Many CFOs preach about keeping up with technological change. But CFO Glenn Hopper went a step further. As ChatGPT talk reached fever pitch he was ready with a proof of concept for an FP&A version of Chat GPT. For good measure he also provided 30 pages of documentation about the process (“Application of ChatGPT to Build an FP&A Tool”) – to acclaim among FP&A professionals.
Hopper, a serial startup CFO, and U.S. Navy veteran who earned a master’s degree at Harvard says: “I wanted to see if someone with just a base understanding of the underlying technology could use artificial intelligence to first off build a tool that would help automate some of the FP&A processes and secondly, get to a way where you could use that tool to interact with your company’s financial data.”
The Chat GPT project was a perfect technical project for Hopper, playing to his technological mindset and skills foreshadowing technological transformation for finance professionals as explored in his book Deep Finance In the Information Age.
In this episode Glenn Hopper discusses
- His unusual transition to CFO, moving from journalism in the Navy to product manager.
- How a passion for writing helps to formulate projects and challenges in finance.
- His experience with financial transformation in finance companies and the lessons he learned from the experience – as set out in his book Deep Finance.
- How having a mathematical and analytical mindset is crucial for anyone in FP&A.
- How his biggest corporate failure (missing a $1.5million invoice in budget preparation) was essentially his “origin story” to becoming a transformative CFO.
- How he wrote a feature film – a low budget horror movie, The Hanged Man which streamed on Netflix.
- His favorite Excel function.
Paul Barnhurst
Hello everyone. Welcome to FP&A Today I am your host, Paul Barnhurst, aka the FP&A Guy, and you are listening to FP&A Today. FP&A Today is brought to you by Datarails, the financial planning and analysis platform for Excel users. Every week we welcome a leader from the world of financial planning and analysis and discuss some of the biggest stories and challenges in the world of FP&A. We’ll be providing you with actual advice about financial planning and analysis. This is going to be your go-to resource for everything FP&A. I’m thrilled to welcome today’s guest on the show, Glenn Hopper. Glenn, welcome to the show.
Glenn Hopper:
Hey Paul, thanks for having me.
Paul Barnhurst:
Yeah, super excited to have you. So let me just give a little bit of Glenn’s background and then I’ll give him an opportunity to tell us a little bit more about himself. So he comes to us from the Memphis area. He has a master’s in Finance and business analytics from Harvard. He has worked for a number of different companies, primarily in finance roles and has served as the CFO for several different companies. And he recently wrote a book called Deep Finance, corporate Finance and the Information Age. So Glenn, can you tell us a little bit more about your background and how you kind of came to be a CFO and write a book about finance?
Glenn Hopper:
Yeah, sure. So I used to apologize about this that I didn’t come up. I didn’t come to the CFO role in the traditional CPA public accounting audit, the traditional CFO path. And the more I’ve been talking about it, the more I’ve been talking to others and especially an FP&A audience, I feel like I don’t have to apologize for anything here because I came up through initially, I mean if we went way back in the stone age before I got into finance, I was a journalist in the Navy which that’s the weirdest transition going from navy journalism to finance. But when I got out of the Navy, my first corporate role was I was a product manager. This is in the late nineties, early two thousands. I was a product manager for a tool that built it was think of WordPress but this was back in the late nineties.
So very few out there. So I started in marketing and my product I always felt like didn’t have enough budget dollars and with my freshly minted MBA I decided <laugh> that I was going to really push for, let me look at the budget, let’s see if we can find some place for it. So long story short on that, I ended up becoming the budget guy for the sales and marketing group. And then after a few rounds and meetings, the COO of the company said, I want that guy. And he brought me out of marketing into operations. So it was kind of a rev ops role. I was the budget guy for the chief operating officer in the company. And from there I went in really cut my teeth on FP&A and early, this is early 2000 still business analytics. And so to me I had the company’s first business intelligence group because we were tracking all the metrics, not just the financial.
And so to me, the analytics and finance function have been joined at the hip for the entirety of my career. And that was, I was in telecom back then and ended up moving to much smaller companies, but I got a bigger title and a lot more responsibility. And that was in maybe about 2007, I took my first CFO role and I’ve spent the past 15, 16 years going from one startup to another as A CFO. And the timing when I come in would be right around the A round or when somebody, if they’ve bootstrapped, if they’re looking to raise money or to be acquired or whatever and they need to go that startup to scale up phase. And that’s when I come in to help them professionalize their back office and finance operations.
