Episode 3 - Benefits of competing in the Financial Modeling World Cup- Insights from World Champion Diarmuid Early
Show Notes
Welcome to Financial Modeler's Corner (FMC) where we discuss the art and science of financial modeling with your host Paul Barnhurst. Financial Modeler's Corner is sponsored by Financial Modeling Institute (FMI) the most respected accreditations in Financial Modeling globally.
In this episode, Paul Barnhurst is joined by Diarmuid Early, one of the best modelers in the world and, the Founder of Early Days Consulting. He is a regular competitor in the Financial Modeling World Cup (FMWC) and is often referred to as the Michael Jordan of financial modeling for his impressive results. He is the winner of the 2021 FMWC and a runner-up in the 2020 and 2022 editions.
Listen to this episode as Diarmuid shares:
· His journey and background into Modeling
· How Financial Modeling competitions help you grow and learn
· The key differences between modeling for competition and modeling for work
· His practical advice on how to build a 3-statement model
· How assumption ambiguity and structural ambiguity impact your model.
· His position on controversial modeling issues including circular references, dynamic arrays, modeling standards and more
Financial Modeler’s Corner is sponsored by Financial Modeling Institute.
Sign up for the Advanced Financial Modeler Accreditation or FMI Fundamentals Today and receive 15% of by using the special show code FMC15.
Visit www.fminstitute.com/podcast and use code FMC15 to save 15% when you register.
Go to https://earmarkcpe.com, download the app, take the quiz and you can receive CPE credit for listening to the episode.
Follow Diarmuid:
YouTube - https://www.youtube.com/@DimEarly/
LinkedIn - https://www.linkedin.com/in/diarmuidearly
· To learn more about FMWC, click below - https://www.fmworldcup.com/
· If you want to know about FMWC Rankings of the top competitors, visit - https://www.fmworldcup.com/rankings/
· Stream the events of FMWC live on YouTube by visiting - Financial Modeling World Cup
· Link to the next streaming session on July 27th at 9 AM MST of which Paul Barnhurst will be a part - Battle VII - Road to Las Vegas 2023 (Excel Esports)
Follow Paul:
Website - https://www.thefpandaguy.com/
LinkedIn - https://www.linkedin.com/in/thefpandaguy/
Instagram - https://www.instagram.com/thefpandaguy/
TikTok - https://www.tiktok.com/@thefpandaguy
Twitter - https://twitter.com/TheFPandAGuy
YouTube - https://www.youtube.com/@thefpaguy8376
Quotes:
“You need to have certain discipline around what you want your model to do.”
“Getting better at excel has been more valuable than getting better known by winning the FMWC.”
“Do terribly struggle for an hour and a half with what should have been a 30-minute question, but eventually get it out. You'll learn a ton from that.”
“The big value in the competition is a learning opportunity and you don't need to train to start learning.”
“The person who is most valuable as a financial modeler is the one who can tell you what are good assumptions to make.”
In today’s episode:
(00:22) Intro;
(00:45) Welcoming Diarmuid;
(01:16) The worst financial model Diarmuid has ever seen;
(04:13) Diarmuid’s take away from the worst financial model;
(07:45) Diarmuid’s background;
(09:45) How Diarmuid became interested in Financial Modeling Competitions;
(11:48) How Diarmuid felt after winning the competition;
(12:51) What Diarmuid enjoys about competing;
(18:05) Financial Modeling for competitions vs. Building it for clients;
(22:37 - 23:23) Validate your Financial Modeling Skills with FMI’s Accreditation Program (ad);
(23:30) Diarmuid’s top advice for competitions;
(25:10) Rapid Fire;
(29:45) Nugget on how to become a better modeler;
(34:40) Connect with Diarmuid;
(35:40) About FMWC;
(37:55) CPE Credit with this episode;
(38:30) Outro;
Transcript
Host: Paul Barnhurst
Welcome to Financial Modeler's Corner, where we discuss the art and science of financial modeling with your host, Paul Barnhurst. Financial Modeler's Corner is sponsored by Financial Modeling Institute.
Welcome to Financial Modeler's Corner. I am your host, Paul Barnhurst. This is a brand new podcast where we will talk all about the art and science of financial modeling with distinguished financial modelers from around the globe. The Financial Modeler's Corner is brought to you by Financial Modeling Institute.
FMI offers the most respected accreditations in financial modeling. I am excited to welcome this week's guest on the show, Dim Early.
Dim, welcome to the show.
Guest: Diarmuid Early
Hey, thanks for having me, Paul.
Host: Paul Barnhurst
Yeah, we're really excited to have you. If you're not familiar with Dim, he's a regular competitor in the financial modeling world cup, including winning it in 2021 and being the runner-up in 2020 and 2022.
So we're really excited to have him today on the Financial Modeler's Corner. We like to start the show with just kind of a fun question that we ask everybody to see their horror stories. So the question we start with is, tell me about the worst financial model you ever saw.
