The Exchanges

Q What were some of the other success factors that you found from your historical investing?

A We regressed earnings growth against returns. Unsurprisingly, it's very correlated. We were curious about whether organic versus inorganic growth would make a big difference. And startlingly, the R squared on total Earnings growth is almost as strong as organic growth only, and I think the reason for that is that when a company is approving acquisitions and we're obviously in control of that decision, we're typically doing it with a lot of strategic advantages. We're doing it carefully, and it's pretty rare that we've seen a systematic acquisition program with the proper discipline be dilutive, and so that was one surprising finding. Another one would be that we used to think that management stability would be very correlated with investment success, and we looked at changes in C-suite management, and there was really not a ton of correlation there, so I'd say that's been an insight that has caused us to be perhaps a little bit less fearful of making management changes when necessary or strengthening management teams. Another interesting one was looking at the correlation between entry multiple And investment success. And in our case, and we really had the data just to look at the hundred or so transactions in our history as a private equity firm, there was really no correlation. If there was one, it was a very weak negative correlation. So lower multiple investments tended to …

AI assessment note: “We regressed earnings growth against returns. Unsurprisingly, it's very correlated.”

Q So from going to college, everyone else is talking about finance, and you weren't interested in it, to in relatively short order, you found yourself in finance. So what was that decision process?

A Well, I'd say it was a series of accidental events that led me there. Around my junior year, I started to question whether a lifelong career in science or medicine was right for my personality. I like interacting with people. I like learning things interpersonally by interacting with people kind of the way we are now. And I've loved listening to your podcast as a way to learn. And so I started to question what alternatives might exist. And at around the same time, one of my lab partners of all people got a job at McKinsey and company, which is a strategy consulting firm. And he said, look, they think about questions that are In many ways, it's complicated as scientific questions, but they apply them to business problems, and so I was attracted to that, and so I ended up getting a job at McKinsey out of college, and then after my analyst stint at McKinsey, I wanted to move back to Boston. I was in the Los Angeles office, and I followed a guy who was my best friend in the analyst class at McKinsey to this job that he was talking about. He said it's private equity. I had no idea what that meant, but I followed him to this job mainly because it was a job back in Boston, and it seemed like it was going to use similar skill sets. Is what I developed at McKinsey, and that's the job I still have now, 27 and a half years later, and that person is my partner, Brendan White, and we've wor…

AI assessment note: “Well, I'd say it was a series of accidental events that led me there.”

Q What are some of the ways that the modeling approach changes how you think about risk?

A In general, I'd say undertaking this process has made us humble to the fact that probabilities are inherently uncertain in private equity and things happen that impact investment. So if you build a regular five-year LBO model of a business that is recession exposed, for example, what you typically do is you build a base case model that doesn't involve a recession, and you say, well, This is just a base case. There's no recession. And then somebody on the IC would typically say, okay, well, what if there's a recession? And the team goes off and says, okay, well, we came back. Here's a recession model. The recession case produces a 1.3 X multiple of capital. And the reason for that is that the team has the freedom to put the recession arbitrarily at any point in the five year time period. And so they'll say, well, we'll put the recession in year two, the company's earnings went down. And then in year three, there's a bounce back and we came back. So we lost a year and it grew a little bit slowly or something. The thing we now do, if a business is severely recession exposed, we have a generic probability of a recession. It's 12% in any given year. We could be wrong about that, but it's better than not putting a probability in. So we say, okay, well, let's just assume that once every eight years, there's a recession. So in each of the first two years, there's a 12% probability of a…

AI assessment note: “has created, I think, a profound shift in how we differentially assess the risk”

Q How did that evolve from the simple heuristic of cheap to how you look at value today?

A First of all, I'd say we always were engaged in a set of very Complex activities to get beyond just understanding the entry multiple of an investment. Now, having said that, the entire private equity industry, for the most part, tends to think about one single metric to at least have a headline sense of what the value of an investment is, and that metric is total enterprise value divided by EBITDA. The reason why that came into place in the first place is that it's supposed to be a proxy for the inverse of The cash yield of owning an asset. So if we were to pause it for a second for simplicity that EBITDA is equal to cash flow, it's not, and that you could in fact dividend all of the EBITDA or cash flow of a business to yourself every year. You can't. Then total enterprise value divided by EBITDA would basically be the inverse of the free cash flow yield of owning that asset. Now, because private equity has evolved so much from the early days where that actually was true, To today where private equity portfolio companies are by and large exited to other buyers, so they're not really being sold on free cash flow yield. There's been a huge evolution in the valuation ranges that are applicable to private equity transactions, and so what we decided to do was to expand our view of what the underlying value of an asset was from the simple snapshot of how much free cash flow is genera…

AI assessment note: “what we decided to do was to expand our view of what the underlying value”

Q Where have you found those types of companies that offer that asymmetric profile?

