Innovative Approaches to Financial Modeling And Risk Management With Raphael Benhamou

Tell me that horror story. The worst financial model you've ever seen. I know you have one.

There's the worst in terms of actual modeling quantity— quality, I should say, or lack of actual knowledge. The model itself was a monster, a beast. There were lots of mistakes in it. But the main issue was actually how that impacted the business and how that affected decision-making further down the line. Being able to derive valuable insights from the models I build and the fact that I built them and understand the model on a very granular level means that I can understand all the moving parts and mechanics of a business. And I can say, That's something that you need to focus on. That's too high level. Let's go and work out how to actually achieve that pragmatically, etc.

With Excel, almost too freeform, right? It's the most flexible tool out there. I get asked all the time, well, will my planning tool be as flexible as Excel? And I always say no.

There's always different levels of granularity. And sometimes you just need a 3-statement model. Sometimes you want 3-statement plus HR. Sometimes you want 3-statement plus HR plus revenue.

One story, one chart in Excel, one table in Excel can do the work of Welcome to Financial Modelers Corner, where we discuss the art and science of financial modeling with your host, Paul Barnhurst. Financial Modelers Corner is sponsored by Financial Modeling Institute. Welcome to Financial Modelers Corner. I am your host, Paul Barnhurst. This is a podcast where we talk all about the art and science of financial modeling with distinguished financial modelers from around the globe. The Financial Modelers Corner podcast is brought to you by Financial Modeling Institute. FMI offers the most respected accreditations in financial modeling. I'm excited to welcome on the show Raphael Benhamou. Welcome to the show, Raphael.

Thanks, Bill. It's good to be here.

Really excited to have you. One of the first questions we start off with every guest before we even do introductions is tell me that horror story, the worst financial model you've ever seen. I know you have one.

It depends how you define worst, really.

I'll leave that up to you.

Well, there's the worst in impact in terms of damage done to the company. There's the worst in terms of actual modeling quantity— quality, I should say, or lack of actual knowledge. So let's go with the first one in terms of impact, because I think that's probably the most important and relevant to people. So I was working on a reinvestment model for a global mining company, and they had a model that was tracking 300 different mines, doing discounted cash flows for all of them and doing an industry accumulation. So aggregation of all that information and then doing some decision-making off the back of it. And they hadn't audited the model in 4 or 5 years. There was some political stuff going on between the person who built the model versus the new manager in charge. And when I came in and audited the model to rebuild it into a new software, I identified something like 30+ line errors, mostly cell reference errors. So basically sort of the gold price tied up with the molybdenum production and stuff like that. And it was causing errors of +10% on a multi-billion dollar investment plan. And it was quite painful to show that to the client because they've been using this to form their investment strategy for years. And so as a consultant, There's all the sort of soft side, which is client expectation, client management, communication, getting them to understand what the impact is, and then explaining that to their managers and superiors. So the model itself was a monster, a beast. There were lots of mistakes in it, but the main issue was actually how that impacted the business and how that affected decision-making further down the line. And there's the scale of the model, because we're talking about billions of dollars of investment, is obviously quite significant. So I would say that that was probably One of my worst experiences is also my first experience. So it's a lot of learning very early on in my career.

I bet that was a great learning experience. What was the key takeaway from going through that kind of nightmare, so to speak?

I'd say, I'd say 2 key learning points. One is regularly audit your model and make sure that if you've got big models like this, you have someone on your team that is sort of technologically literate and financially operational literate so they can actually understand what's going on at every line item and reviewing it on a, well, not daily, but on a regular basis. And the other aspect is in terms of communication, making sure that people actually know what they're using and why they're using it and what the implications are if there are mistakes, that people align sort of expectations. They know that the sort of, even from a risk management perspective, I've got a background in operational risk management, is having some kind of checks and balance system in place to prevent this kind of thing from happening and also communicating regularly to avoid these things happening. Sort of going on for a long time without any kind of check.

Makes sense. Basically having a better check process and audit process to review these models.

Exactly. Yeah.

Can you next tell us a little bit about yourself, your background, what you're doing today and how you ended up where you're at?

Going back a little bit, got an undergrad in physics and languages, did a master's in engineering and finance. And for me, that really sort of molded the world that I see in terms of finance strategy and modeling overall, because I learned very early on how to model the microscopic and then go out to the macroscopic and see how small things can impact big things. Actually, ironically, as the case study I just described shows. So being able to understand things on a very detailed level and the engineering degree afterwards helped a lot in understanding the operations and the mechanics of a business. And then combining that with the finance and seeing what the long-term impacts is, has really molded the way that I think about the world and the way I think modeling and business in general. I've gone, I've worked in consulting basically my entire career, apart from 2 years or so in industry, working in a sort of a fintech and then a mining project in fertilizers. But most of my career has been in boutique consulting companies and then at EY in the risk management, data analytics, and regulation compliance and robotic process automation teams. I went independent about 4 and a half years ago, just before COVID Started off as FP&A. Well, actually, I started off as a generalist startup. Business consultant, just helping with business planning, go-to-market, fundraising, getting people into incubators, that kind of stuff. And then sort of migrated from that into more FP&A consulting, helping out with financial modeling. So helping with people with, again, with their fundraising, but modeling aspects, because modeling had been a big part of my career, but it just wasn't sort of what I started off doing when I went independent because of connections. So helping out with modeling, with investment, with bespoke solutions like hiring planning, optimizations around remuneration packages, all that kind of stuff. And eventually migrated again from that into fractional CFO services. So today I do a lot more fractional CFO services. I work on retainer with a couple of clients, sort of essentially doing the modeling aspect. And then afterwards the advisory and insights and strategy afterwards, once the analysis has been done, or at least the first part of it's been done, because I never consider a model to be finished. It's always iterating and sort of being updated. So being able to derive valuable insights from the models that I've built and the fact that I've built them and understand the model on a very granular level means that I can understand all the moving parts and mechanics of a business. And I can say, that's something that you need to focus on. That's too high level. Let's go and work out how to actually achieve that pragmatically, et cetera. So that's more or less where I am today.

