How did the internet fundamentally change business? It's all about zero marginal costs, and playing a different game: blanket the world, capture customers, and only then, monetize them.
Note: This interview originally aired on The Acquirers Podcast on November 15, 2019. Tobias and I discuss a number of topics: Zoom, Slack, Atlassian; Ben Graham's investment in GEICO, in 1948; how GEICO was a disruptive innovator in the Clay Christensen sense; what factors made it so, and how can we identify similar companies today, and much more.
I'm particularly pleased with how it turned out, because a lot of our discussion remains relevant today.
Tobias Carlisle:
So just before we get started, tell me a little bit about Heller House.
Marcelo P. Lima:
We started in 2010. I left college and started out as a software developer for several years. And then I worked in real estate finance and worked at a hedge fund and started the fund in 2010, really, because I wanted to pursue some very specific ideas that I had. Back then it was the aftermath of the great crisis. There were a lot of these Ben Graham-type situations that were very attractive and a lot of them traded in foreign markets. Specifically for us, we found a lot of things in London. And the reason for this is London has a lot of these closed-end funds.
What happened is that on the lead-up to the crisis a lot of promoter hedge funds and investors were raising these closed-end funds to go and buy European real estate or go buy some other type of hard assets. There was even a fund that we invested in that had carbon credits for pollution offset. Some of these were pretty complicated.
The overall theme though, is that after the crisis, after 2008 and 2009, the asset values dramatically collapsed. A lot of these things had leverage, in the case of real estate. You had this double whammy effect: imagine that you have a closed-end fund trading close to NAV, but then all of a sudden the NAV gets destroyed. There's leverage on it. And then the spread widens.
Now it's trading at a 30% discount to a much discounted NAV. And then you lose liquidity. The stock doesn't trade as frequently. There's no sell side coverage because now the sell side can't really make a lot of money promoting these things. And so that created a lot of opportunities in London. We exploited that for several years. When it works out, it works out fantastically.
There's a very strong intellectual appeal to this Ben Graham type of investing. It's very intellectually stimulating because when you can mathematically prove that something is below liquidation value, maybe this building isn't worth what they're say it's worth, maybe it's worth 50% less, but even then it's impossible for me to lose money. I know what the rents are, I know what the market is like, et cetera. On top of that, there were a lot of these situations where these little closed-end funds had their board of directors taken over by a hedge fund and they had put in place a process to liquidate the investments and return cash to investors.
It was in fact liquidating and there was a hard catalyst in place. So that worked out really well. But as the years went by, the drawbacks of this strategy started appearing. These things are obviously pretty illiquid once you get in, but then as they liquidated, they become smaller and smaller and they become even more illiquid. So it's counterintuitive, but as the thesis plays out and becomes sometimes more de-risked, it becomes harder to participate in the opportunity.
Tobias Carlisle:
Just before you get too much further. So your background was software and then real estate financing, did you say?
Marcelo P. Lima:
Yeah.
Tobias Carlisle:
So how do you make the leap to value investment from there?
Marcelo P. Lima:
When I was in the real estate finance job I was trying to raise money for a consumer products company that I had started on the side with a friend. I came across this real estate investor who was doing tremendously well. And he said, hey, come work with us. It's a great opportunity. And that's when I jumped into real estate finance. Well, the first month on the job, I read Warren Buffett's biography.
Tobias Carlisle:
The Lowenstein one?
Marcelo P. Lima:
Exactly. And I completely fell in love with this idea that you could invest in different businesses. You didn't have to be tied to software or consumer products or real estate. You could analyze things and pull the trigger when it made the most sense to do so. I just became a huge Buffett fanatic and learned everything I could first about Ben Graham, and then Buffett. I read all of his letters and started going to Omaha and attending the Value Investing Congress.
In a way that was a very positive experience, obviously because you learn a lot. In another way, if you invert it, it also, I think, crystallizes in your brain a way of thinking that sometimes is hard to break. I'll give you an example. I saw, and this is something we'll get into later, perhaps, but I saw more than one short pitch on Salesforce at the Value Investing Congress. Famous investor goes on stage and says, this company is dramatically overvalued, look, they're losing money. This makes no sense. It's a short.
Of course, now we know that was incredibly wrong. But all of us were nodding and saying, wow, yeah, of course this thesis is right, this guy is so smart. This thesis is airtight and makes a lot of sense.
