When thinking about flexibility and scalability, each individual will have a different lens on what that means to them, their team, their business. That lens is informed by their role in their organization.
Today, we talk with ADAM SMITH, head of originations technologies, and CHRIS BRIGHT, director of application architecture. The two have been with defi SOLUTIONS for just under and just over six years respectively. They have been integral in the success of defi, most specifically in terms of product development. Chris has worked extensively on two of defi’s loan or lease originations products — defi LOS and the new defi ORIGINATIONS being rolled out in 2022 — and Adams oversees product development for multiple defi products including – defi ORIGINATIONS, defi LOS, and defi XLOS – as well as defi ANALYTICS and defi’s digital interface, defi CONNECT.
Their responses were edited for clarity and brevity.
When you think about flexibility, what do you think of? When you think about scalability, how does that differ?
ADAM: So let’s talk about scalability first because I think that’s probably the easiest for us to define. Scalability from a technical standpoint means that I’m able to provide a system that lets clients bring on as much business as they want. The system can grow and meet demand without me or anybody in our operations team having to do anything.
So if a client is pushing through …say… ten applications an hour, it just hums along, and as they ramp it up to 10,000 applications an hour, our system and our architecture, our deployment, and everything that goes along to make our product “our product,” can recognize more load is being put on the system. A scalable system can grow at a very controlled rate in a controlled manner so that our client doesn’t feel any issues when they start putting more and more load onto the system.
Do our originations systems take full advantage of the technologies that allow us to be the most scalable?
ADAM: The defi LOS and defi XLOS both have it now in varying degrees. They were both designed and built for a different type of architecture. We didn’t have the tools and technologies available then that we do now and neither one of them are fully utilizing the architecture that’s available to realize the complete value of being in the cloud.
It is a goal of ours to be truly cloud native and able to scale at will. It is part of the new defi ORIGINATIONS.
Compare scalability with flexibility and what that means to you.
CHRIS: Flexibility is providing the tools necessary for clients to implement their business objectives with the least amount of friction.
ADAM: It’s inherently built into our products. What I mean by that is they’re all configurable and while they have a variety of base capabilities and features and functionality, clients are not locked into a single workflow or a single way of doing business. We offer them flexibility to configure the system to work for them and to work for their business.
So case in point: configurable credit flows for sub-prime versus prime versus super prime clients. The systems are flexible and can be configured for finance companies, large banks, or small or large captive lenders, using the same tool but configured to their unique needs. They’re flexible enough to have a multitude of systems to integrate into. And to allow our clients to configure those integration points as needed or integrate to various systems or vendors or third-party data providers as needed, in order to meet their business requirements.
CHRIS: I would describe flexibility and scalability as two opposing forces in some ways. If you have infinite flexibility, it is very challenging to anticipate the scalability needs of the system without sophisticated processes and tooling in place to understand how our clients might use the tools in a way that we didn’t necessarily anticipate.
I think something can be configurable but narrowly defined in such a way that it’s not very flexible. For example: you could have a piece of configuration that might have predefined values for A or B or C. But what happens when you need D?
Flexibility to me is being able to modify the system in such a way where you can achieve D without necessarily needing coding changes. Or, maybe, allowing a client to implement that D scenario but if they try to go beyond D they have to come and talk to us so we can understand what they’re trying to achieve. That doesn’t necessarily mean a long-drawn-out SOW process.
It’s just more about us being the experts in the system, knowing what they’re trying to achieve and whether or not it’s conducive to the stability and scalability of the system.
What has limited our ability to be as flexible and scalable as we (and our clients) have wanted to be?
CHRIS: I would say it is maturity and processes for managing the operations of the system. (It’s the shift of focus of investment into evolving our core system beyond its initial success.)
ADAM: Our two originations systems were made for completely different clients and completely different client bases. defi LOS clients were typically smaller finance companies. They wanted flexibility built into the system. They wanted to be able to change on a dime. The defi LOS has always valued agility first. Getting features into production and spending the time to get production right was much more valuable to the LOS and to the client at the beginning then was a stable environment that was very well tested and that would be able to handle work at volume.
defi XLOS was built initially for, arguably, the number one auto manufacturer in the world and the number one auto finance company in the United States. It is flexible, but it was built more for stability. It’s like a big diesel engine. It’s not as sexy but once you get it running it is like a freight train. It’s just going to keep going. There’s a lot of stability built into it.