Great. Appreciate the background. Just so our audience knows, one of the things we're going to be talking quite a bit about in this episode is Causal. It's one of the tools you're using that goes beyond Excel. So I just want to kind of let everybody know that's going to be a focus here.

So sure.

If you're curious, yeah, a little bit out there.

Go ahead.

Yeah.

Should I give you a bit of context around my connection to Causal then?

Let's get to that in 1 minute. We'll get there. Give me a second. There's a couple other questions I want to ask, but we're going to spend a lot of time on Causal. Raphael's going to share why he's using that tool, how he uses it so people can learn about some other tools. But before we get to that, in your background, there's a question I want to ask because I'm kind of curious. You mentioned how your background really helped you, but you started with physics and languages.

Yes.

In undergrad. What made you become interested in getting a metals and energy finance master's? That's quite a switch.

It was a big switch. It was a very big adjustment at the time. I understood nothing about the mining industry or the energy industry at the time, apart from the general impact it had on society. I guess, so during my physics degree, I did a year abroad in Switzerland as part of my language component, and I was of studying in French. And I was working in— as part of that, I had to work in a lab. I was supposed to do just what's called travaux pratiques, which is just practical laboratory work. But the lab work was actually really boring because it's stuff I'd done in first year in the UK because they've got a different curriculum in mainland Europe. So I signed up to work in a PhD lab on high-temperature superconductors for a year, which really gave me a taste for— and there was a specific goal for this, which was to rather than just doing general theoretical research was to actually have an application, which was to build maglev trains. So magnetically levitated trains for people who don't know about them, basically reducing resistance and moving fast and sort of reducing energy consumption. And it really gave me a taste for practical applications of physics. So once I finished my undergrad in physics, which happened to be in 2009 where there was no, you know, the economic crash, I started looking at sort of masters in different more engineering-related roles as a master's. I applied to a couple in the environmental space and in space because those are 2 areas that I had a lot of interest in from my undergrad in physics. Both of those were complete. And as a backup, I kind of applied to this Metals and Energy Finance course mostly because I thought it'd be interesting to learn about, because I sort of knew about the applications, as I said, to industry and to infrastructure around the world. And I thought that it had a lot of courses on engineering, accounting, finance, strategy. languages and stuff like that. I thought it'd be a really interesting, well-rounded course to do, even if I didn't want to go into that specific industry later. I'd learn a lot from it and it'd gimme a lot of skills for later. And it turned out it really did actually. As it turns out, I ended up loving the mining industry. I thought it was a really interesting one. And another sidebar on that is that I got very, very passionate about environmental issues, food security and stuff like that. And now less so, but especially when I was still studying, like 80% of the world's energy comes from coal. conventional energy sources, I can make a small difference on the renewable side and it won't really make a difference. But if I can make even a smaller change on the conventional energy side or mining side, it could have a much larger impact in my lifetime. So I thought, okay, it's a way to potentially help the environment, help the world later on from the other side. So that was kind of my motivation. It turns out that I actually really enjoy the mining industry as a sort of an area of professional interest. It's a very varied industry. Every mine site, every geological site is different. So comes up with, there are lots of different problems and challenges that you have to resolve. There are obviously a lot of sort of more current issues around sort of a skilled workforce, around environmental impact, around really doing, using renewable energy, smart technology, having automated machinery. So there's lots of moving parts to it that are really, really interesting and sort of having the opportunity to work on sort of a mining project that was going to produce 20% of the world's fertilizer and feed, essentially feed the world, have a material difference in feeding the world. It's pretty exciting to do. So there's some definite upsides to working in a less than popular industry.

I can see you've given it a lot of thought. You really enjoyed it. It sounds like it's had a lot of variation. One other question I'm curious, 'cause I know you've studied, you know, in Europe, you've lived in Europe, you've lived in Israel. I know you speak a number of languages. You studied languages. How many languages do you speak fluently?

Sadly, only 2 now, French and English. I used to be fluent in Spanish, but lack of practice has meant that I can speak it, but a lot of— I have a block in my brain sometimes. Spanish and Hebrew comes out, or vice versa. I used to think that I was very skilled at languages because I picked up all the romantic languages very easily— Italian, Portuguese, French, Spanish, and so on. But once I moved here and started learning Hebrew, I realized that actually I've got very good at picking up languages that connected to each other. But once you take a language that's completely different, like Hebrew, it all falls apart. So I'm fluent in French and English. I'm more than passable in Spanish. I understand Italian, I understand Portuguese, and I get by in Hebrew.

Got it. So there's a number in there that you can at least carry on a basic conversation.

Yeah, I'd say 4 or 5, I can have a basic conversation.

That's more than I can do. So good for you. It sounds like, you know, you've spent a lot of your career building models. What attracted you to building models? What is it you enjoy about putting a model together?