Guest: Diarmuid Early
I think there's two categories of this, that go in kind of completely different directions.
The one is the small model where someone has just been kind of using Excel as a scratch pad. You know, they found a number somewhere, they found another number somewhere, they just type the two in and multiply them together or something. And then like 10 more calculations, I don't know, like a market sizing or something, but there's no labeling. It's just like, you know, one cell equals 4,735 times 367, the next cell is like that divided by 313 is like, where did these numbers come from? What do they mean? You know I've come across a number of times like I started my career in consulting where you know people are quite focused on making the outputs, the slides look very pretty but not always the best about keeping consistent workings in the background, so that's one category is just I have no idea what's going on here and you know that might only be, you know, a half dozen cells or it could be bigger. And then the other category is the model that is too large for its own good, where it's, you know, spread over 40 different tabs and each one of them has hundreds of rows. And it's just like a spaghetti kludge of everything points to everything else and good luck unwinding what's going on.
Obviously key feature of all the worst models I've ever seen is that nobody has documented anything, what's the logic, what's the flow. I've had a couple of, assignments in my consulting business where the brief has basically been, can you make this model any good? And nobody really knows what exactly the model does or how it does what it does. So it's like, can you just work through this whole thing and unpick it all? I'm like, oh, I have to.
Host: Paul Barnhurst
Yeah, I've had to unpick a few. I mean, not so much model. I had one time more going back to, I'll date myself, Microsoft Access and somebody left, and no instructions and trying to unpick like 50 queries and streamline this process because using like five different databases and it was just a huge cluster and it's a lot like that massive spreadsheet, right?
Cause you're just trying to unwind one thing after another and slowly work through it and trying to make sure you can get back to the answer, like, okay, why was it producing that answer? Is that right? Or is it not even right, and I should ignore it?
Guest: Diarmuid Early
Yup. It's a nightmare.
Host: Paul Barnhurst
It is definitely not one of my favorite things to do, but it's rewarding when you actually get to that output and you see something that goes from being a total nightmare to actually adding value for the client.
Guest: Diarmuid Early
Yeah, I mean, I feel like there's definitely some value in forcing people to inherit models a few times early in their career because it gives you a sense of why some of the best practices that people talk about like labeling and documentation and that kind of thing are actually helpful.
Host: Paul Barnhurst
It's definitely something I've, learned over my career, you know, the importance of that. You know, speaking of those two categories that you talked about there, the overly complex, no documentation, just a huge sprawl cluster, so to speak, and the Scratchpad. What have you learned? What are kind of your takeaways from seeing those kinds of models that you take to heart as you work?
Guest: Diarmuid Early
I mean, for the Scratchpad one, it's, it's mostly label things. Ideally, well, especially if you're going to actually, you know, put this into a client output or something, label them, including a source, like whatever here is the number of cars in Australia came from this website. It can be very simple, like, you know, take 10 extra seconds to do it, but then it means that somebody can reproduce it or audit it or whatever.
I mean, I guess that's kind of the driving principle behind it, right? Is like in theory, when you hand over your model, somebody else should be able to check your workings and, you know, in the Scratchpad situation where it's like, all right, I'm trying to whatever, like classic consulting market sizing thing, I'm trying to, you know, figure out X while I'm, I'm pulling together a bunch of different, you know, inputs and data from lots of places. Like the whole idea of being able to reproduce that is where did you get these numbers from? And you know, are those sensible places to get the numbers from?
I mean, for the enormous spaghetti version, I think the biggest thing there is you need to have a certain discipline around what you want your model to do, because even with the most impeccable design, if you want your model to have 100,000 different outputs, then it's going to have to be huge and complex, and that's going to bring with it a lot of, you know, execution risk or, you know, risk of things becoming inconsistent, that kind of thing.
You can get around a certain amount of that with, you know, an elegant design and that kind of thing, but actually, there's almost a more basic problem to deal with first, which is, are you tasking this model to do more things than it's reasonable to do? Would it make more sense to, you know, go 80-20 on it? What are the, what are the 20% of things that will give us 80% of the value, and could we have a one or two-tab, simple model that will give us what we really need rather than trying to answer every different question under the sun with a level of precision that's maybe misleading because it's just assumption layered on assumption.
Host: Paul Barnhurst
Totally agree with you. I've been there on, you know, that side of that 80-20, finding the balance. I once built a model for a customer and he wanted the ability to adjust something like, I think I had over 10, 000 variables. I tried to convince him to go lower, but that's what he wanted. And so that's what I built, but I'm just like, why? You're going to have errors throughout this model, and it's not going to be the model it's going to be that you have old inputs that you don't even know about.
Guest: Diarmuid Early
Yeah!
Host: Paul Barnhurst
Because there's too many. How do you validate that 10,000 inputs are right? You can't in a reasonable manner as a human.