Paul Barnhurst:
Great. And I appreciate that. And funny enough, when you mentioned Navy, that’s where I started my career. I was civilian, but I worked for the Navy on a Navy base for four years, if you’ve heard China Lake. So I spent four, I’ve spent four years out of China Lake. And another kind of interesting note, you mentioned not many people saved finance from background journalism, the navy, the guy I have on next week, which is a fractional CFO, my next interview, he started his career with the Navy. He was a power plant, nuclear power plant supervisor and now he’s a practical CFO. So
So when you said that, it just kind of stuck out to me of wow, there’s some coincidence there. But that’s definitely a different background. I really like how you talked about how business and analytics, it’s kind of always been together in finance, the finance and analytics side, similar for me. And then I have a master of science information management. I started in a report writing and putting in a BI dashboard type role and then switched to FP&A. So I’ve always been closely tied to the analytics side of the process, so I can appreciate that. So maybe next, can you tell us why did you decide to write a book? I think it’s corporate finance and the information age, so how’d that come about?
Glenn Hopper:
So the desire to write I think comes from the journalism background and it’s sort of writing is the way that I sort of work through a problem. You can go through and try to come in to a place where you can solve an answer mathematically, but just to, while I’m thinking something, my approach to it is to write. And once a journalist, always a journalist maybe. I am compelled to keep writing blog posts. And I’m a contributor for several different websites and it’s one, I’m kind of an evangelist for what I do with the automation and data collection and using machine learning and algorithms and really combining data science and FP&A together for modeling and for all the analysis that we do and everything. And so for several years I’d been writing all these blog posts and articles for various websites about using data science and finance.
And I looked back and I thought, this is a lot of content here and it wouldn’t take much to make it a cohesive single unit of work that would capture everything that I’ve been talking about. And so the mission of the book is for people, whether they’re new to finance or have been doing it a while, but I was picturing someone who’d been doing finance and accounting in a corporate setting for a long time and very good at that, very much a professional in that area, but we get stuck in the way that we do things and we’re pretty good. All of us are very good at Excel and we know how to do crazy stuff in Excel and it’s thinking about trying to do things a different way, can be intimidating and trying to keep up with the latest accounting changes and what’s going on, everything you have to measure there. And then, oh by the way, learn this new technology, it sounds daunting and why would I do that? What I’ve been doing work. So the mission of the book was twofold. One, to make this new technology more accessible to break it down and just explain it at a base level and two, to show the power of it. And well, I guess the third part of it is I’ve got a big section of the book that is how to basically lead a digital transformation from the finance department for an entire company.
Paul Barnhurst:
So sounds like there’s a few things in there, what it is, why you should do it, kind of how to think about it and then how to lead a transformation.
Glenn Hopper:
Yes.
Paul Barnhurst:
So maybe on that part on the transformation, because most we’ve been hearing digital transformations and finance for a decade now, we hear a lot of them fail. What’s your, what advice do you offer someone in leading a transformation? How do you think about that?
Glenn Hopper:
So first off, I just said digital transformation, but I’ve been kind of called on the carpet for that for the exact reason that you said people are tired of hearing digital transformation. And it goes back, it is more than 10 years because it goes back to think about the first time that accounting moved to the cloud. So everything is a digital transformation and transformation suggests you, it’s a one and done. So what I keep meaning to train myself to do is not say digital transformation anymore, to call it digital evolution in that it never stops that we are constantly trying to evolve the company and technology and the advancement of software and hardware that’s out there and the capabilities are not going to change. So people get mad when they think, okay, we’re going to do this and then we’re done. It’s like, well that’s one step and then we’re going to have another step.
And then as fast as technology’s moving now we’re going to have to do it again. So yeah, digital evolution is the new way I’m thinking about it. And the reason for it is if we look back to when I first started my FP&A career everything was in Excel. Maybe I think maybe we brought in an access database early on and we felt super tech focused then because we weren’t just doing Excel. We had a database. But the database actually, that was my first financial transformation, is moving beyond Excel and getting sort of direct access and manipulating data in a database instead of just in spreadsheets and in the technology that’s come on since then. I can’t think of a single function in accounting and finance and on the corporate side that isn’t, if not fully automated, made much simpler by just off the shelf software out there right now. I mean everything from expense management to bank reconciliations to software packages out there now that the continuous close. And so if you’re not keeping up with the technology, I mean you’re going to get left in the dust.
Paul Barnhurst:
I agree with you, as I heard someone say, and I really like the way they phrased it is AI isn’t going to replace finance. There’s all this talk about ChatGPT , which we’ll get to here in a minute. But the way they put it is finance people aren’t going to be replaced by AI. They’re going to be replaced by another finance person that’s taking advantage of AI and goes beyond just . AI sometimes gets used as a term for everything and technology and it’s like, no, there’s some lines, but it’s so true. The technology’s getting much better. Companies that are not being cutting edge and continuing to evolve, not transform because like you said, it’s an evolutionary process, it’s constant. There’s always something to improve and ways to bring in technology are going to get left behind, in my opinion. I mean I think we’re seeing it rapidly start to change. And speaking to that ChatGPT came out here a few months ago and when I reached out to you to the podcast, you were working on a project there around chat GPT that you recently released. So could you maybe talk a little bit about that, what it was you were doing, a little bit about that project?
