Make Spreading Financials Easier with Technology

Mitch Woods: Welcome to Baker Hills podcast. Lending Made Easy. The show where we demystify the world of commercial lending and bring you up to speed on everything you need to know to make lending easy. I'm your host Mitch, and I'm here to help you navigate the complex world of lending and finance. Whether you're a seasoned banking professional or new to the industry, this podcast will provide you with valuable insights and information that will help you succeed.

So join us as we explore the exciting world of banking. Grab your coffee, close your spreadsheets, and let's dive into today's episode of Lending Made Easy.

Welcome to today's episode of Lending Made Easy. Today's episode, we're gonna really dig into how to make the loan origination process easier with statement spreading technology. Key idea is process efficiency. So to kick things off, kick this over to to Bryan Peckinpaugh here. How does statement spreading technology improve the overall loan origination process for a bank or a credit union?

Bryan Peckinpaugh: Yeah. So I think there, there's a few layers to this, right, Mitch, where, you know, we, we tend to take for granted, especially here at Baker Hill, where we, we effectively created this genre, right? If you go back to our founding, uh, we were the first company to bring digital spreading to the market. We, we take for granted that a statement spreading solution with manual data input

can be a lift for a lot of organizations and, and can drive a lot of efficiency through standardization of that data input and the analysis of that data. So I think far too often when, when we start talking about technology in the world of financial analysis, people want to immediately go to automating financial analysis.

They want to talk about OCR, they wanna talk about these other ideas, and those are great and they have a place, right? And don't get me wrong, but we, we tend to skip, ask some of the critical building blocks and, and those are building your view of solid financial data and, and how you want to structure that to, to be able to run your analysis, to be able to do that over time.

Uh, right. So a lot of that standardization plays directly to how is this organization performing quarter over quarter, year over year. Being able to have those standardizations allows me to look at that, right? If, if everything's all over the place, if I'm allowing free form, Excel data entry, it becomes very hard to see what's changing because people may put it into different buckets, right?

So having that standardization is key. Now, where do I see some challenges and struggle with technology? I think, and especially in today's day and age, people are jumping far too quickly to automation, and we are losing the institutional knowledge. We are losing the value of the people that are doing the analysis.

I'll use a very simple example around tax returns, right? So tax returns are a great example of an easy place to automate. Pretty much these days, they're all computer generated and printed, so they're very nice clean documents for an OCR to read. They are highly structured. I, in fact, you can even, uh, train those OCRs and update them on a yearly basis.

As the tax codes change, you can get very, very good read results on tax returns. And if I'm talking about a simplistic individual W2 and somebody who has a singular job, They don't have any outside hustle where they're bringing in money from, uh, driving an Uber or delivering food for DoorDash or you know, running an Airbnb.

If I just have a singular W2 job and I automate that with an OCR type technology and I plug that into the standardization of my spreading template, that's gonna work really, really well cuz there's not any decisions to make. All I'm looking at is very basic income and ex, you know, kinda standard expenses that people might have.

If I start to think about an individual, so let's, let's say David here owns 15 rental properties and is effectively operating as a business, but is filing, you know, what amounts to an individual tax return with a bunch of rental property income stated, again, very easy to automate. Very, very simple. OCR can do that all day long and put it into my templates.

I still gotta have to do the analysis. I, I don't have the benefit of a system telling me that that's a good structure or not. You know, that that set of properties that David has, maybe two of those properties make up 80% of the income that David brings in from his rental properties. Well, now what? You know, do I, do I really want to use that a as part of making a loan to David to go buy a 16th one?

Am I gonna get the third performing property or am I gonna get the 14th under performing property . Uh, and, and that, that matters, right? And, and the technology's not gonna do that for you. And I think far too often I hear automation of financial analysis, which I think we need to break. That's, that's not what we should be , you know, solving for on the broad scale.

Yet it, it's a great aspiration. I think there's a lot of great technologies and capabilities out there that we can strive to get to, but we, we shouldn't look to replace that real solid human intelligence in the process or sake of automation for sake of, you know, speed and things like that. So I think there's a lot of great here, right?