Guest: Diarmuid Early
Absolutely. And where I've seen this a lot is, you know, we want to model each country separately and each product within each country separately. And we want to have separate assumptions for all of those. Aside from the difficulty of maintaining it, it's almost certainly not realistic to say that you have independent assumptions for your level of demand growth in every different country in the world. You know, and if you pretend that it is, then whatever, you know, rather than having some simple assumptions that like, I don't know, we'll assume developed markets will grow like this and developing markets will grow like that, or by content or by reading, like whatever grouping makes sense, if you just make them all separate assumptions, then the country team over here makes one assumption about how they're going to grow, the country team over here makes another assumption about how they're going to grow and they're not really logically consistent. And so you're kind of missing out there.
Host: Paul Barnhurst
Yeah, it makes total sense to me there. So I appreciate that. We, you know, spent a little time there on worst models and how to think about, making sure you don't end up on the websites, right? That you see worst model ever or worst visual or those types of things. You know, we haven't had a chance to ask you, can you give our audience a little bit about your background? Just tell us a little bit about yourself and how you ended up, you know, where you're at today in New York.
Guest: Diarmuid Early
Sure. So, I grew up in Ireland. I studied math and physics in college. I then did a PhD in Computer Science and I started my career at the Boston consulting group in London as a junior consultant. And I guess, you know, when I was at BCG in London, it was like an office of 150 people or so, so it was like not quite at the level where everybody knew everybody, but pretty close to that. And so I guess I acquired a reputation relatively quickly for being the go-to guy for Excel in the office to the point where, you know, my last six months or so there, I would basically have almost a drop-in clinic at my desk every Friday where people would just like cycle bikes, you know, people were often traveling Monday to Thursday, I would just have like a line of people being like hey, how can I do this?
It's a great way to learn because, you know, the nature of the business is that people were, were asking all kinds of completely different questions. So I did that for a couple of years, and then I moved to Deutsche Bank in London. I was like a junior business manager in the corporate finance business, and I moved around various flavors of that role and becoming a little bit more senior over the next many years and moved to New York with them a little over 10 years ago.
And then after doing that for a while, I started kind of working less like shifted to part-time when my second son was born. And then at some point we thought, okay, well, Dim wants to work even less and at some point, you can't call somebody an employee anymore. So we basically made a deal that I would keep like a little sliver of my, the sort of most valuable pieces or the most hard to replace pieces of my job and turn that into a contractor position and so then I went and set up my own firm, and I've been a sort of freelance Excel guy ever since.
Host: Paul Barnhurst
Got it. Great. I appreciate you sharing that background. So, that leads me to the next question. Obviously, you quickly became known as the Excel guy around the office, but how did you become interested in competing, you know, in financial modeling competitions and these Excel competitions? How'd that come about?
Guest: Diarmuid Early
So, it's funny. I took to Excel very quickly as soon as I discovered it. It came relatively naturally to me. I enjoyed it, and like I said, I had a kind of reputation within the office. But, you know, that was an office of 150 people. You quickly get the sense when you look around that anytime there's a group of three or more people in the world, there's one of them who's the Excel expert and that might be, you know, among the sort of retirees in the church group, well, Bob is the expert because he knows how to change cell colors and font colors and, you know, most other people can only change one of those two, but literally, no matter where you are, there's someone who's the Excel expert. And I guess the thing that appealed to me about the idea of the competitions was that I had absolutely no idea where I stacked up versus anybody outside my office.
And, you know, was I just kind of Bob and the church group on a slightly different scale? I mean, it didn't take much to convince me to give it a go because, you know, Excel was fun and I'm a competitive kind of person, but that was the original appeal. Well, that and the prize money, because, back then there was, I think the first year I did it, there was like a 30,000 dollars first prize for modelers. So that was also quite appealing. I did not expect to do, well, I didn't even expect to do as well as I did, I certainly didn't expect to win, but it was an interesting idea. But yeah, mostly it was just, I wonder, am I just like a big fish in a small pond or am I actually quite good at this?
I don't know. Let's find out.
Host: Paul Barnhurst
That's a great reason to do it. And so now, did you win your first year competing?
Guest: Diarmuid Early
No, I didn't. I totally blew it my first year, year , I mean, like I said, I did better than I, I did not expect to make the finals, which I did.
Host: Paul Barnhurst
Yeah. I can't say you blew it if you made the finals, that's pretty good for a first year.
Guest: Diarmuid Early
Yeah. Well, I blew the finals, but I did okay in the qualifiers. But yeah, I won my second year.
Host: Paul Barnhurst
Cool. Well, that's exciting. So you won, that was the old one, and then you've also won under the financial modeling world cup, right? So you have the two different wins.
Guest: Diarmuid Early
Yep. That's right.
Host: Paul Barnhurst
So what was it like, what's that, you know, feeling like winning the competition? How has that felt for you?
