work with the intricacies and nuances. And when we built it, we can reuse it.
And the other is we've worked with CloverDX's professional services teams too to help us build out and just augment and add capacity and expertise where we have some really complicated and difficult challenges. And that's proven its effectiveness over a couple of years for us.
I think at times you work with technology partners that aren’t the right fit for how you're trying to build that. Or maybe don't share that vision. Or maybe they'll say, well, here's this piece of software, but I don't want to get into the messy, ugly stuff you really have to do. I'm just going to do this one really cool part and then kind of let you handle that. But we've been able to kind of jump right into some of the hard stuff and solve it, and that's made a big difference.
"We've been able to kind of jump right into some of the hard stuff and solve it, and that's made a big difference."
One last thing - is there some lesson that you've learned, something that you would have maybe done differently? Or some advice that you might have for someone who might be seeing similar challenges like trying to get to different data from legacy systems and new systems?
I think what we've learned is maybe the way that you think you need to tackle a problem, or the things that you think you might have the most expertise on building yourself, maybe challenge your assumptions about how your technology teams execute and what you build. And kind of look at things from a different angle and in terms of where you're spending your time.
I think when you start to build a technology stack, what happens is you start to have older things that you become an expert about and you maintained over years and years, and next thing you know you might have 80% of your development time doing what it feels like is maintenance, and the amount of time you can actually invest on new things that help you build into the future gets smaller and smaller.
And I think that's where you have to look at platforms and teams and products like CloverDX - to look at where you can get more efficient with that and reduce the amount of time you have to spend on maintaining, working, and doing things that don't add business value.
And if you can say ‘We have this set of data we feel really comfortable about, our bridge between that data and the use cases that we need. Now let's focus all this time on building something new and thinking about what can actually add value.’
Sometimes you can feel it in your gut when people start coming to you with problems and opportunities and you don't feel like you can execute on it or solve it adequately or effectively. That's a great tell for ‘We might need a partner in this, or there's probably something that can help us build through this or get this done in a new way.’