Glenn Hopper:
Sure. And the project was really meant to be a proof of concept because as I evangelize about what we can do with the technology that’s out there, it would be a true unicorn to find someone who knew everything you need to about finance and accounting to work in or run a finance department and could also <laugh> architect and develop software. So asking someone to have both of those skills that’s asking a lot. But I think that what’s happening is the technology that’s out there is making it easier for non, I don’t want to say completely non-technical because I think we do have an obligation whether you’re in finance, accounting, sales and marketing, whatever department or career path you’re in, you’ve got to keep up with the technology. But the reason for the chat GPT project was I wanted to see if someone with just a base understanding of the underlying technology could use artificial intelligence to first off build a tool that would help automate some of the FP&A processes and secondly, get to a way where you could use that tool to interact with your company’s financial data.
So what I did with the project was I created a fictional company and had three years of financial statements just in Excel spreadsheets. And my idea with this was I don’t want to on my own write a single line of code. Or even figure out where I’m going to put this or how I’m going to do it. So basically if someone isn’t going to write code, how are you going to get these financial statements moved into a database where you can run some automation on them? So opened up Chat GPT basically outlined with a prompt the project I wanted to do and I, well, I did tell it because I’m thinking about the architecture, I said I want to do it in Google CoLab and use SQLite databases. And the reason I picked those were they’re both free and for a SQLite database you can even connect to your Microsoft not Microsoft, your Google Drive and just put CSVs in there and or put your database in there.
So I gave it to parameters and said, now what do I do to get this information out of the CSVs and into a database? And through a series of prompts, you see I set up the tables, import the data and then run some basic FP&A on ’em and do some kind of cool things. And it’s really, it’s not meant no one could take the code that Chat GPT and I wrote and run that in production, but it’s in an amazingly few steps where I wrote no code where chat GPT wrote all of it you can see the future of where this is going and there’s a little, right now I compare it to a AOL in the dial-up era, you’re kind of it, it’s watching the sausage be made. I can remember back in the day, the screeching modem while you’re waiting to connect and it’s a little of that going on.
This is not for everyone, but it’s moving so quickly. I think there’s only a matter of time before you’re not seeing how the sausage is made. You’re giving the prompts and it’s doing it all in the background and you’re just seeing the finished product. But if you want to be able to capitalize on it, then this was that first shot out there for me to show AI can do all this stuff for you. And it ended up being a fascinating project and probably, I don’t know if I’ll do another paper on it, but I’ll probably keep just pushing this and seeing how far I can take it with chat GPT writing code and what kind of stuff we can come up with.
Paul Barnhurst:
Yeah, I enjoyed reading it. I, Glenn sent me the article before that came out and I appreciate that as you’re getting ready to post it. And I’ve read through the whole thing this week in preparation for us chatting and it was really impressive. I know it was Python, a lot of what it used to write that code, it was the language that chat GPT, but I was really impressed. A couple things in there I saw is I know you also had to do some financial ratios and then at the end you asked, you created a simple bot and asked some natural language questions. So why was that kind of the path you chose of what you wanted, the tasks you wanted Chat GPT to do the bot in that?
Glenn Hopper:
Yeah great question there. So <laugh>, the first off, when I started the project, I was envisioning an API with chat GPT, where I would just put all the data into these tables and then have chat GPT go direct to the tables and answer questions. Now there may be a pro version or whatever, but on the publicly available version, you couldn’t do that. You couldn’t just have Chat GPT, go look at your databases. But the whole point of the project for me was to have this automated system that could basically be, I don’t, maybe like a junior analyst or whatever. So if the CEO or CRO or anybody else in the company has a question rather than reaching out to FP&A asking for a report or whatever, how cool would it be? And I actually did a project like this in my book too using Lex, the Amazon language, but it, it’s not as I had to actually in that load the questions for it to ask so it would know what to answer.
And I wrote the book two years ago. So in two years the technology that’s out there has gone from, okay, I’m making a chatbot using the Amazon Lex service to now on the doorstep of just having chat G P T do this. But the whole idea is all these about depending on the size of the company and the resources that are out there when people want reports, depending on how backlogged the report group is and how much data is democratized, you could have to wait a day a week or you’re stuck in the backlog for some complex report, wait a really long time to get data that’s needed for a business to make these data driven decisions. So my proof of concept there was can we find a way to automate some of these easier reports? And it’s really, it’s about data democratization and how if we’ve got a system that can do this, we’re not taking any human resources and can answer these questions. To me, the next wave of finance automation,
Paul Barnhurst:
That makes a lot of sense to me. What you said there about kind of the next wave of finance automation and wanting to demonstrate that you could create a bot where people could go ask questions, use natural language and get responses versus having to wait from a report department or a more senior finance analyst or whoever it may be. So question there, how long do you think we are away from having this in most our tools, getting to that point where you almost have, as you put it, the junior analyst.