Guest: Diarmuid Early
I mean, it was super exciting. Partly because I was, you know, a young guy and not exactly broke, but not exactly rich either. And so the idea of winning 30,000 dollars was extremely exciting. But also it was, like I said, just the format I found like engaging, energizing, stressful, but super enjoyable, and so, back in the model off days, there was like a live final event. And then, you know, later that day they would announce the winners. So, with the Financial Modeling World Cup, it's sort of a league over the years, but even by the middle of the year, you've got a pretty good sense where things are going to shake out. Whereas with that, you just have no idea.
It's like, this felt like it went pretty well for me, but you don't know. Maybe that's because the questions were easy and it went well for everyone, and it's just like total surprise.
So yeah, it was very exciting.
Host: Paul Barnhurst
I bet. I bet you're thinking about, all right, how could I use this 30,000 dollars? This will be fun. Among other things, that's really cool.
So, you know, you mentioned you like, the form you're a competitive guy, but what is it you find most exciting about competing each year? What's kind of the, maybe the things that you enjoy most in that?
Guest: Diarmuid Early
I would say what I enjoy most is just, it is exciting. There's not a lot of things in my life where, A I'm trying to, you know, achieve something hard in a like fixed time, like in two hours in this case, but like in a fixed short time period, B- it's like just in that sweet spot of difficulty where I know that I, theoretically would be able to do it, but it's going to be really hard to get it finished. I have a lot of friends who do this competition. So there's a real kind of social element to it. It's like, oh man, this is going really badly, I wonder how Lawrence or Andrew or Michael or Villum or whoever is going to be getting on with this one. That aspect of it is fun as well.
Host: Paul Barnhurst
I'm sure you got a lot of people that do a repeat every year. I mean, I've seen some of the top names and you get to be friends with them, and so, yeah, you're like, there's a competition against your friend, and then there's also just wondering how they're doing as you're like struggling, or you blow something and you're like, okay, did they have the same challenge?How did they think about solving it?
I think that's one thing that anyone who spent any time in Excel knows, there's pretty much always multiple ways to solve it. There's pretty much never just one way to do something.
Guest: Diarmuid Early
For sure, for sure. The one thing I'd add, by the way, is my motivation for still doing it after, you know, 10 plus years, it's intrinsic, I like it, it's me, I'm a competitive person, I like it.
But when I think about other people considering it, like I talk to people a lot who are like, oh yeah, you know, it sounds like fun, maybe I'd try it sometime. I'm not sure, I've got a lot of enjoyment out of it, but the big value that I've got out of it is there's also like a training development angle that's kind of huge, which is if you think about, like, if you read about deliberate practice, if you've read, Malcolm Gladwell's very pop-sciency take of the 10,000-hour rule, a key part of the practice that helps you get better is immediate feedback and like quite specific detailed feedback. And it's the kind of thing that's very difficult to get on the job because, it would be odd for someone to ask you to build a detailed model to work out, you know, these, you know, 15 different financial variables and this ultimate kind of net profit or IRR or whatever, if they've already built the model and they already know all the answers and they're going to tell you whether you got it right or wrong.
Like you might do it once in training, but after that, someone will kind of sense check it like, does this seem crazy, but it's very unusual to have pure assumptions laid out clearly enough that there are exact right answers to all these questions and here are the exact answers and you can see, did you get them like, not kind of close, not passes a sense check, but did you get them exactly right? And if you didn't, something went wrong.
And that kind of feedback in, you know, these cases are typically designed to be solved in anything from kind of 20 minutes to maybe an hour at the very long end. You know, you do it, you see if your answer is among the choices, you see if you got it right, and, within an hour, you've got a lot of feedback. And then if you look at the solution model, or you just kind of try again, you rinse and repeat like two or three times on one problem and you learn a lot and you think about, over 10 years of model off plus financial modeling world cup, the number and the range of problems is huge.
So, you know, if you're like, oh, I'm kind of struggling with three statement modeling, you can easily grab like five past questions that deal with three statement modeling and try each of them, try them again, within like a couple of weeks, you can up your game enormously. And so like, that's not why I originally entered it, but that is like, I've gotten enormous kind of career value, so to speak, out of that kind of training and feedback.
Host: Paul Barnhurst
Sure. Yeah, that totally makes sense. And I would imagine you've also probably gained some, clients over the years that have learned about you through your competition.
Guest: Diarmuid Early
Yep. It definitely, it helps with the profile. Although honestly, I would say the getting better is, has been more valuable ultimately than the getting better known.
And the nice part about that is, you know, you get that value whether you come first or 75th. And this is again, one of the things I always tell people is like, you don't need to win or to care about winning for that matter, for this to be a smart way to spend some time.
Host: Paul Barnhurst
I've thought of doing it a few times and you're making me think that I really need to find the time one of these days, but we'll see.
Guest: Diarmuid Early
I recommend it. I mean, what's the worst that can happen? Often I think people's unspoken response to that, well, the worst that could happen is my name will be up there on the leaderboard with, you know, four points out of a thousand or something, but, they post the results anonymously if you want, and unless you score kind of pretty highly in FMWC. So you can get the benefits and remain under the radar, if that's what you're worried about.