A One of the areas that we have found a lot of that type of profile in would be companies that have a specialized human capital services offering that is combined with a revenue model that has strong recurrence. As we all know, from 20 10 to 20 20, software as an asset class came into real prominence in our industry. It's really transformed the way capital has gotten allocated in private equity, and the reason for that is the software business model was simply put very misunderstood and underappreciated. In this decade, we think there are classes of companies out there that offer human capital services that are highly specialized, that have pricing power, that have the ability to generate very attractive margins, that are incredibly asset light and very capital efficient. That would the right ownership model generate similar types of returns on capital as software and are fairly misunderstood? Oftentimes these are companies that participate in gigantic industries where there's really relatively low private equity penetration, so a lot of opportunity for private equity to have a focused role in consolidating the industry. And so an example of that would be the USCPA industry. It's a Forty billion dollar industry before accounting for surrounding advisory revenue that typically goes along with having an audit or a tax practice. Within that, tax accountants tend to have very, very s…

AI assessment note: “One of the areas that we have found a lot of that type of profile”

Q What are some of the challenges that you've run into either internally or with portfolio companies of trying to inculcate this thinking?

A Seth Klarman talked about this on your podcast, Andy Duke talks about it, is that you could make the right decision and tail events can happen, and so one of the important things for us is when something happens that's unexpected, think about whether it's bad luck or good luck in spite of the decision making or whether something about our decision making was flawed. We've been working on this particular type of thinking for the better part of a decade. Where we now believe we have enough data points to start backtesting and refining the model. But sometimes we end up in tail scenarios relative to our distribution and just being clinical and then giving our teams permission to say, gosh, something about the way we did the inputs was wrong, or this was a tail event and just being really intentional about that. That's one generic challenge. The other one is just logistical. It's just hard to move an investment operation. Even of our scale, and I don't consider ourselves a huge firm, to go from one mode of modeling to another mode of modeling, and in the beginning it was just taking up so much time in these processes. A final one I'd say is our initial inclination when we built this kind of different way of probabilistic thinking into our process was to overcomplicate it, and we would put too many input distributions in, and we'd get an output distribution that we couldn't look at …

AI assessment note: “That's one generic challenge. The other one is just logistical.”

Q So under Jack Meyer, what was that dynamic like when these internal asset managers all started spinning out at the same time?

A It was fascinating. I mean, it was great to have a front row seat to that evolution of Harvard's endowment. From a timing standpoint, we were right at the leading edge of a part of Harvard's endowment operation, balkanizing into private investment operations. So when I joined, Jack was the boss of every investment team. He would show up at our weekly staff meetings. He was an exceptional investor and had this amazing ability to just see through the risk adjusted return that people were debating across a very wide range of asset classes. So it was an amazing learning experience. And then when we started to spin out, it was our private equity team. There were some public equity teams that spun out. There were some private credit teams that spun out. And so there was a lot of news about why it was happening, whether it was better for the university, not better for the university, but it was an amazing time to be part of that transition.

AI assessment note: “It was fascinating. I mean, it was great to have a front row seat”

Q I'd love to walk through the lens of that manufacturing of decision-making. And at the highest level, how have you thought about the optimal decision-making unit?

A We have a phrase at Charles Bank. It's called mission atomization. We regressed our investment results against all these different metrics, going in multiple, exit multiple, inorganic growth, organic growth. One of the things we analyzed was, gee, you know, what percent of our human workloads go into what type of activity? And as we analyzed it, one of the things we observed is that there's just a lot of workload creep. Five people get into a meeting. There may be a stated mission for that meeting, but the meeting will evolve into something else. We'll start talking about things that weren't really related to that initial Mission. And so atomizing workloads where we say, look, X number of people are going to do some amount of work, but the goal of that activity is a tangible thing that we're going to try to accomplish. We now sit down in meetings and we say, okay, what's the prize available in this meeting? We try to be super explicit about it. And as we zoom out to how we try to systematize that we operate in sector teams at Charles bank and within each sector team, there are stage gates of work. After which people have to check back in. So a stage one workload is one where a couple of people can spend a certain number of hours exploring an idea. But after those certain number of hours are done, they have to come back to a group and say, gee, you know, we did this work. Here w…

AI assessment note: “We have a phrase at Charles Bank. It's called mission atomization.”