And a lot of great strides that can be made through standardization and using technology for that. I, I do caution people for, for thinking that, you know, the AI based financial analysis is all of a sudden gonna solve your problems and you don't have to do the hard work of understanding the data that you're getting as, as part of financial analysis.

David Catalano: Yeah, so I think what you're saying is the critical thinking that can be delivered by a human analyst is really what's required for the deals you're talking about. Because we had a segment of this show specifically talking about auto decisioning and scored loans, and the example of a loan you just provided is not, is not appropriate for a scored auto decision.

It's definitely requiring some financial analysis. And when you have a person working on that and seeing the 25 properties or however many 16 pro 15 properties that you were talking about, understanding, the two of them make up 99% of the profit and we're gonna buy another one that that's critically important to, to be able to see that, understand that, and it's just a better application of a credit analyst time in a deal like that

because it is a complex financial situation, even though those, uh, financials are not complex. So it does require the critical thinking of an analyst. So I think what you're saying is, is just be careful and not over automate something that really requires or, or really is not ready for, for that level of automation.

Bryan Peckinpaugh: It, it get gets back to David right, what you and I talk about quite a bit, which is you shouldn't buy technology to buy technology. You, you shouldn't get enamored with feature function and bells and whistles and the shiny new object. You need to go back to just good business practice, which is define the problem you're looking to solve and figure out the best way to solve the problem, and in very few instances is the problem truly.

I need to spread faster, right. I , right? I don't know if you've ever spread financials, David. It's been a long time since I have, but I did back in the day. You can get really good with a 10 key , especially with consistency of what you're being given to spread and, and I think it, it boils down to, you know, as you look at your process defining what you're trying to solve in it, is it a data accuracy problem is it a pure cost reduction play well, but you better figure out where they're doing the intelligence tasks and solved for them somewhere else in your process so that doesn't get lost. A and also thinking through the unexpected consequences and benefits of the decisions that you make and, and saying there's a lot of different ways to go about solving these ideas

and each of 'em are gonna have pros and cons and which one best fits. You know, I think there's some really interesting applications of solutions in the market. A partner of ours, Validis as an example, and what they do in integrating to financial accounting packages that can help with automation of financial analysis.

But I think it also opens up bigger doors if you think about what you can do with the technology they bring to the table that's going to be different than the pros and cons you would get example, another partner of ours, Flash Spreads and their work with OCR of tax returns. Again, you can look at both of them and say they are, quote unquote, automating financial analysis.

They're doing 'em in a wildly different way and, and they are gonna bring you ancillary benefits that are different from each other not to mention, we've got even a third partner Finagraph that has, again, a slightly different version of how they solve or for the same problem. They're all gonna be an appropriate fit for an institution depending on what business problem they're truly trying to solve and what other aspects of their process they care about.

Right? And, and what those solutions will, will bring to bear. Is it, do I care a lot about experience and what it's like for the borrower versus. , do I care more about the rich data that I can get over and above what might go into a spread? Lots of different things to weigh out as you look at what might be out there to, to help solve this, you know, maybe immediate problem of automating the input.

Mitch Woods: And Brian, I think something else to bring up here that it's important, especially when we start talking about process improvement, especially around load loan origination processes. It's not always just about speed I think that's important, but I think to your point, it's that speed to decision and not just the decision of the loan, but the speed to be able to make a decision within my process to move to the next piece of the puzzle, right?

To move to that next part of my workflow. Uh, and I think to your point, we think about automation in that, in that sense, but I think there's some other, other things to consider just from modernizing that statement, spreading experie. to standardizing it. Right. So to your point, a credit analyst can do the job of analyzing credit and not just data entry.

So talk a little bit more about that. Like, when you start to talk about automation and standardization of this data, what does that really do for, for a financial institution and, and helping them make better decisions?

Bryan Peckinpaugh: Yeah, I think the standardization is key, right? Because now if I've got a singular individual credit shop, right? I have one person that looks at all deals and they've been with me for 25 years. I don't need to worry about standardization and consistency. They are gonna take care of that on their own. Uh, if I have a larger organization and I, maybe I have 2, 3, 5 people in the financial, uh, analysis role, the underwriting role, those roles may turn over.