Host: Paul Barnhurst
Sure. Yeah, no, for me, it's been a matter of just making the time commitment. If I show up last or middle or wherever, it's all good. It's a learning experience is kind of how I look at it, so one of these days I'll just have to make it a priority, cause it does sound kind of fun. I mean, I'm sure I would be probably that one that's four or three out of a thousand, but that's another story.
Something I want to touch on a little bit, you know, we've talked about kind of the competition side and all the benefits you've got about that and how it's helped make you a better modeler, but how is financial modeling for competition different than when you're building it for the corporate world, you're building it for a client? Cause I know, there's speed, there's some definite different elements, there are right answers, and usually, it's corporate, it's a lot of assumptions…
So maybe kind of talk through some of the key differences between the two.
Guest: Diarmuid Early
So I would say, the number one difference you touched on is clarity, because if you're going to, you know, have hundreds of people do a financial model and tell them whether they got the answer right or not, then you need the assumptions to be clear enough to say this is the exact right answer. Those cases are always written very clearly, very exact assumptions, and, you know, if you compare that to the real world, if you've ever found a case where someone says, these are the exact assumptions and I just need you to turn them into a financial model, you will not need any clarifying questions. I have not omitted any details that are relevant or necessary. You've got yourself a keeper.
That client is the best.
Host: Paul Barnhurst
I want them to tell me what stocks to pick because they can forecast the future. If they're that good.
Guest: Diarmuid Early
Exactly. So it's one of the things I've always said that, you know, when people talk about, winning FMWC or winning a model-off, I've always shied away from any implication that that makes you like the best modeler in the world because, in a real-world sense the person who is most valuable as a financial modeler isn't the person who can, you know, turn the crank on a set of assumptions and tell you what the numbers will be, it's the person who can tell you what are good assumptions to make, what are plausible assumptions to make, is this the right approach. For me, that's the big thing. You know, it's in two respects. One is not just the kind of numerical assumption, like, okay, there will be some interest rate on this debt, well, the model is going to be structurally the same, whether it's 5% or 8% or 117%, but the sort of structure of the model, like, you know, we talked earlier about if you're modeling something for lots of different countries and lots of different products, are you going to have independent assumptions for every country and product, or are you going to have, you know, group them up in a certain way, have something that's driver based, have a sort of single economic forecast scenario that impacts everything in related ways, that kind of thing. You know, those structural decisions as well as kind of numbers that can be more easily tweaked are already baked in what you get in a competition question.
That's, I would say, the biggest difference, like both assumption ambiguity and structural ambiguity, is like the core of building a model for a client because, you know, most of the clients who are commissioning a model like that, part of the reason that they're doing it is because they're not the expert on model building. So, you'll spend a lot of your time asking clarifying questions. You know, have you thought about whether you want it to be built like this or built like this? Have you thought about how much you need to vary this assumption? Have you thought about, are you sure that it's always going to be six categories or do we need to make it flexible so that there can be more categories added easily?
That kind of thing.
Host: Paul Barnhurst
You know, as you mentioned on the assumption side, I was really grateful I had a professor in college, in grad school that when he grade our models, one of the first things we were required to do on every exercise is we did a Porter's Five Forces just to understand the landscape and to think of it strategically.
Then we'd go through and, you know, you take the case, but making your own assumptions. And he really graded more on our analysis around how we thought of things and our assumptions, and, hey, did you get every single math, right? And did you get the answer I expected to see? And that really helped me realize that rarely, in the modeling competition is different, but like I said, in the real world, rarely is there a right answer. It's really about those assumptions and can you support them and do they make sense and what do they tell you?
You know, the rest is making sure you have a good structure and you don't have errors in the model, but those assumptions is really what makes it valuable.
Guest: Diarmuid Early
Yeah, 100%. And you know the challenge is that assessing someone's, you know, how skillfully have they designed this, how artfully have they analyzed it that kind of thing obviously somewhat subjective but also just enormously more time-consuming. So if you look for example at the FMI, their questions are often quite similar to competition questions, except that there is more ambiguity because they don't just say, did you get the net present value of the investment is 117 million? You know, someone looks through and kind of assesses every piece of the model, how logically did you build it? How clearly did you lay it out, all that kind of thing?
If you can get someone to check your work at that level of detail, that's much more helpful, but obviously, in a competition, that's not going to happen.
Host: Paul Barnhurst
Yeah, of course.
Commercial break:
In today's business world, financial modeling skills are more important than ever. With Financial Modeling Institute's Advanced Financial Modeler accreditation program, you can become recognized as an expert in the field by validating your financial modeling skills.
Join the Financial Modeling Institute's community of top financial modelers, gain access to extensive learning resources and attain the prestigious advanced financial modeler accreditation. Visit www.fminstitute.com/podcast and use code FMC15 to save 15% when you register.
Host: Paul Barnhurst
So last question. I think you've touched on this a little bit, but if someone out there is listening and they're thinking of competing, now, what's the top advice you'd offer someone?