Q portfolio accounting to reporting to reconciliation, trading, compliance, and more. In the AI era, asset and wealth management firms moving to Ridgeline gain a decided advantage. That's why customers call it miraculous, game-changing, and an awakening. If that's not how you would describe your investment management tech, request a demo at ridgeline.ai. And now back to the show. How did using the model inform where you search for investment opportunities?

A The model that determines our pricing tolerance or valuation tolerance of an investment, that exercise happens at the final investment committee meeting after we've been working on something for months. What's been fascinating to see is how the change in that end modeling process has filtered upstream to influencing how people choose to spend their time. One simple thing that we've noticed, and this has been one of the most exciting changes at our firm over the last 10 years, is even way before the Fan of Outcomes model is built, when teams are out there sourcing and looking for investments, the language that they use to think about and filter investments involves the Fan of Outcomes type thinking. So we will often hear somebody say, we found this idea through our research, It's a company that does X, Y, Z type of business. And when we build the fan, we think it's going to have a really attractive asymmetric shape. Now that's a very interesting thing that's happening. It's a tool that we are using at a final stage of an investment process, really influencing the way somebody is thinking about potential investments at the very early stages. So as that got amplified, what we have noticed is that there's been a shift away from filtering investments using availability. There's always a part of somebody's brain when they're looking at investments around, how do I make sure that I do…

AI assessment note: “how the change in that end modeling process has filtered upstream to influencing”

Q Once you make an investment, how do you apply that thinking towards the toolkit that you're going to use to help improve the operations as an owner?

A That's one important and profound improvement that we're looking to make is to really use the modeling tool as a cognitive motivator for high urgency, higher quality portfolio management. Because we have Focus our model on a two-year period, because our rallying cry now is, unless we're winning out of the gates, we're losing, because that's just a simple way to articulate the outcome of all this regression analysis we've done, because the underwriting process of generating this fan of outcomes model involves such specific inputs around what we intend to do to improve the business. The other thing I would say is, We consider ourselves producers of decisions. We're professional decision makers. It's not just a final investment decision. It's really every decision that we're making that we think of as part of that process. And so in portfolio management, we are using probabilistic thinking to try and improve the quality of our decisions. Should we invest in a new branch office? Well, what does the fan of outcomes look like in terms of upside EBITDA contribution relative to risks? Of that decision. We don't have to build a model to think about that, but we can use that framework to discuss that decision more intelligently with our management partners. What does a fan of outcomes look like on making two or three critical hires in the go-to-market organization? Well, the downside, if…

AI assessment note: “in portfolio management, we are using probabilistic thinking to try and improve the quality”

Q As you've increasingly applied this to your decision making, how do you get a sense of whether it's working?

A Great question. Well, the number one report card for us is the performance of our portfolio and portfolio construction, and we've been in business for over 25 years as Charles Bank. I think we've generated a very consistent track record in North American private equity, and so if we look at our portfolio construction, one of the things we're seeing is that over the last, I'd say, Six or so vintage years, we are starting to see a higher concentration of investments that have real breakout potential. We don't think that's accidental. We think it's because we're being much more focused on all the things I talked about by integrating our value creation planning into the underwriting by going to work with a maniacal level of focus even before we close the investment on accomplishing that two-year enterprise value growth. So this non-accidental concentration of more breakout performance into our portfolio is something we're seeing We've been working on this tool for just under a decade now. Private equity is, it's a little bit like watching paint dry. Things don't happen very quickly, and so we're now just starting to be able to look at the different vintages of investments and see this type of effect.

AI assessment note: “the number one report card for us is the performance of our portfolio”

Q How do you use that decision process to consider your business strategy at Charles Bank?