Glenn Hopper:
So honestly I think the bones and this project kind of helped prove that in practice for me, not just theory, but I have been saying, and I’ve thought about this at least the past couple of years, someone in, because I’ve been in the small medium enterprise space under $50 million a year in revenue. And in that space there’s not a lot of you don’t have as much data as the big companies do. You don’t have as many resources as the big companies do. And I think that we are on the verge of something like when QuickBooks came out, I mean IT revolution and QuickBooks Desktop and then QuickBooks online, it revolutionized accounting for small businesses. Now if you’re at one of these large companies right now that has access to an incredible amount of data and teams that can work with it, you can do some amazing things with FP&A and looking at whether you’re forecasting or trying to explain budget variances or whatever, you just have so much more data to work with.
I think someone is going to come along and have a product that is basically the QuickBooks of machine learning that brings this capability to the small businesses that otherwise couldn’t have it. And I think that there’s probably a way that once you had enough customers you could actually, whoever came up with this software could aggregate data and they could anonymize it and all that, but be able to use a broader dataset that any single small business would have. And I really think this is going to be a next wave. And so is, to answer your question, I think the base technology is there right now and there’s so many people doing so many things with this the large language models and all the advances that are going on right now, that’s just on the chat side, but just in, the tools out there to do drag and drop ML (machine learning) and if you combine all these it’s just takes someone that’s going to have the focus on doing it for FP&A and it’s, as soon as someone finds that, I think there’s going to be an off the shelf product that offers this today.
And I’ve seen, and I’m not going to come and shill for any particular software, but there are components of this in SaaS products that are out there right now and it’s, it’s only going to, it’s an arms race right now. So there’s between existing products and some startup that’s really going to nail it. I mean as soon as someone gets to focus on this, we’re going to start seeing this on the market.
Paul Barnhurst:
No, I totally agree. And I’ve even seen a few very small companies that have implemented Chat GPT and some of the things they’re doing. So I a hundred percent agree with you we’re seeing it and they’re going to figure out how to package that and it’s going to change, it’s going to be in many ways a game changer around many of those basic tasks that for companies that, especially when your day, if your data’s fairly structured and clean, I mean it’s amazing what the answers you can get out of it. And I think it’s an extra incentive to keep that data cleaned. Exactly right. Because you get really messy data and think, and you don’t have good master data and it gets really hard regardless of whether it’s AI or a human trying to make good sense of the data.
Glenn Hopper:
Yeah. And 80% of data science is just cleaning the data. So
Paul Barnhurst:
60% of FP&A is cleaning data. If you’re an analyst, you spend a lot of time cleaning data. I mean, I worked at a job where I had to do almost all kinds of things in Excel and it was just doing something like this was, as I was looking at your project, I’m like, man, I could’ve done this all without having to know Python using the scripts and versus power query and a lot of those type of things. So yeah, it’s exciting to see where the technology’s going.
[Datarails ad]
If someone was to read your project, what do you hope they take away from it? What do you want them to, what’s the one thing you want them to walk away with from the project?
Glenn Hopper:
Yeah, I guess mean, you know, and I are geeks like this, so you said the <laugh>, the paper was interesting and I was fascinated by it for the project as it stands right now, it is, again, it’s a lot of looking at how the sausage is made and so there’s, there’s big code blocks in there, so it’s 30 pages long, but a lot of that is just the code that chat G P T, yep. Generated. But I think, yeah, I
Paul Barnhurst:
Will admit I didn’t read all the code and skip those parts.
Glenn Hopper:
So the takeaway for readers of it is if you can endure that and honestly the audience we’re talking to, this isn’t like a broad we’re just talking to anyone who’s doing accounting and finance, we’re talking to our kind of geeks right here, we’re talking to FP&A people who like to get down into the weeds and do all this stuff. So hopefully some of your listeners would find it interesting as well. And also because of the nature of our job and what’s out there a lot of your audience will be better coders than I would ever be and just kind of understand the fundamentals of this. But I guess what this, I can write a SQL query, it takes so ridiculously long to do it and it’s the worst use of my time is to actually be sitting in front of a blinking cursor trying to do that.
And I guess the thing that I want people to take away from this is whether it’s written a SQL query, it’s for some reason it’s not working, I can dump that into chat GPT and it’ll QA it and find out where you left the comma out and save so much time over what we were doing. But what I really was amazed at and it took me several weeks to do the paper, but primarily because Chat GPT was crashing all the time because of such high use. So I’d have to wait for chat GP to come up, but what I hope someone would take away from it is that I didn’t write a single line of code for that, which means they wouldn’t have to. So it’s all about sort of prompt engineering and chat GPT and then seeing, going from having these three CSVs of financial statements to a chatbot at the end that you could ask specific questions of it without writing a single line of code. I mean that to me is amazing and it shows the potential of the future and I hope it gets people excited about what’s out there.