I've always liked puzzles. I've always liked strategy. I was very much into sort of, well, Risk, yes, and more chess when I was younger, thinking 5 steps ahead. My father and my grandfather both taught me chess when I was about 4 years old. So it's been something, some of a family activity for years. I like I mean, it comes actually back to physics, breaking things down into their core component parts and then building back up again and understanding what the overall behavior trends are and how something, a small change here can have a massive impact there. It really talks to sort of physics on a mechanical level, but also chaos theory, stuff like that. Sort of how a tiny little change here that you intellectually don't think will have a connection to something over there. But if you, but I mean, you know, maybe unconsciously that there's a connection. But you don't make the mental leap because there's so many different moving parts along the way. And building the models to understand that mathematically, creating logic to decouple that and break it down into component parts of things that I find really interesting. And then applying that to business is where I went from enjoying it from a physics perspective to enjoying it from a finance strategy perspective. I like application rather than theory. Well, I like application more than theory, I should say. I always, I love physics, but I prefer using it in real-world environments.

No, I get it. You like to see the real-world application. That makes a lot of sense. The idea of being able to break it down, build it back up, understand the different parts. There's definitely correlation with modeling and physics in that sense of how you're breaking things down. I know you started modeling obviously with Microsoft Excel. I know you've done Excel—

A little bit, but not really.

What did you mostly start with? I know it wasn't Causal, which we'll talk about. No, Causal wasn't around.

What did you model with earlier? Ironically, it was something very similar to Cozy, which is how I ended up enjoying Cozy so much. So, in my first job, the founders, so I was working in a boutique consulting company. There were 2, basically 2 directors and me as the first hire, or the first consulting hire they had in HR, basically. They were essentially building solutions in a software for the mining industry developed by mining, as a mining software tool in South Africa. The director had a connection to the founder of that company. They'd worked together in a previous project or something like that.

I can't remember.

And so I started my professional career sort of working obviously in Excel, but also in parallel on this other software that's called Corent. And it's kind of like Causal today, but more programming in the language. It's more math. The language is more akin to MATLAB, which I used extensively obviously during my physics degree. So it was very familiar to me and it had a lot of the structure of it was very similar to Causal today. It's object orientated rather than cell references you create. A single variable and then create dimensions and it brought up categories. The naming nomenclature is slightly different. You can create scenarios very easily. They had a sort of integrated dashboard. And that's what we were using in my first job. And I used that every day for 2 and a half years while I was in that first job. And I actually felt like it really, it massively influenced the way that I then later started building models in Excel and Google Sheets because I started thinking a lot more from an object variable perspective rather than a line, line, line perspective that a lot of people use Excel in. It helped me to structure models a lot better because you can create hierarchies in the model, you can create templates and you can say, okay, this is, I'm going to think about building a general template that I can use, like with the first project I mentioned, so 300 times. So creating sort of logic that is easily repeatable and stuff like that. And it helped me to really think about how I can do that on an Excel perspective as well. So creating templatized logic very easily using obviously dollar signs to sort of block things and copy things down easily. Stuff that seems obvious to a sort of a competent Excel user, but to a beginner, not so much. So it really gave me a great grounding in best, I feel, best practices for modeling across all modeling sort of platforms, especially Excel. 'Cause one of my big nitpicks with Excel and Google Sheets is that it's too freeform. It's great if you're a competent or good user, but if you're a bad user of Excel and you don't know you're a bad user of Excel, you can make lots of mistakes really, really easily. And a lot of the softwares that I've ended up using in sort of since kind of take Excel and put limits around it to prevent that kind of behavior from happening, which is one of the reasons I think that they're really useful because they enable people who are not so sort of competent, if they're not using on a daily basis, it allows them to use them and still get insights, which is fundamentally the important point of modeling.

So there's a lot to unpack there. So yeah, that's helpful then.

It's a very long answer.

Started with mining for a couple years using a mining software. You use MATLAB obviously in college. You had that programming background. You came from the object kind of variable approach versus the freeform spreadsheet nature. And one thing you said there, right, that I really like is with Excel, almost too freeform, right? It's the most flexible tool out there. I get asked all the time, well, will my planning tool be as flexible as Excel? And I always say, no, you're trading off some flexibility for structure. That's not a bad thing. You just have to recognize that. You know, because if you're talking large planning tools, you need the database, you need all those other things. And it sounds like many of the planning tools you've used have traded off some of that flexibility for the structure and rules to make it easier to understand, to make it less likely to make mistakes. That has worked better for you. Is that kind of a fair way to kind of think about that?

Yeah. And it's not necessarily because I'm making the mistakes so much as the people that I'm working with make mistakes because I'm a consultant. I've worked as a consultant basically my entire career. So my daily activity is seeing other people's bad models. I know that most people don't spend their days working in models in Excel or Google Sheets. And so any tool that can enable them to still have the insights and the ability to carry out the analysis that they need to do to be able to make informed decisions, but reduces the risk of having, of creating, inserting mistakes into them is a win for me because at the end of the day, as long as you're not trading on core functionality that allows you to do the analysis, limiting your freedom is not a bad thing, I think, in that respect. And I'm talking just about modeling, by the way. I'm not talking about society. Just for instance, like that.

We won't go there. We'll stay away from that. Yeah. So I know about 4 years ago you came across Causal. Were you looking for a tool to kind of replace Excel with your— as you started your business? Or walk through how that happened, how you came across Causal. And maybe a little bit of why you switched to using them. I believe you use them, what, about 95% of the time now?