Guest: Diarmuid Early
I'd say my number one piece of advice is just start. And what I mean by that is I've talked to so many people over the years who have told me some variant on, yeah, you know, I'd really like to try that, I just need to kind of brush up on my skills a bit first. In my mind, that basically gets it backward. It's like, I'd like to start training, but I need to train a bit first. I was like, no, no, no, no, like just, just do it. And that is your training. You'll do terribly in some rounds if, you know, if you haven't kind of done this kind of thing before, but that's fine, that's part of the process.
Like do terribly struggle for an hour and a half with what should have been a 30-minute question, but eventually get it out. You'll learn a ton from that. If you eventually don't get it out and you end up dipping into the solution model, okay, how did they get to here or right now can I finish it off from there, that kind of thing, you'll still learn a lot from that.
I think people hurt themselves a lot with this idea of, you know, I'll get ready first and then dot, dot, dot, because, you know, it's a competition. And so people naturally think of it from a competitive perspective. But I very much think, like I said, that the big value in it is a learning opportunity and you don't need to train to start learning. That is your training. So just do it. That's my number one piece of advice.
Host: Paul Barnhurst
That's really good advice. I like that. Cause yeah, I've, I've had the same, oh, I need to train for it if I'm going to do it. I think everybody thinks that. And I really like how you said that because you're going to learn through the competition. You're going to probably learn more than you would by spending five hours because you've got started and you've seen what it's like, and it can help you get more deliberate in your practice as well. That's really good advice.
So we're coming up on, this is probably one of my favorite sections that we have. We've interviewed a few people and we have, it's kind of our Rapid-Fire Questions.
So what I have is I have about, I think it's seven or eight questions here. For these questions, assume you're building a model for a client where they're going to actively use it and update it. So, not competition and you have to pick one side or the other. So we don't allow “It depends”, because that could be the answer to all of them, I recognize that. And then at the end, we'll give you the opportunity to elaborate on one. What you're thinking was there.
Guest: Diarmuid Early
All right, so it's like just a yes or a no.
Host: Paul Barnhurst
Yeah, it's just kind of, okay, if I had to pick one, which one would it be? Realizing you could find a situation where you could use both. So, you know, kind of thinking from the corporate side.
So, circular or no circular references?
Guest: Diarmuid Early
Hard no.
Host: Paul Barnhurst
VBA or no VBA?
Guest: Diarmuid Early
No.
Host: Paul Barnhurst
Horizontal or vertical model.
Guest: Diarmuid Early
I don't feel strongly about that one. Uh, I guess horizontal.
Host: Paul Barnhurst
Dynamic arrays in the model. Yes or no?
Guest: Diarmuid Early
Yes.
Host: Paul Barnhurst
External workbook links, yes or no?
Guest: Diarmuid Early
No.
Host: Paul Barnhurst
You said that with authority, most people do.
Named ranges versus no named ranges?
Guest: Diarmuid Early
Yes, named ranges.
Host: Paul Barnhurst
Okay. Do you think using a formal standard like, you know, smart or fast, yes or no?
Guest: Diarmuid Early
No. My dirty little secret is I don't actually know any of the formal standards.I suspect I sort of mostly follow some of them, but I'm sure there's lots of things that I don't follow and I don't know.
Host: Paul Barnhurst
There you go. And so the last one we have here is what is your lookup function of choice? VLOOKUP, Index Match, XLOOKUP, or Choose?
Guest: Diarmuid Early
I question whether those are the only four options, but I'll go with XLOOKUP out of those.
Host: Paul Barnhurst
I realized there could be more, but I wasn't going to come up with 20 options because I get it, there's a lot of ways to do it. And what are some of the others? Just go ahead and throw a few out.
Guest: Diarmuid Early
Well, I mean, I was being a little cheeky with that, but in the land of dynamic arrays, you can use things like SUM and CONCAT as de facto LOOKUP functions.
So if you have like a two-dimensional array of text with like row labels and column labels, and you want to extract the intersection of, you know, row A and column B, then you can do something like, CONCAT if blah, if it's text or SUM if blah, if it's a number. There's a whole other world out there with dynamic arrays.
But yeah, I'm an XLOOKUP fan. I still have VLOOKUP in my arsenal, which I know is a controversial topic, but it's mostly XLOOKUP.
Host: Paul Barnhurst
Well, I'm a big fan of Oz Du Soleil and what he says on this. He's like, it's not a religion. Just use what works. You don't need to have a battle about this. And I've had with many people and I'm like, I like XLOOKUP myself. I use it most of the time, but I'm like, I can see where you can use all these. That's totally fine. As long as it works and you understand, and what I like to say is you need to understand the pros and cons so you don't have mistakes in your model. Outside of that, use what works. It's really about being educated, not about there's a right way to do it.