Paul Barnhurst:
And I definitely think it will. And along those lines, in my last Excel training I did, I brought in chat GPT and had it write a number of formulas and showed people, look, it’s another research you can use in addition to Google, in addition to Microsoft’s help. You just need to be careful and audit it, make sure that what it’s giving you is, cause I showed, hey, in these cases the formulas had some issues and here’s why. Anyhow, is that really good reminder that you can use it? You can use it today. There are areas that can help you in your job, but don’t just blindly trust it.
Glenn Hopper:
That is a great, I’m so glad you brought that up. So in these generative models they call when I mean Chat GPT will give you no matter what it’s answering, it will give you with such certainty the base reality, this is the truth of everything. And they call mm-hmm I heard this the other day. They call those hallucinations. So when this predictive text model gets it wrong, it’s just hallucinating something and there’s seen all the people go out and hack the prompts and get chat GPT to have an existential crisis, amazing what’s going on. So that is such a key point with it and whether you’re using it to write content or if you’re a middle school kid trying to use it to write a paper for you, just generative text, it’s not sentient. And when it writes the code as well, it’s kind of like these drag and drop machine learning tools.
There are things out there that it’s amazing. You can bring your data in, tell it what you’re trying to predict on and it’ll go do some kind of random forest and come up with these predictions. But if you don’t know what a random forest is or how it works or what model would actually be the best to use, then you’re putting yourself at risk and you’re going to confidently report on something that is dead wrong. So that is a key point and you’ve actually crystallized it better than I did in that it is a tool that’s out there to help you. But in that paper, if something were wrong, if you were just relying on chat FPT to do queries and you didn’t have a base understanding of it, it could be querying the wrong thing. You’re putting together a board a report for your board of directors or whatever, just like you wouldn’t turn over if you’re writing text messages, if you’ve ever done that where you just completely turn it over to the predictive text to respond and you just sort of follow the trail of what I mean that would be the same thing.
And we’re not there yet at all it with letting Yeah,
Paul Barnhurst:
I’ve got myself in trouble more than once, not noticing the auto correct, change the word and I’m like, ooh, did I just say that in the text? Yeah.
Glenn Hopper:
Or sorry, chat. G P T is a great tool to have out there as long as what you’re doing and it’s a human machine collaboration and it’s like a really cool financial calculator or knowing R or something. It’s just another really cool tool in your tool belt
Paul Barnhurst:
That that’s a really cool way to think of it. Kind of like that financial calculator or when we first got the pc, how much did it simplify our work? I think chat GPT is a lot like that. Or when we first got a spreadsheet, as you mentioned QuickBooks, what are those moments that have revolutionized the way we work? Getting a calculator, revolutionized it, going to a computer all of a sudden having a spreadsheet, QuickBooks on the accounting side and now for finance and other industries chat, GPT and AI in general is going to have ways that it can help revolutionize the way we work if we’re, and so it’s important, like you said, to understand enough to be able to use it. You don’t need to be a tech geek, you don’t need to write code. I’ve always been a big fan of people learning basic SQL learning Excel. Well, I came from a little of a data background because I did report writing for a while and it’s served me a lot of value. So people need to understand it, but you don’t need to be a tech geek. And I think there’s a clear difference between the two.
Glenn Hopper:
Absolutely. Yeah. And the really cool thing is that chasm is getting narrowed more and more by how quickly and how far beyond where it started the technology’s going. Yeah, back to your earlier question. I mean we’re there now, we just need somebody some if you have any young entrepreneurs in your group looking for a startup I really think that building a tool that focuses on this, that the base technology’s there, it’s just applying it to what we do.
Paul Barnhurst:
Great. So the kind of question, we’ve talked quite a bit about technology here and I have one more question along those fronts. So think out five, 10 years, how do you see the FP&A departments different from today and maybe how do you see beyond just chat GPT, but technology in general shaping kind of FP&A and finance?
Glenn Hopper:
Yeah, I think so. I mean there’s a couple different ways to go on where the technology is. And I think the two things that I want to look at that I would say are going to be vastly improved are sort of the whole back office system, the software that’s out there, how connected the software is and how data moves from one system to another. And once all your different, whether it’s a single ERP actually covers everything and all your data lives in that, or if it’s multiple systems that are linked together and you have this unique identifier that identifies records and transactions all the way through all these systems and you have more and more data, that’s the base that’s happening with the data that’s available to us. So if maybe your CRM didn’t used to speak to your accounting or your project management or your fleet management, whatever your system is that’s out there, you have these independent siloed records that are not connected.