Is that—

walk through that journey.

It was very spontaneous. So, I went independent just before COVID I sort of signed on for a contract with a startup that needed help on the business, essentially business consulting, business analysis perspective. And when COVID happened, they dropped me obviously, because they're in some crisis, in a crisis and worrying about sort of cash and all that kind of stuff. And all my leads obviously for, so everything that I was in my pipeline disappeared as well. So I basically spent about 6, 7 months just trying to reinvent myself, work out how I was gonna communicate myself in an online world to the world about what my services were. And I really struggled with it because I've got a very wide range of experiences. I've worked in everything from risk management to business optimization to M&A and everything in between. So I couldn't just say I'm a business analyst or business consultant. No one understood what that meant. So I was having these calls with people. I was on this platform called, well, I am actually still on this platform called Lunch Club. And I was meeting people from all around the world because as it went from in-person meetings to online because of COVID And I was connected up with a guy in an investment company in the US. And I was telling him about my background in modeling and strategy and finance. And he tells me he just heard of this software that someone that he knew had invested in or started using. I can't remember exactly what he said and that I should try it out. I was like, okay, sounds interesting. Why not? I've got nothing better to do at the moment. I didn't have any clients. I checked it out, happened to be Causal. I tried out the free trial for a week. Within about 2 hours I was getting comfortable with it because it was basically the same but a more user-friendly version of what I used at the beginning of my career, Qurrent. I started building out, as it happened, I was taking a mortgage brokering course. I built my mortgage model on there as well, just for a bit of fun, just to see what I could do with it. I love the platform. Like really quickly I could see the potential of it. I could see where it was going, sort of functionality it had already and where they could potentially go with it in terms of added functionality capabilities. Because I'd seen this other software 15 years earlier that was very very similar and I'd sort of been part of their development process. So a little bit of background on that is they were obviously constantly iterating the software and when we were pushing the bar with some of our clients, we would go back to them and say, hey, can you add this functionality? Can you put this in? Can you do that? And so I'd seen, I'd been part of that product development process, that company. And so I could see where Causal could go potentially because I could see how they'd structured the software. So I reached out to Lukas, who's one of the co-founders. At the time it was just Lukas Thamer and his brother who was working on it part-time. So they were, I think, still bootstrapped at that point. Just the 3 of them. I reached out, I said, look, I love your product. I think it's great. I really see the potential and I've got a background in consulting, financial modeling and so on. Is there anything that we can do to help each other from a business perspective? And he said, look, we're a SaaS platform where 95% of our clients sign up and then say, can someone help us build our models? Because we don't have the time or resources or expertise to build stuff. So if we don't have the time to do it because we're building out the product and we also don't have the expertise because we're software developers more than anything else. So When TemuScale is more product, but if you want to take that on for us, great. So that was basically the start of a really beautiful relationship with them. They would get people, they would do their marketing, they'd sign up clients, and then when someone asked, can I get some help with that? They would then send it to me. Eventually they even sort of started saying, so Raphael can take the first call with them and then sell them on Causal. So I was also sort of showing them the added value of Causal as well as selling them my consulting services. But yeah, that's how it kicked off really. And it went on like that for probably about 18, 24 months maybe before they got their first fundraise. And then afterwards they started hiring people internally to do customer onboarding. And I started focusing more on my existing clients and people who'd come back to me for repeat work and doing my own marketing. But at the beginning it was great 'cause I had sort of 15, 20 new clients a month without ever having to do any marketing. It was fantastic.

Yeah. So it sounds like it kind of was, Born out of COVID and trying to find something and having free time leading to opportunity to build on that.

There's a lot of grafting. There's a lot of sort of networking calls and sort of I mean there's a lot of stuff obviously that I tried before something happened with Corsair. I mean I even yeah I mean I was speaking to contacts in Australian mining companies and in South America and all kinds of different things just trying to see if anything. could work. And because it was COVID, nothing was working because everyone was panicking. So it was really sort of a serendipitous moment and it really worked out, I think, very well for both of us as a very symbiotic relationship because they really needed someone to onboard clients at that point. Because otherwise I have a feeling that a lot of people would've signed up and then lost interest within a week because they wouldn't have been able to use it. So it really helped them to lock in their early clients and it helped me obviously to build up my consulting practice.

Yeah. So let's talk a little bit about Causlo. I want to spend maybe about 10 minutes kind of diving a little further into that. So first question is, what kind of models are you primarily building in Causlo? Are they 3-statement, M&A, operational, a little bit of everything? What are you mostly doing?

A little bit of everything. So as a fractional CFO, obviously there's a 3-statement model, which is core to everything, but because I've got a background in, in engineering and operations and to consulting, from consulting perspective, a lot of operations. I've always fundamentally believed that a business model that doesn't include operations isn't an effective tool. Because if you try to build everything from essentially the revenue line on your P&L and you don't have anything going on in the background that's saying, okay, how are you driving revenue? How are your costs driving revenue? How is marketing expense driving your revenues? All that kind of stuff. or hiring decision-making, you're essentially fighting with one hand behind your back. So I've always, I mean, even before as a fractional CFO, long ago, I basically, probably since my first job, to be honest, I've always built models from operations and finance. So breaking down a model into what the revenue drivers are, how the expenses work, what your working capital is, what your HR planning is, and driving the 3-statement models through the With that, then afterwards adding on valuation models, sometimes having a blank. It's a problem with sometimes with being multilingual is you have a blank in a language, but sort of having a cap table in there. So having that, how that affects the balance sheet. So I try to build a unified business model, if you will, and add granularity where it's relevant and juggling the return on investment and building that with the insights and advice that it's going to give. So sometimes it's just not relevant. So one of my old clients has just come back to me yesterday and he asked me to build out a high-level model that he wants to go to investors with. I sort of have to then clarify, okay, what do you define by high level? Because then he says, I want to integrate Xero into it. So it's already more complicated than he probably imagined. So there's always different levels of granularity and sometimes you just need a 3-statement model. Sometimes you want 3-statement plus HR. Sometimes you want 3-statement plus HR plus revenue. Sometimes you want a separate marketing model, a salesperson model.