But we were fundamentally misunderstanding the change that came about with the rise of the internet and the delivery of software as a service. So that is all to say that being in this crowd of value investors and going to Omaha and listening to Buffett is a great education, but it can also perhaps prevent you from having a more open mind and embracing something that's new.
Tobias Carlisle:
Buffett's own style though, has always been more towards finding something that compounds and finding something that's at a very high return on invested capital sustainably. And then you look at the growth of that over an extended period of time and try and buy that at a meaningful discount that you get a margin of safety. So what has software as a service changed to what Buffett had traditionally done?
Marcelo P. Lima:
Well, I wouldn't say that it's software as a service. I would say it's the internet. The internet has created a world where all of a sudden you have something that never existed pre-internet, which is zero marginal costs. I can create a product. Let's say in the pre-internet world, I am Procter & Gamble, and I create a new type of detergent. Because I'm Procter & Gamble, I have my distributors, I have the slots in the supermarkets and I own some shelf space. I own perhaps the end caps, and I can place that new detergent. I can advertise it on television.
And just for the purposes of this discussion, let's be very extreme to make a point. Let's go back to the 1950s. There's like three TV channels, call it. I can advertise that. I have prime time advertising spots. I control all of them because there's only three channels. And I also advertise on radio. I advertise in print and the customer will go to the shop, go to the supermarket, buy the product, have that brand recognition and go home and rinse and repeat. And so the cycle continues.
Now, imagine all of a sudden the internet happens. And we have an explosion in the number of channels available. In fact, the whole concept of channels disappears when you have YouTube and Facebook and Instagram and Twitter, et cetera. The barriers to advertising are no longer there. The concept of primetime advertising to a large extent disappears. A lot of it is programmatic on Facebook and Instagram, et cetera.
Now you have, of course, this theme of software taking over more and more of the goods and services that we consume. Software is eating the world, as Marc Andreessen says. And software can be delivered at zero marginal cost.
Procter & Gamble is a great business. I'm just contrasting it with this new world. The skills that Proctor & Gamble has in terms of controlling distribution and the slots at the supermarket are less and less relevant or valuable perhaps in this new world.
In ecommerce you have infinite shelf space or practically infinite shelf space. And so the winner isn't necessarily the incumbent. Business dramatically changed with the advent of the internet. It's created a lot of dislocation and disruption is a very overused term, but it created a lot of disruption of incumbent businesses. And so I think that's what really changed. It's not something that Buffett was really onto until, until very recently.
Tobias Carlisle:
So how would you characterize something like Microsoft, which existed pre-internet and sold an operating system that locked you in? And they sold that on floppy disks that were virtually costless to them. They had some distribution and they had some minimal cost of producing them. But as any business school student will tell you that the first disc that comes out costs whatever $1 million. And then every disc that comes out after that costs 30 cents or something like that. And that was selling them for hundreds of dollars?
Marcelo P. Lima:
Yeah. That's a great point. And so that was a software as a service pre-internet, if you will. And it's funny that you bring this up in the context of our conversation about Buffett because there's this famous email that I'm a big fan of. I think it's just fascinating. Jeff Raikes, an executive at Microsoft, he also read the Lowenstein biography. This is back in the late nineties. And he wrote Buffett an email and said, hey, Mr. Buffett I just read your biography, and this is how Microsoft is similar to See's Candies. This is how we're similar to Nebraska Furniture Mart. He’s making all of the analogies that Buffett would understand. And he says, “I know you like Coca-Cola, but we at Microsoft have this money printing machine.”
We have all these PCs that have a Microsoft license just coming out of the factory. And he says, look, I'll concede that as far as the long-term moat, it's a little bit cloudier. It's hard to tell whether Microsoft we'll have such a strong moat 10, 20 years from now as Coke will. And Buffett replies to the email and says, look you guessed it, that's exactly how I feel as well as far as the moat. I can predict Coke a lot better than I can predict Microsoft. And what I find very interesting is that right around that time, the volumes for carbonated soft drinks were really starting to fall in the United States and around the world. So consumer tastes were changing dramatically.
Tobias Carlisle:
What were they shifting to?
Marcelo P. Lima:
I think there was this explosion of drink-
Tobias Carlisle:
Bottled water.
Marcelo P. Lima:
Sort of... Yeah, not only that it's funny, I think it was... I forgot his name. It's escaping me now, but the founder of Boston Beer [Editor’s note: Jim Koch]. He was talking recently on a call and he said, look, when I was a kid you had water and Coke and that was it. Now you have kombucha and ice tea and Starbucks. So there's huge variety of drinks that you could go out and purchase. The only thing that we knew is that carbonated soft drink volumes were going down and customers were really seeking alternatives, whether it was water or La Croix or coffee or who knows what else. That traditional advantage that Coca-Cola had was becoming less and less relevant.