And so if you’re trying to get a true picture of your customers and of the transactions that’s going on, it’s very difficult. So we’ve seen over the last decade certainly, but even back before that, trying to tie the information from these systems together. So as the connection between the systems gets more, as they get more integrated and the data becomes more cohesive, whether you’re doing a data lake or data warehouse or whatever or whatever you’re doing to capture this data, when you can identify it from end to end and you have this much more data. The second part of that is the data science component of it. And the more data you have access to from all these systems, then the more predictive you’re going to be able to get because you’re going to have more correlations and more features to build out these models.
So right now, if you don’t have access to maybe very early customer data from the CRM or maybe you don’t have access to information about when the customer churn and you can’t see, so building models will get better. And then the way you build the statistical models and find correlations, the more data you have. And I briefly touched on this in the paper, just finding things that are related. How does the cost of sales relate to the inventory levels or to whatever it is. But I think the more data you have, sure, the better you can train these models. So the two fronts that’s going to go on will be the system integration. And I’m seeing more and more now, and we talked about this a little before we went on air, the combination in when people are getting financial degrees or MBAs right now, the combination of analytics with that.
So we’re going to have a more trained group of people that is trained more based on machine learning and the what’s out there. So it’s going to be these two things moving together and FP&A and automation will increase with it too, but the people who are using the automation, everything will be more informed. So we’re going to see some pretty significant changes. And then one thing, and one last thing I’ll say about that is I think back to the beginning of my finance career, the amount of data entry that went on and sort of the mindless work that we had to do. And if you’re entry level finance person, you’re doing a boatload of data entry. And I think as this automation increases, that mindless work goes away. And the real value that we give to the organization is this mindful work that we do and we spend less time, like you said, 60% of FFP&A is data cleaning. Well, we spend less time on that and we’re adding a real value and hopefully keep showing more and more how we can drop that cost center moniker that we get stuck with a lot and say No, we’re how we’re providing real value to the company.
Paul Barnhurst:
Yeah, that’s exciting times because that’s where every finance person wants to be is really in those strategic discussions being viewed as a value creator, having that seat at the table versus, all right, well I got Delta Air and Delta Airlines over here, are they the same customer? And how do I make it all work? We’ve all been there, whatever the analogy is, Ford and Ford Inc. I could name any big company and you could probably have four different ways if four only probably doing pretty good, four are more ways that it’s spelled between your different systems. And it’s like how do I get this all to end to end as you talked about, really understanding that full life cycle of your customer and what that brings. And so I think you’ve brought a lot of great points there. So I know you’ve been a CFO of a number of different companies. So kind of question there, as far as the CFO, when you’re looking for somebody in FP&A, you’re looking to hire somebody or bring someone in to help without budgeting and forecasting, what skillsets do you look for today in somebody? I
Glenn Hopper:
Mean, it is so much more the analytics now and obviously the importance of understanding the fundamentals of finance, they have to have that too. But I guess going back to what I just said is I’m seeing more and more the finance comes with the analytics as well but whether or not you’re officially, even if someone I’m talking to, if Excel was the only tool they used, but they get the difference between linear regression and polynomial regression, it’s just the understanding of the basic math of it. And actually a solid, and when I started, I didn’t realize this when I started in finance, but an understanding of statistics is so key to me because I will say, and I guess that all the people I’ve hired have had a finance background, but if you took a statistician and gave them the basics of what we’re looking for here and whether in the income statement or whatever and had them apply their statistical knowledge to finance. To them the number is the number and to get them started, I, I’d say a really good statistician could learn the finance side. Just really if you’re a finance person, you could learn the statistics side as you go. But I think just that having that analytical and mathematical mindset is really the key part of it. And if someone has that in that foundation, you can almost teach the rest. It’s like the Southwest Airlines when they hire, they’re hiring for a certain personality type and an understanding but if you have that foundation, I think you can train on the rest
Paul Barnhurst:
I agree with the idea of foundation, whatever that foundation is, you’re looking for an employee, you can train the rest. You want to get those basics. And I think when we talk about statistics, I do think there’s an opportunity for finance to use that a lot more. And we’re starting to see it. In fact, this week I was working on an article that was writing around Montecarlo simulation, and so that’s something that isn’t used enough. I haven’t used that in my career, I learned it in school. And then you get in and it’s like, oh, you do some simple weighted probabilities. And it’s like, right, why should I wait three scenarios when if I have the right tools, I can look at 10,000 different scenarios and get an expected probability and a lot more information to reduce risk and increase accuracy? Because that’s really what it’s about is you’re trying to make sure you’re best deploying that next dollar of capital and managing the risk because the risk’s going to be there regardless whether you use Montecarlo statistical or just scenarios or however you build the model, it’s really helping you to manage the risk and to be more accurate.
At least that’s kind of how I think of bringing in statistical modeling into finance. Would you agree? Is that how you kind of see it?