Sure.

It really depends. But generally I try to go for as much as is useful and insightful and relevant to the client, because at the end of the day, a model is only as good as the insights it's giving. And if it's only giving high-level insights that don't actually drive accurate, don't give accurate advice, then there's no point.

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 PODCAST to save 15% when you register. So the next question I want to ask is, with the models you're building, sounds like it's across the spectrum, a little bit of everything. What's maybe— what do you like most about doing it in Causal? Like, what's the main reason you're using Causal versus Excel?

There's a couple of different aspects. As a consultant, again, I need to juggle my preferences as a professional model builder with my activities as an advisor in terms of strategy. And then in terms of the client relationship, how the client is going to onboard it. So one of the big challenges with any model is, is the client actually going to use it? Are the team leaders, if I'm talking to the CEO, the decision maker, and they're saying, we want to build this model, we want to develop KPIs, OKRs, et cetera, for the team. If the team doesn't want to onboard them and actually accept them, then again, there's no point in the project. So all of that is being juggled in terms of the decision on what platform to use. Sometimes they're open to using a new software like Causal, sometimes they're not. Sometimes you have to create a hybrid solution. In terms of what I really like in Causal, I would say that from my perspective as a model builder, I love the fact that it's, the language is easy. So rather than use Excel references, use the names of the variables. which makes it much easier for me to audit and sense check things, but also easier for me to explain things to people because they can, I can show them model, I can show them the equations. I can say, no, that doesn't quite seem right. I think you've missed out parts of the expenses or whatever, or you've, yeah, you've captured everything that needs to be done. Then there's the aspect of how it communicates to the client and how the client is going to use it in terms of the fact that Causal integrates with a lot of different systems. So I can use it, connect it to QuickBooks, to Xero, obviously from a bookkeeping and accounting perspective. I can connect it to a lot of HR systems, so BambooHR and a whole load of others, CRMs, all that kind of stuff. So you can easily get connections to other softwares and get that live information to be able to show value in terms of comparing budgets versus actuals and all those kinds of insights. You've got dynamic dashboarding, which is, I think, great. I mean, you can do some things with Excel, but it's a lot easier and faster and it saves a lot of time for the client in terms of chargeable hours when I can create one chart instantly that can simultaneously allow me to see things on a monthly, quarterly, and annual level while also seeing whether it's the delta change between previous months, previous quarters, previous years, and filtering out between the different categories and filtering out different things on the bottom or seeing an aggregate version. I mean, one story, one chart in Excel, one table in Excel can do the work of an entire dashboard on Excel. And I think that that from just from a time perspective is amazing. But then actually showing that to a client and seeing how they can then see things that even I didn't think about because they're coming from a slightly different perspective. It's really, really exciting. Yeah, just there's a lot of functionality in it that's really useful. Like I come from also a risk background, so Causal is, as far as I know, the only easy accessible— well, Corinne actually does it as well, but they're an enterprise tool. It's the only tool that does probability distribution modeling really, really, really well. So in Excel you can do it basically with Glass, what's it called? Glass Ball or something. I think it's been a while since—

Crystal Ball or whatever.

Crystal Ball. That's it.

Yeah.

Crystal Ball. So Crystal Ball, I don't know if they've developed it in the last couple years, but it used to be that you could do one input to one output. So you're just really seeing the range of the distribution rather than the behavior, cuz the behavior's gonna be the same every time. In Causal, you can build uncertainties and probability distributions on any input assumption that you want, and then build out your logic with sort of your max, your min, your distributions, your sort of buckets or whatever you want, and then see what the concatenated effect is on your outputs. And that from a risk perspective, from being someone who has been a strong proponent of quantitative risk analysis for the better part of a decade, is gold. You can't get any better than that. And just on that perspective, just explaining that to clients, being able to show them that just the way they change their terms with clients, whether they get paid in 30 days to 90 days, or whether it takes them 2 months or 6 months or 9 months to close a client in their pipeline, and how that can impact their cash flow and all those kinds of questions. But using risk and showing, okay, this is where you risk going under $0 in your bank account and so on, is really, really important and really, really insightful. And again, providing those kinds of insights and value is what I think a model should be about rather than the tool being the tool for the sake of the tool, essentially.

A couple things I want to kind of recap there, if I, if I heard it right. First, you like the fact that the variables, right? You're referencing variables, not cells. It's not the traditional 2D spreadsheet. This is a multidimensional modeling tool where you can have the multiple variables listed.

Yeah. But it's also the language. So you say, rather than saying C56 times VLOOKUP, whatever you're saying, for example, revenue is production times price. You're literally writing the name of the variable every time.

Correct. Now you're doing natural language variable. Like if you use the table in Excel, you're referencing those column names. Similar to like a variable versus referencing C2 times B2. Yeah, totally get it. That's a common thing you'll see with a lot of planning tools.