CHRIS: And in some cases the data requirements differed quite a bit. On XLOS, the bread and butter are captives, prime and super prime applicants, people that are generally not going to get turned down for a loan. It’s more about “Can I offer the best deal to be chosen by that applicant?”. Whereas on LOS the initial clientele or target market was subprime, which requires a lot more risk analysis and a lot more data points to conclude whether or not one of our clients would want to fund the loan.
The LOS system had to be more flexible to accommodate the additional data and for that data to be piped through the systems, whether it was being necessarily used or not, and that impacts scalability. The XLOS has a different approach that tries to anticipate all of the data points that a client may need and provide those out of the box.
Is there a different fundamental architecture that could be implemented to make this more efficient?
CHRIS: The architectural style nowadays is more “microservices-like” where you can scale different pieces of a system more independently. It’s easier to observe those pieces to understand where the performance bottlenecks may be, to see what’s going on at a more granular level and have a better understanding of when a problem might be coming up. Because parts of the system are isolated, we can measure their usage utilization at a much more granular level.
ADAM: The idea is to take these large applications, break them down into their core components, develop them into self-contained services that are not tied to a database and that can be called and scaled independently of one another. So, using an originations system example, as the credit portion of the system is getting hammered with applications, all the credit stuff can expand and scale as needed, but the funding part of the system that isn’t getting hit as much can stay exactly where it needs to be.
Why is now the time for defi to go for flexibility and scalability in a single originations product?
ADAM: Neither originations platform in its existing state would be able to meet future demands without serious investment. Over the past year we have doubled the expense and doubled the size of just about every defi LOS server it runs on in production. We doubled the SQL capacity. We set a 2021 road map that was frontloaded with feature development, stability, and performance improvements and we’re seeing the dividend of that. But if LOS was going to be our go forward platform and maintain our market leading position into the future, we would have additional investments to make. The same is true on XLOS.
BRIGHT: So, we’re taking the experience of building the flexible LOS, the flexible XLOS workflow, the operations maturity of the XLOS, and its ability to efficiently process massive amounts of data and bringing them all together. Also, the new ORIGINATIONS platform will be moving to more of a cloud-native hosting and management style.
ADAM: And, now is the time as our clients are changing as well. Speed and stability remain paramount to them. They still have to turn around that initial decision as fast as possible because the first one back to the finance desk, or the first one back to the client shopping from home on their iPad, is usually the one that’s going to get the deal.
But they are also looking to us for innovation and thought leadership how and what technologies to implement to solve their business problems. And being cloud-native gives us the availability of tools such as true near-real time messaging; machine learning, AI, and big data; and all of those tools that are the big buzzwords right now.
We can take the lessons learned from both products and really develop out a best-in-class origination product.
What we’re doing now sets the groundwork for other products and services?
ADAM: Everything we’re doing now is setting up the originations product for the long haul. It is taking full advantage of tools available to us in the cloud. It is putting forth design patterns that allow us to make changes much more strategically, tactically. It’s allowing us to develop that system that scales automatically and it’s providing capabilities that each product had individually that differentiated it, now together in one product.
A configurable UI and Java scripting. Fully configurable, fully customizable credit flow options. A fully open API layer, a modern API layer so that clients can build onto our system and build those capabilities outside of if they want. Maybe low code/no code eventually sold along with this. And then you look at some of the other stuff with pricing capabilities, with the defi STRUCTURE product that’s coming online here shortly. All of those capabilities will be much more mature, much more enterprise level, much more scalable, much more performant and let us provide new capabilities to our clients.
Nothing is “flip the switch perfect.” Why should a client stick with defi through this change?
ADAM: It really comes down to a system that is incredibly flexible, incredibly stable, incredibly performant, with features such as Auto Structure where they can adjust their credit bureau information on the fly. And with an open API layer so they can do development into the system on their own.
We are working hard to make sure there is no loss of functionality, feature, or systems capability. We can’t promise that things will work exactly the same way, but we’re working to help ensure the same outcome.
We understand that our defi LOS and defi XLOS originations systems are in use by our clients day in and day out, eight or more hours a day. In some cases, they’ve been using them for years and are very much used to doing things a certain way. There could be resistance to trying something new.
But we’ve learned a lot over the years, from trial and error, from our clients, and are in the position of being able to continue to learn more and share what we’ve learned through our products and services.
CHRIS: When we pull this off, it’s going to be exciting. I really truly believe that this product will be the best of both worlds and best in class therein enhancing our already market leading position.