Glenn Hopper:
Absolutely. And I love that you said that because that was actually, I took CS 50 the Harvard Computer Science David Mallon’s course, that’s, it’s on edX and all the free platforms out there now. It’s probably the most popular undergrad course at Harvard and I think one of the most popular online courses. But my final project for that was a Monte Carlo simulation for FP&A, and I loved it. It is amazing. So yeah, instead of doing just your worst case mid and best case, I mean we’re going to do 10,000 of these and see what comes up. And it’s pretty amazing out there. And I think having that statistical understanding opens up all these new tools that are out there that you can use to improve forecasting.
Paul Barnhurst:
I agree. Well, I could probably talk all day on the tech stuff with you, but I’m not sure our audience wants me to go that long. So we’re going to move into some of our more standard questions here. I have about four or five questions for you and then we’ll let you go. So first one, this is one we ask everybody. I’ve always been a big believer that failure, as the world classifies, it leads to success, that failure is a learning opportunity. So can you describe a time you’ve, had it experienced a failure at work? What did you learn from the failure?
Glenn Hopper:
Actually, I love this question because my biggest corporate failure is kind of my origin story. Million years ago when I was in telecom, I was the budget director and I was managing $150 million budget and it was me and my procurement guy and that was it. And we worked like crazy and it was a lot to do and just no resources to do it. And my procurement guy was great, and he knew more about procurement process and all that than I would probably ever know. But he was very, very old school. And when I say old school, this man kept everything on a paper ledger. He had a, there’s a big CRT monitors, he had one of those on his desk. It never even turned on. He had just piles and piles of paper on his cube.
So in this era and this, everything was siloed. I was on the op side. Our controller did not want to give us any access to a system. If we wanted something from him, we’d ask for a report, someone on his team would get a report to us. It might take 24 hours, it might take two weeks. So we had no real-time visibility into data and we were coming up at the end of the year. I think our CapEx budget was like $17 million, well, $15, 17 million in there. And didn’t, my procurement guy tracked all the invoices in his paper ledger and we’re coming up towards the end of the year, we were VC backed company, and we had board meetings come up. We were really tough times for telecom in the early 2000s. And we knew this, yes, board meeting was going to be contentious and we’re coming up towards the end of the year trying to just really laser in on the budget and on say, a $15 million budget in November where there’s no time to do anything we find there was one extra invoice from one of our suppliers that we’d lost track of. It was only $1.5 million and it had been paid, so the finances of it were fine, but when 10% of your budget is off by one invoice that you didn’t have any visibility into, it was this ruined my Thanksgiving that year. Me, the CTO and a bunch of other people worked <laugh> all the way through Thanksgiving weekend trying to figure out how are we going to save face , we’re about to get killed. We were sure we were all going to be fired, and I was ready to go teach high school by the end of the weekend. I was like, well, my finance career is done. I’m going to go just do something different. But we got through it and we had a really painful board meeting that we had to get through with all that.
And it was such a disaster that once I came out the other side of it, I realized I’m never going to operate where I don’t have full visibility into all the financials and the ability to see all this again. And nothing is ever happening on paper. We’re going to have a slug trail for everything that we do. And anyway, it was the most painful part of my career, but it really, and this would’ve been, I dunno, 2000, 2004 or so probably, but it really changed the trajectory of my career. And I got the support from some senior management at the time, and actually from that got to build the company’s first business intelligence team. I had back in the day crystal report writers and everything that came out. And we took over the metrics for the whole company. And really from then, the things I got out of it were visibility, data democratization, and being able to see things from end to end. And that’s also where I got my idea that finance, and this is a whole other conversation, so we’ll save that for another day, but that finance needs to be the owner of all the KPIs for the company, not just the financial metrics.
Paul Barnhurs:
We’ll save that one given where we’re at for another day. But I tend to a agree there’s a lot of benefit when you have cross-functional metrics to have them all sit in finance as kind of that owner of them. So I definitely think there’s some value there. And I, you brought back some memories when you said crystal reports, so I remember that system as well. But it sounds like it really helped you on your path and helped define your career. So it was a defining moment you ended up l learning a ton from, but I’m sure not fun in the moment.
Glenn Hopper:
No, I it was horrible
Speaker 3:
You sharing
Speaker 4:
That example.
Paul Barnhurst:
Yeah, not fun at all in the moment. I can relate to some of those type of things. So I think you’ve talked a little bit about probably the biggest opportunity for FP&A moving forward with the way technology’s going. What do you see as the biggest challenge for FP&A moving forward?