Sure.

Is they try to use those variables because it is more natural if you're learning from the beginning. Most people are used to B2 times C2, but if you're teaching somebody, revenue times quantity makes a lot more sense than B2 times C2.

Exactly.

So that, that natural language referencing the, the name of the variable. Is one thing. 2, the integration, the way it easily brings in different data sources. And then it sounds like one you really appreciate is the risk modeling. Yeah. Data of the probability distribution. There's some things you can do there easily that you can't do in other tools. And then it's dashboarding, the dynamic dashboarding. Are those kind of summary of some of the key things you really like about modeling?

Yeah, I'd say some of the other ones which are really useful is things like scalability. So Using categories or dimensions, you can very quickly build something that is single-dimensional. So, and then after, add products, add currencies, add scenarios instantly in the home. And unlike Excel where you'd have to go and manually add it to every line, you can literally add it just at the input and then Cores will automatically feed it through the whole model. And that is, again, it's all about saving time and making it easier for people to use things and scale models easily when a business inevitably evolves and changes over time. The sidebar on that is Causal does cohort analysis really well. You can build in— there's a unique cohort category in Causal which you can build in cohort analysis for any subscription SaaS model or anything like that. I've developed my own, actually ironically based on my first job, developed my own formula, a single formula that controls the entire cohort. It can save me hours of work just with that one formula.

Great. Cohort analysis is a good example. Yes, that ability to drop in a new variable and not have to sit there and update the entire model to get it to work. Yeah, I've seen that a lot with different, different tools that take a similar approach to Causal.

Sure.

What's your least favorite thing about the tool?

I would say it's probably the, uh, communication, the interface human to model in the sense that like the, you, it comes back to what I was saying earlier about sort of, uh, onboarding challenges. So this, the, the decision maker in the business might be really interested interested in using Causal. The rest of the team doesn't want to use it, then you have to work some kind of hybrid solution, which is a problem. But then another factor of that is when you're talking to investors, for example, and they want to see something in Excel because that's all they're used to. And so you have to then export it and you lose all the added value of having built it in Causal. So that can be very frustrating because it's more about people's prejudices or biases or comfort zones, which sort of irritates me. It's not so much Causal, but it's so different. in many respects from what other people do. You have to work around it sometimes to get to something that people want to see. And it's something that Causal has actually tried to get closer to. So it used to be that in Causal you'd model with cards. So you'd see just like a card for the variable, which would just show the data for the current time period that you're in. And you'd write, you click on the card, you write the formula, and then you close the card and you just see sort of essentially a tree diagram almost of cards. And they moved from that to a spreadsheet view because people said, We are more comfortable with spreadsheets. And I was a big railer against them doing this because I found the card view to be much more intuitive to actually building things and moving things around and understanding how the architecture worked. But again, it's what people are comfortable with. So it's that fitting the product to what clients need. I just wish that people were a bit more open to change because people are so ingrained, so used to the cell spreadsheet mentality of Excel and Google Sheets that they find it difficult to move away from it. And I find that frustrating sometimes because it's fundamentally Excel was built as a database solution. It wasn't ever meant to be a modeling solution when it was designed. So it's been adapted to that because there was nothing else better in the market. But if there are better tools in the market—

So it sounds like you wish people would be more, I mean, it's not so much a reflection of the tool of what you dislike, but the challenges that comes from Using it to help people and how they got it. Yeah. All right.

Because fundamentally, I think the tool is amazing. I mean, it's got, it's had problems in the past. Any software does when you do updates and there are bugs and problems and stuff like that, but that's part of the sort of a startup's journey. So I'm not going to critique it on that because every software has that. I mean, Excel, even I, how old is it? Like 40, 50 years old now. I break Excel on a semi-regular basis. It's normal. the software breaks down or has bugs and stuff like that. But fundamentally the tool is, for me, leaps ahead of where most platforms are in terms of modeling.

So if somebody wants to start learning Causal, building models, you know, they listen to this and they're interested, how would you recommend they start?

So you can sign up for a free trial. There's a Causal University in the, on the website. So once you've got a model, you can go into the documentation. The Causal team are fantastic about creating short-form videos to explain how different functions work. There's a lot of documentation as well. So you can look at that. They've also got a lot of templates so you can download it or not even download, you apply it to your account. You can get one of the templates, start playing around with it. A lot of them are built for businesses and a lot of it is automating now. So there's a fantastic integration with Stripe now where you connect your Stripe account and with an AI, it actually builds out a Stripe analytics tool for you without you even having to do anything. You just sit there and wait for 2 minutes and done. And then you can go and investigate the logic. You can have a look at how it's built the formulae, how it's done the default charts and stuff like that and play around with it. So there's a lot of functionality that's sort of already built in. It's really built, I mean, this is what I was getting back to, is that they've moved away from a software that people need a lot of work to learn how to use or get an external consultant to use, to trying to make it as easy as possible for people to onboard on. And so they've they've made huge steps and efforts in this regard. So there's a lot of documentation, a lot of material out there to learn how to use it.

So it sounds like sign up for a free trial, just start giving it a try.

Yeah, like any SaaS product really. Yeah.

Well, like any tool, you know, Excel or whatever. Get started and start learning. Best way to figure it out.