Guest: Diarmuid Early
Yep, I agree with that completely. The one I would come, of the ones we ran through, the one I would come closest to saying that there is a right and a wrong answer on, while acknowledging that there's not a right and a wrong answer, is the circular references. I think they're, the way that I've most often seen them used, which is like, you know, calculating interest on an average balance where the interest feeds into the closing balance, that kind of thing.
If you look at the size of the difference it makes versus just not calculating interest on interest, it's absolutely tiny. And if you look at some of the errors that you can introduce, like, basically, if you introduce an error in a model with circular references, even if you, you know, hit control Z or clean up the error source afterward, every cell in the cycle is now a source of error for the next cell, and so you have to like actually break the circularity, which can, if you haven't set it up for it, and most people don't involve like changing the structure of the model temporarily and then bringing it back, it's a nightmare.
I think circular references are dangerous because they're so easy to do. Most people do not understand the logic of them, like, do they always converge? What makes them converge or not? How do they handle errors? That kind of thing. You know, how do they interact with data tables, with goal seek? It's very easy for people to use them, and they generally don't know all the risks and downsides.
Host: Paul Barnhurst
That makes a lot of sense. And I had someone who used one in a model I had, and I'd always get a circular reference error. And I'm not sure I ever found exactly where that error was, but it was big enough, the numbers were close enough that the model worked and I didn't feel like spending 20 hours to try to chase it down and that's probably not good, but that's an example of what happens when you have it.
So we have just two questions left here and then we'll let you go.
First one is we'd like to ask everybody for kind of a nugget, some kind of learning that they've had around financial modeling that you could share with our audience and kind of that piece of advice, so to speak, or that technical nugget that you think would really help people.
Guest: Diarmuid Early
I don't know if this is the best one, but the one that comes to mind, which was a bit of a revelation for me the first time that I sort of clicked it, which is when you're building a three-statement model, I always used to think that as like I could go through the pieces. So like, okay, here's a bunch of assumptions about revenue, let's build a revenue schedule, here's a bunch of assumptions about costs, let's build a cost schedule. And then, you know, as I'm going along, I'm building up an income statement, a balance sheet, whatever. And the challenge when you do it that way is, by the time you actually put together the balance sheet, if something doesn't balance, there's like 45 places that that error might be coming from.
And so the habit I've got into, and I recommend it for anyone, if you're building a three-statement model from scratch is basically build the plumbing with no schedules at all first, before you touch anything else. So in other words, build an income statement where the revenue is zero, the direct costs are zero, depreciation is zero, the interest is zero and so on, but all the pieces, in other words, you know your gross profit is linked to your revenue and your cost, and your net income is linked and so on. Link it through to the balance sheet, link it through the cash flow statement and then like the whole logic of the three statements is, no matter what your business does the balance sheet should balance and the three statements should flow together.
So, in other words, if you’re a business that has no assets and doesn't do any transactions, congratulations, you just built that model and you know, you can actually even kind of store that template if it's a thing you use regularly. And then if you have a business that sells 100 units in the first year and then doesn't do anything else forever after that and incurs no costs for it and so on, that's still a valid model and it'll still lead your balance sheet to balance.
And so the idea is at every step of the way, as you layer in each additional schedule, your balance sheet should balance at every step. And so then if you're doing some particularly complicated thing, for me, it's usually Deferred tax and you get the sign wrong, you know, one step back, this balanced and now it doesn't balance, so I know exactly where to look. And you know, if it's off by whatever, twice the amount of your Deferred tax assets, like, okay, yeah, I think I know what's going on here. I guess the part to operationalize that is, you know, add in a row of checks where, you know, check if the balance sheet balances and then add in a master check, and I link that to some conditional formatting.
So, basically, as soon as the balance sheet is out of balance, the top bar of every sheet turns red. And that way, you know, if it happens, if I break something in a way that I'm not expecting or, like, sometimes I'll say, okay, well, it's turned red now, but that's because I've added this thing in here, but I haven't yet added it in here, fine, you know, work in progress, but if it turns red when I'm not expecting it, then it's like, whoa, something's going on.
Host: Paul Barnhurst
That's really good advice. I like that. And I can see how having it all linked, you can check every step. Anyone who's dealt with the balance sheet has had a balance sheet that didn't balance, I am convinced of that. I don't think there's a single person out there who's had to always balance, unless they've always just stuck a plug-in, to force it to balance. That's the only way it's ever going to always balance. That's good advice because I know I've spent a few hours trying to figure out why is this not balancing? Is it retained earnings? Is it what sign did I get wrong? You know, where's the issue?
Guest: Diarmuid Early
Oh, yeah, and there's you know, when you think it through there's a lot of different places it can go.
Host: Paul Barnhurst
Yeah. And the more complex the model, the more places.
Guest: Diarmuid Early
I mean, the one thing is when you do it that way, it makes you realize how limited a check the balance sheet balancing is in the sense that, you know, the balance sheet balancing should give you zero confidence that you've forecast revenues correctly because if your revenues are growing twice as fast as they should be, that's still a business scenario and the balance sheet will balance. And if you've depreciated too quickly or too slowly, it'll still balance.