Glenn Hopper:
I mean, I do. I think it is, and it’s not just for FP&A, it’s for all fields. We become specialists and experts in a field, whether it’s law, marketing, finance, you become an expert in that. And then because technology is being deployed so much out there right now, the challenge is now we have to become experts or at least have a very sound working knowledge of something that’s different than what we do every day. But if we don’t keep up with that technology, if you don’t kind of get on board and ride the wave of this AI and machine learning and all the transformation that’s going on technologically, you’re going to get drowned by the wave rather than move forward on it. So I think the challenge is going to be how do I keep up with whatever the latest accounting rules are on treatment of leases versus, and at the same time keep up with whatever the latest technology is that’s out there. And I wish there were an easier way to do it, but I’m climbing to the top of the mountaintop and just preaching to anyone who will listen. You can’t ignore this. You have to embrace it, dive deep. And I, does that mean maybe you need to go on if you haven’t already, go on Coursera and take a SQL course? Yeah, sorry. It probably does, but FP&A guys, we all knew. I mean, I think I am preaching to the choir here, but
Paul Barnhurst:
Yeah, I hear you. So it sounds like that biggest challenge is just being prepared and riding the wave versus being swallowed by the wave of what’s coming. And yeah, I tend to agree with you knowing the basics of SQL, I’ve always preached that FP&A should at least know the basics. You don’t need to be hard code and be able to write big detailed SQL, but be able to understand it well enough that if you need to pull something or review what Chat GPT wrote for you to pull something, being able to do that. So I think you and I are on the same page there. I’ve had that discussion with a number of FP&A people and get different answers. So the next question, this is where we get to the personal question of our interview, what is something unique about you that you can share with our audience? Something we wouldn’t find online?
Glenn Hopper
I wrote and produced an independent film in 2007. It’s called The Hanged Man. It’s horrible. Like, nah, I don’t know. That’s not fair. Well, I don’t know.
Paul Barnhurst:
You’re not giving an ringing endorsement. If I see this on Rotten Tomatoes, I’m probably not watching it. It’s
Glenn Hopper:
Horrible. Yeah. So I’ll say this, for a first time filmmaker who went from <laugh> nothing to completed film that was distributed and out there on Netflix for a while, and I think you can still get the DVD somewhere. Don’t, it’s not streaming anymore. So for a student film, it would’ve been pretty good, but it was I don’t it. Yeah, so it’s called the Hanged Man. It’s about some social misfits who are drawn in by this cult-like leader to a barn in the middle of nowhere and hijinks happen and Hijinx happens. So that’s way off the beaten path for probably most FP&A folks to have done that. But yeah, I
Paul Barnhurst:
I Can’t say that’s what I expected. So I appreciate that. That’s the first we’ve had, so I like that. So now this is one of my favorite questions I ask everybody. Our sponsor Datarails is a platform built on Excel. So your favorite Excel formula feature kind of function, what’s your favorite thing about Excel?
Glenn Hopper:
I probably the one are, so I’m going, I’m trying, kind of running through the laundry list right now thinking about, ooh, index matches cooler. And really for me though, it’s probably <laugh> the closest because I use it so much is just when you don’t have that unique identifier or when you are trying to combine data from different sources. The thing I use in Excel, well now I’m thinking pivot tables. Okay, I’m just going to say VLookup, which was my initial answer I was going to say, because <laugh> taking so much information from different tables and trying to match those and bring them together for the people who can’t do that, that’s probably the one that I use the most. It’s not the sexiest but I am using it daily.
Paul Barnhurst:
Good old faithful and lots of people use it. I totally get it. A good lookup, whether it’s V lookup index match lookup, yeah, X lookup, as long as power query, whatever. Plenty of ways to do a lookup, but it’s one of the most valuable things in Excel is to know how to do it. So I can totally understand that one. So next question here, what advice would you offer to someone starting a career in FP&A today? So if you could give ’em one piece of advice, what would it be?
Glenn Hopper:
Yeah, so and we touched on this earlier and I think the technology’s going to change. So having a technology angle is good, but going back to what I said earlier, a good knowledge of statistics is going to go a long way. So beyond just the finance and accounting, and that’s probably part of most finance curriculum right now, but I think pay a lot of a attention to it and really think about applying that and building statistical models. And because then it’s no matter what the technology is, and if you’re going into more of a data science angle with it, having that just sort of basic statistics understanding will dictate what machining learning algorithm you’re going to use or how you approach a problem or how you look for features to put the model in correlations and ways to really hone the model. And it would really be take statistics seriously, and if you’ve got some electives, maybe spend them on a couple more of of statistics.
Paul Barnhurst:
Great. I think that’s the first one that’s given us the statistics answer. So I like it. I like to get different answers and I’ve really enjoyed the time with you here. Last question. If somebody wants to be able to get ahold of you, if they’d like to reach out, what’s the best way for them to contact you?
Glenn Hopper:
I think LinkedIn is probably the best. It’s really the only social media I’m active on at all. And yes, I think LinkedIn is probably the best way to get me.
Paul Barnhurst:
Okay. And then if anyone wants to maybe read your book or see your chat GPT, we’ll put those in the notes so people can access that. And again, just wanted to thank you for being on the show. We’ve really enjoyed having you. I know I’ve enjoyed chatting with you, and I think our audience will really enjoy this episode. So thank you for your time, Glenn.
Glenn Hopper:
Thank you Paul. Really appreciate it.