Yeah. My, generally my recommendation for people is, and this is the way I did it as well as when I started using it, is I just think, Okay. If I was to have a fictitious business, or if I've already got a business and I just wanna start at the top down and say, okay, let's build it very simplistically. Let's just say, okay, how do I generate revenue? Okay. I need an assumption about price. I need an assumption about production. Okay. What am I producing? What are the products? So dig deeper and deeper in that way. And the more you build out your thought process and put paper to pen, metaphorically speaking, the more you'll hit roadblocks in your learning that you'll then have to learn how to do things to actually do it. And that's the way you, well, at least that's the way I learned. I think this is a very effective way of doing it because then you're practicing what you're doing rather than just learning theoretically. If you just read a document, it'll go in, well, at least at the moment, it'll just go in one ear and out the other. I need to actually practice it quickly.

What we're going to do here now is we're going to move to the rapid-fire section. I'm going to run through the questions twice, I think, is what we're going to do. First time, I want you to answer them for Excel. A few of them we won't ask the second time because some of them can apply across both. Some are strictly to Excel. So I'm going to ask all the questions related to if you're modeling in Excel traditional spreadsheet, then I'm going to ask you to give an answer for Causal. The goal here on each of these is, you know, on the Excel side, just a couple seconds. It has to be one or the other, not an it depends. I know it can, it depends for everything. Same with the second. And then at the end, we'll let you elaborate on a few of those that may be really different or that you're passionate about. So just to kind of—

Sure.

Keep it a little bit more of the rapid fire. First time we've done this where we're going through 2 different lists, so I'm thinking in my head, what's the best way to do this? We'll see how this works. So think traditional spreadsheets here. Circular or no circular references?

No circular references.

VBA or no VBA?

Sometimes, yes.

We'll go with the yes. I get it. Excels dynamic arrays in your models, yes or no?

Not so much.

Workbook links, external workbook links?

Definitely not.

Yeah, that's usually what people say. It's like, no, no, never. Named ranges, yes or no?

Selectively, I'd say.

That's usually— everybody hesitates on that one. They're like, please don't go crazy.

Exactly. Yeah.

Do you follow a formal standards board for modeling like FAST or SMART or any of those?

Not really. I've learned from my own experience, really.

And do you think Excel will ever die?

Unfortunately, no. I'm kidding.

Do you think AI will build the models for us?

This is an interesting question. I think it can. The question is whether it should.

Interesting. Yeah, there's definitely a question of that. So you think it's capable. There's a broader question of really, should we allow it to do it?

What's the human— Like all tools.

Yeah.

There's the tool can do something and whether we should be using it for that is another question entirely.

Understood. So sheet cell protection, should you be using it in your models?

Depends on the client. Generally, I prefer not, but it depends on the client.

We'll say generally no, but there is use cases. Got it. All right. Do you believe financial models are the number one corporate decision-making tool?

I don't think they currently are. It depends on the business. I'd say they're not currently, but they should be.

Fair enough. What do you think is then? What do you think most often ends up being?

I think it's less about a physical thing and more about politics, conversations, relationships between people. Yeah, I think unfortunately we're not in as much a data-driven world as we should be yet.

Fair enough. So then what is your lookup function of choice? Do you prefer CHOOSE, VLOOKUP, INDEX MATCH, or XLOOKUP or something else?

INDEX MATCH.

INDEX MATCH.

For me, it's the most versatile, but I know that there are a couple of others that are more versatile now, but I'm used to it and I like it and I don't see the point in changing something that I know works very, very well.

Yeah, it's still incredibly flexible. I like, I'm starting to use INDEX XMATCH because XMATCH gives you a few more options. Options than math.

Right.

Yeah. All right, we're going to run through that list. Obviously not all of them apply, but we're going to ask some of them and you could just say, hey, doesn't apply at all, or whatever that we talk about at the end. So this is for causal. Circular or no circular references?

No, none of them.

VBA, I'm guessing doesn't apply.

Doesn't apply.

This one I forgot to ask for Excel. Horizontal or vertical model? Let's go Excel.

You prefer Excel?

Okay, so Excel. Horizontal or vertical? Let me explain what I mean. When I say horizontal, I'm talking multiple sheets. Vertical is kind of stacking it all. Sometimes people think I mean the time dimension being vertical.

Yeah.

No.

So in Excel, I'd say horizontal. In Causal, up until now, I've only used horizontal because, and there's reasons for this, which are more about functionality and sharing and stuff like that. Causal's recently developed a new functionality that allows vertical modeling or much more effective vertical modeling, which I've only tried out once so far, and so far it's been great. But I don't have enough of an opinion either way at the moment. I mean, to be honest, both of them work more or less the same way in Excel, in, in Quarto, because you, because it's object orientated, you create objects for blocks of logic either way. So whether it's across models or within sections of the same model, it doesn't really make a difference.

Yeah, because of the object nature. Yeah. I'm assuming dynamic arrays don't apply here.

I mean, they're automatic in a sense. Yeah, because you create a variable and you've dynamically got everything inside it, and you either refer sort of like to like or a section of it or an aggregate or whatever. So it's almost like they're built in naturally.

Named ranges, you're naming the variables.

So yeah, yeah, it's not really— it's again, it's automatic.

Again, I don't know that the formal standards have generally been designed for a spreadsheet, so they wouldn't really apply.

Yeah, but again, actually, like I was saying earlier on in the conversation, The way that I've learned how to build models has been heavily influenced by object variable structures from the Brent when I was, when I was in my first job. So I'd say I've learned how to build models even in Excel using an object attribute approach.

Sure.

If that makes sense.