Basically, if any of your assumptions are wrong, the balance sheet will still balance. It's only if something is structurally wrong, but at least it lets you isolate that bit.
Host: Paul Barnhurst
Yeah, I like to say there's really two things to a good model. There's good structure and design and there's good assumptions, right? Something balancing doesn't mean the model's good in and of itself. It means you've built a model that, hey, it balances, good, congratulations. But does that mean it makes sense? No, I don't know. Maybe, maybe not.
Guest: Diarmuid Early
The good thing about, the sort of discipline you build in the competition is when you're able to very quickly and precisely execute on a set of assumptions, it means that you can spend more of your time on, you know, frankly the side of that balance that's more important. If it takes you days and days to build the model and figure out why your balance sheet isn't balancing and develop and iterate and whatever, then you have you know half an hour at the end before you go into the client meeting to say, oh and those assumptions we put in on day one, do those make any sense? Whereas you know if you can build it in a few hours then you know you spend the rest of the week thinking like okay now let's do sensitivity analysis let's take a step back let's figure out if this is the right approach, that kind of thing.
Host: Paul Barnhurst
Make sense to me. I can agree with you on that one. So, now we're at the end of our time here.
So, last question for you. If our audience wants to learn more about you or get in touch with you, what's the best way to do that?
Guest: Diarmuid Early
I mean, if they want to learn from me, then my YouTube channel is the best place for that. I post videos there fairly regularly.
If they want to contact me, then LinkedIn is probably the best way to do that.
Host: Paul Barnhurst
All right, perfect, so Linkedln, if you want to contact, if you want to learn more, check out his YouTube channel, we'll put that in the show notes and if there's anything else you want us to put in the show notes, let us know, but thank you so much for carving out some time this morning, really enjoyed having you on the show, Dem and good luck with, you know, the World Cup this year. We hope to see you at the top of the podium again soon.
Guest: Diarmuid Early
Fingers crossed! Fingers crossed!
Thanks for having me, Paul. It's been a good chat.
Host: Paul Barnhurst
What a great episode with Dim or Diarmuid Early. It's a lot of fun having him on. He's often referred to as the Michael Jordan of modeling with two world championships coming in second a couple of times. He's a great modeler. He's one of the best out there.
And so, you know, after talking with him, I made the decision to go ahead and give it a try. I'm going to do the Financial Modeling World Cup myself. I went ahead and went to fmworldcup.com. I think it was 25 dollars for a ticket to participate. It's going to be at the end of this month in July. I'm doing it with a friend, AJ Patel. He's known as the Excel Ninja. He and I are both going to compete.
And for my session, Andrew, the president of FMWorldCup asked me if I wanted to be part of their streaming. They stream four candidates or four competitors every session live. And so mine will be streamed on the 27th. You can go to the Financial Modeling World Cup YouTube channel to watch that. This episode will come out a couple of days before that competition.
That competition will be streamed live, as I mentioned, from 9 to 11 in the morning on July 27th, on the YouTube channel. And so then I'll share some of that as well. AJ will talk a little bit of our experience and what we learned. You'll see those things on LinkedIn.
So, I thought the next thing I would talk about is, where do the contestants stand at this point? You know, they call it the road to Vegas. Having a little bit of fun. Kind of like the road to the Final Four. The championships will be in Vegas, believe it's December 9th, 7th through the 9th with the finals being on the 9th, and right now where the standings are for the top three are as follows.
So, Lawrence Lau is in first place with 3,654 points. He's won the last three stages. He's got a thousand points. The winner gets a thousand points.
In second place, two points behind is Diarmuid Early. He's done very well in all of his stages, so he's been very consistent at or near the top in every session. He hasn't won any of the four, but he's been near the top every time.So he's in second place.
And then in third place is Andrew Nye with 3,320 points. All three of those are regularly near the top of the leader-board. They're some of the best out there. So it sounds like it's going to come down to the wire, going to be an exciting competition. We're on stage four, with four more stages to go, and obviously that last, the ending will be that championship in Vegas, and I look forward to bringing you that episode and talking to some of our top guests and getting their experiences.
So again, thank you for joining me. And as a reminder, you can earn CPE credit for this podcast. All you have to do is go to earmarkcpe.com, download the app, and answer a few questions and you can earn CPE credit for your continuing education.
If you enjoyed this episode, please share the podcast with your friends as we're trying to increase the podcast and we'd love for everybody who enjoys financial modeling to get the opportunity to listen. So, share it with your friends, subscribe, and leave a review on your podcast platform of choice, Apple, Spotify, Google, or whatever that may be. And thanks again for joining us for this episode because if it was not for you, the audience, I wouldn't be able to do this. And I love bringing you these episodes.
So again, thank you for joining me.
Financial Modelers Corner was brought to you by Financial Modeling Institute. Visit FMI at www.fminstitute.com/podcast and use code FMC15 to save 15% when you enroll in one of their accreditations today.