Consistency becomes wildly important right? If I get a deal in and it goes to David versus going to you, Mitch, I want to see those analyzed in the same way. I don't want the nuances of your, your individual personalities impacting the credit decision. Right? And to do that, I need to make sure that the data is making it into the same buckets every single time that the ratios that come.

Are representative of exactly what I wanna see and care about, that the resulting cash flows fit with how I think about your ability to repay loans and things like that. And, and I need the system to help me drive towards that, you know, help take some of that human inconsistency out of the process. So that making the decisions on not just the same data points, but consistent data points.

And, and what I mean by that is without a tool, and with more freeform data entry, what rolls up into a specific ratio or a calculated data point may be different if you put that into different buckets, right? So, you know, if I'm breaking out certain assets in different classes, they may roll up to a number and David enters it in differently and I get a smaller number.

I need that to be consistent. I need it to always be the same data entry based on what I wanna make my decisions on. So you, you really have that focus of, you know, driving consistent data so that I make consistent decisions so that I can adhere to my credit policies so that I'm making the right decision every single time. Both to say yes and to to say no.

David Catalano: Yeah. And in some instances, you're just trying to pre-flight a deal or determine if the deal has the ability to get approved. And if you can quickly get into the solution and spread. And if you're a relationship manager and you have that capacity and you can, you know, understand the spreading and the tool set up the way in which, uh, the bank wants it spread, that's a fairly, uh, efficient way of determining, just, at least from a cash flow perspective. Does, is this steel gonna work?

Because if you think about it, most of them are cross-sell, right? We're selling another loan to an existing customer. You have those spreads in the solution already. Adding an additional period or additional spread is not going to be that that challenging if they're all set up and you can see trends, you can see it, whether or not it's the similar to the last spread that was, that was completed, are there any material changes there and get a sense of if this deal has legs?

Cuz uh, you don't want a good customer getting along. No. Uh, you wanna understand what, how do I get a yes? What am I gonna do to get a yes for this existing customer? Mine.

Bryan Peckinpaugh: But you're, but you're also right David. A fast No is infinitely better than a slow No.

David Catalano: Oh my God. You're right.

Bryan Peckinpaugh:If I could tell you no and why right away, that's, that's valuable to, to those existing relationships.

You know, we also, you, you have to think too Mitch about different types of lending. You have different drivers of consistency and being able to trend over time to be able to project having that trending that David was talking about, having period over period analysis, you can't do it if it's not the same buckets, right?

If they, if, uh, different components of the spread are moving all over the place, I can't, I can't analyze income period over a period. I can't analyze debt period over period unless it's all rolling up to the same, you know, same chart of accounts. So it, it, it is vital to, to the health of my credit department to be able to see that make that analysis cuz as David mentioned, a lot of the work that's done is with not just existing clients but existing books of business, especially in the commercial world, right? I'm coming back and I'm doing annual reviews, uh, or more frequent, depending on the performance of the loan. And I need to be able to see that how is this trending, how is their ability to repay changing period over period?

And that's all due to the consistent input of the data that, that allows me to see how that that is trending and then be able to do things like shocking the portfolio to say or shocking the individual loan to say, you know, especially in today's, uh, economy, what happens if all of a sudden the world shuts down for two, two weeks and there is no foot traffic through the door of a retail store? Like, what does that do to, you know, their ability to, to make money and repay the loan? How much cash on hand do they have, et cetera.

Mitch Woods: Yeah, no, that's, that's great. And I think something that I heard over and over again is the word consistency, and I think that is critical, especially when we start talking about spreading financials and really making that loan origination process easier for everyone involved.

So I think some great takeaways from, from both you, Brian and David today, and really understanding what's that incremental lift that implementing some type of statement spreading technology's gonna have on that loan origination process for my bank or credit union, because it's not about going from Excel to full automation it's taking that first step and what lift is that gonna give me? How is that gonna create more consistent processes and standardization in my data? And then how is that gonna help me make better decisions? So I think some great takeaways for, for anyone out there listening. So thanks everybody out there for tuning in to today's episode of Lending Made Easy.