Yeah, no, I get what you're saying. And then you don't have lookup functions? No.

Again, redundant. It's automatic as part of the software.

Way of having the database and everything and the way it works. All right. So I think that covered our rapid-fire section. That was kind of a unique hybrid there. So hopefully people enjoy that. One last question I'll ask kind of for this section, and then we're just going to wrap up. What's your favorite shortcut in Excel?

I'd probably say F4. That's high on my list.

Top 10 for everybody, for sure.

Yeah.

Alrighty. So if you could offer one piece of advice to our audience to be a better financial modeler, What would that advice be?

One is difficult. I would say, so my perspective on this is what I was saying earlier about connecting operations to finance. I think to be a good financial modeler, you need to have a good understanding of what's going on in the business as well, not just at the finance level. If you want to be able to provide insights, be able to provide strategic insights, value forecasts that actually make sense, you have to be able to connect it effectively and accurately to your model, so to your operations. So To give an example, I've seen, I can't even tell you how many times I've seen models where you've got the first line is revenue and it says, okay, the revenue in the first month is X and we are going to grow revenue by Y% every month. And that's how we're building our model and we are gonna reach this goal at the end. That doesn't really tell the story because if you then take that as a decision maker and say to your team, okay, we need to increase revenues by month on month by this amount over the next couple weeks to reach our goal of this revenue by the end of the year or this profit by the end of the year, they're gonna say, okay, how do we do that? So if you can actually give them some insights about this by actually building an operational model that actually says, okay, we're going to increase revenues by focusing on this one product line that's got a higher ROI than anything else, or we're going to increase our marketing spend to drive it, or we're going to optimize our marketing spend because at the moment our business is seasonal, we'd be spreading money too much across the low seasons or Anything like that. We're gonna hire a new salesperson. We're going to be a bit more lean in this respect. We're going to push out this product 3 months earlier, whatever that is. That's what should be driving your revenue growth. And the more you do that, the more you build interrelations between what's on your 3-segment model with the rest of the model, the more you're building a tool that will help you connect strategic insights to operational goals, which will at the end of the day really help the businesses because you'll be able to say, Okay, my goal is to reach $10 million in ARR by December 2025. What do I literally have to do to get there? Rather than just saying, that's my goal, you go out and work out how to do it. And that from a management perspective also speaks volumes because it means that you're actually interested in the business and you're not just sort of top-down telling people what to do and not really caring about how they actually go about doing it. Again, it's that interaction between the business model, the values it's providing, and the management and the use of it in business. Is really, really important. And I think that to be a good and effective business modeler, that you need to really internalize that and understand that. You can be great at auditing, you can check all the numbers, you can build, be an expert at the financial statement and creating your 3-statement models. But if you don't know how to actually use that to derive insights and give advice or to give people the information that they need to make the decisions, it's just gonna be a tool on the shelf that's never gonna be used. It's gonna be used once when you built it, and then afterwards it's gonna be forgotten. And I think the model should be a living, breathing part of a business in the decision-making space. They should be reviewed on a regular basis. They need to be checked and sort of see what the KPIs are and everything. And obviously that's what the ideal is for every business, but most businesses don't do that. So I think that that aspect of it, the softer side of it is very important and it's an often overlooked side.

So understand the operations of the business. I'll boil that down.

Yeah.

Sorry.

I prefer to give context because otherwise it's very like Tagline and not actually giving any insights.

And actually, no, you definitely— I get what you're saying. You gave some great examples there. And just, you know, I, I fully agree. I hear it all the time for FP&A professionals. What's the number one thing? It's like, learn the business. Don't spend all day in a spreadsheet. That applies to a model as well. So great advice. So if our audience wants to learn more about you or what you do, the services you offer, what's the best way for them to, to learn about you?

So they can look me up on my LinkedIn, Raphael Benhamou, R-A-P-H-A-E-L Benhamou, B-E-N-H-A-M-O-U. I've also got my website, benhamouconsulting.com. Also on Instagram, fractional.cfo, I think. I need to check that one. I'll send you that one afterwards. Yeah, I'm not really active on that one, to be honest. I did it as a bit of a lark, but mostly LinkedIn or email. You can email me at raphael@benhamouconsulting.com as well.

Well, thank you, Raphael Benhamou. Really appreciate you joining us today and getting the opportunity to chat with you and We'll put that information in the show notes. If you want to share anything around Causal, feel free to send anything you want me to put in the show notes. We'll definitely put a link to the software as well.

I thought about this actually. On my website, actually, Causal, you can embed models onto websites. I've put 2 demos of Causal or templates really of Causal on my website. If you go on my website, there's a section called demos, I think. You can go in and have a look and there's 2 models on a sort of a subscription-based modeling bit for subscription-based modeling and also for sort of CRM, sort of big client contract kind of transactional modeling approach as well. So you can see what Causal looks like, you can play around with it, you can fiddle around with the assumptions and see how it impacts the dashboard all without actually having to learn how to use Causal. So it's quite cool.

Great. Well, we'll definitely put a link out for that. Appreciate you joining us and I hope our audience enjoys this conversation as much as I did. I've— I'm familiar with Causal. I've used a little bit, but it's always fun to see somebody who's using it all the time and, and get their perspective. So thanks for joining us.

No, it's my pleasure. Thanks for reaching out.

Financial Modelers Corner was brought to you by Financial Modeling Institute. Visit FMI at www.fminstitute.com/podcast. And use code podcast to save 15% when you enroll in one of their accreditations today.