A We've just simplified our thinking around what do we do to which decisions offer the most asymmetric upside to our firm and muted downside. When we do that, unsurprisingly, it's back to the basics, making sure that we are being best in class human capital managers. At the end of the day, we're just like every other professional services business that we like investing in, our greatest asset is our people. And so just thinking through The upside of being a great human capital manager relative to the downside of having attrition of star performers. So that leads to a whole set of investments that we've made in terms of just being much more disciplined human capital managers. We have a in-house talent management function, which is quite robust to help us with that. Really thinking through not getting attracted to shiny new objects. If you think about expansion into new territories, expansion into new business lines, those all have risks that can be Understated if you're just simply making decisions because you're growth oriented, but just being really judicious about assessing risk. Anything new has to, by definition, have a high level of generic risk. So I think it's just made us generally more prudent decision makers, and I'd say the main effect of it has been for us to really stick to our basic knitting around just being great people managers, sticking to industries and compani…

AI assessment note: “simplified our thinking around what do we do to which decisions offer the most asymmetric upside”

Q What did you take away from this kind of movement at age eight?

A If I had to pick one profound influence, I would say it's the fact that frameworks are something that I consider to be relative. So I think if you grow up in one culture, the educational system that you grow up in, the language that you use, the customs with which you interact with people, it becomes a given and you don't question those things. But because I was displaced from one system at a pretty young age and had to learn A pretty different system. I mean, Korea was somewhat westernized at the time, but in many ways it was not. And the way they teach math is very different. The way they teach history is very different. The way they teach language is very different. There's a lot more rote input of information. And so understanding that the way an entire society does things doesn't necessarily mean it's the only way to do things. That sense of relativism where there may be a bunch of different ways to systematically approach something. I think I started to learn that at a pretty early age.

AI assessment note: “I would say it's the fact that frameworks are something that I consider to be relative.”

Q What concerns you about the industry going forward?

A Because of the way the leverage markets work, and certainly now an enormous amount of capital is getting allocated to private credit, and a lot of it is going into the hands of a few very large players. Those types of excesses, just because of the organic sequencing of psychology in our industry, can create valuation bubbles, and so I think that's a generic risk. We certainly saw one in the 21 time frame. As a firm, because of our long history, we've lived through many of these bubbles, and a lot of them don't end well. We think a lot of the leverage excesses of 21 are likely to result in first liability management exercises, LMEs, that will then involve some amount of restructuring over the next few years or so. Those are generic risks to our industry. Because of our fundamental research-driven model, we tend to look to find opportunity in that type of chaos, so we're getting ready for that.

AI assessment note: “Those types of excesses, just because of the organic sequencing of psychology in our industry, can create valuation bubbles”

Q When you've run the regression analysis over all the different aspects that go into your investments and investment success, how did that translate into how you started to focus the firm?

A If you go back a decade or so, our primary modeling tool for assessing the attractiveness of an investment was a five-year LBO model. Now, even back then, we recognized that the five-year LBO model is highly flawed. Let's just say that you and I could review every five-year LBO model that got approved at an IC five years ago. Let's just say we could look at that today for all North American buyouts, and let's say we could compare The dispersion of the outcome of those models with the actual outcome of those companies that receive the investment. I am pretty sure what we would find is that the actual outcome dispersion would be way greater than the five-year LBO models that got those investments approved. I'm also fairly certain that we would find that the base case outcome of those five-year LBO models would be something like two and a half to three times multiple of invested capital. And then there would be some scenarios for all these firms that approve these models, and some of the scenarios would be conservative cases, and there would be some upside cases, but we know that the actual dispersion would be way wider. So the question for us is, well, we're using a highly flawed tool that actually promotes all kinds of predictable human biases. There's anchoring bias, there's familiarity bias, to make a decision that is inherently very difficult to make. Because what we're reall…

AI assessment note: “we tried to develop a tool or a modeling regime that would be closer”

Q How has the fan of outcomes impacted your investment decisions?

A The model is just a tool for us to think better and work better together. The model doesn't give us any answers that we can't generate as human beings. We just think it makes that process more efficient, and so if we were to just crack apart the different benefits of this approach, the first one is talking about two years versus five years. When we ran our regression analysis, our ability to grow the pre-tax earnings of a company within the first two years of our ownership was highly correlated and very predictive of ultimate investment success, no matter how long the hold period was. Now, that's a useful insight, As opposed to some less useful insights such as, gosh, our investment success is really correlated with management quality. You don't really know management quality when you first come to an investment. You actually develop your view of their quality depending on how that investment's going, so that's sort of a spurious correlation. The fact that two-year EBITDA performance is actually predictive of ultimate investment success is a very useful insight because we can think about two years way better than we can think about five years, especially if we force our teams to crack The first two years into year one and year two. And one of the things that we witnessed happening was that there is a much higher degree of accountability to the team's modeling. So if you were my…

AI assessment note: “if we were to just crack apart the different benefits of this approach”