It was fascinating because Coca-Cola the company in 2000 was trading at something like 40x earnings which is a very high multiple given the subsequent very low growth that it experienced. Microsoft was also trading at a very high multiple and that stock underperformed for a very long period of time after 2000. But it turns out now with the benefit of hindsight, twenty years later, that Microsoft was indeed a much more prosperous business than either Jeff Raikes from Microsoft or Buffett believed because perhaps they were in very high margin software. They were able to take those cash flows and invest in adjacent businesses like gaming and search. And they're not huge in search, but they do have, I think, 7 billion of revenue from Bing, which is kind of unbelievable, but gaming is a big business for them; all their entire office suite, Microsoft Azure, their cloud offering, of course is growing dramatically now.
It turned out to be a much better bet than Buffett believed. And I find that kind of fascinating. So to answer your question, the internet accelerated that process of delivering software at very low marginal costs. Whereas the marginal cost was a floppy disk back then, all of a sudden it was just the cost of transmitting the bits over the internet.
Tobias Carlisle:
So one of the companies that we were discussing before we came on air: Zoom. So would you like to just tell us a little bit about Zoom and then I just want to use that as a way that we can discuss software as a service?
Marcelo P. Lima:
Yeah. So what's fascinating is... And this ties into a broader discussion of disruption and what it is, but a company like Zoom, I think it helps if we talk a little bit about disruption, but the traditional theory of disruption envisioned by Clay Christensen is a theory that involves a new startup company attacking an incumbent at the very low-end. It's typically a low-end offering; it's a lower margin offering. It's attacking customers that the incumbent doesn't necessarily care about because they are so low margin.
Tobias Carlisle:
What would be an example of that?
Marcelo P. Lima:
The classical example that Clay Christensen uses in his book, back then he talks about hard disk drive makers and backhoe excavators. A more recent example, I think, is Netflix, which everyone in the audience will be familiar with. Netflix, when it started out mailing DVDs, it was in many ways an inferior experience. You didn't have that instant gratification of going to the Blockbuster store and getting the tape that you wanted. In some ways it was superior because you could have this long tail of the catalog available to you but –
Tobias Carlisle:
And also you didn't have time to go to the store.
Marcelo P. Lima:
You also didn't have to go to the store. You did have to wait though. You’d have to wait for the thing to arrive in the mail. I remember clearly when Netflix started offering streaming, it was horrible. It was low resolution. It was choppy because the bandwidth wasn't really there. It was clearly an inferior experience, but for a certain cohort of customers, it was good enough. And Netflix moved up market, so to speak and started improving their offering and then eventually started creating their own original shows and bypassing completely the existing value chain of the movie industry, which means not showing movies in movie theaters, going direct to consumer, et cetera.
By virtue of being effectively a software as a service offering, not in the traditional enterprise sense, but software delivered over the internet, they could then expand globally at a very rapid clip. So that's an interesting example of that. Now, there are other examples that do not fit the Clay Christensen definition, like Uber. And so Clay Christensen came out and said, look, Uber is not disruptive. And the reason it's not disruptive is they didn’t come in and offer an inferior service at a cheaper price and then moved up market. In fact, Uber started at the high-end. It was a replacement for black cars in San Francisco, the very expensive high-end.
Tobias Carlisle:
That was the original pitch for Uber, that it would be a way to call a limousine ride. There's that very famous deck that floats around. But most people probably became aware of Uber when it was in fact, that's probably not a bad description of it, it was a low-end, you got into somebody else's car, you didn't get into a taxi and it costs less. And that was kind of the idea.
Marcelo P. Lima:
Exactly. So, so even though Clay Christensen claims that Uber is not disruptive because of the way it began, the aftermath looks very much like disruption because to your point, they use the same technology that they developed for the high-end, meaning, the app, the routing technology, and this idea of matching riders and drivers. They were able to segment that across the low-end as well. Now they have a number of offerings; all sorts of things between low-end and high-end. And that's the difference. Now, when you have the internet, you're actually able to serve everyone because your marginal cost is essentially zero.
And so bringing the discussion back to Zoom. Let's think of a pre-internet business, if I'm delivering a conferencing service without the internet, hard to imagine what that looks like, but and I'm offering it to enterprise customers. They're paying a lot of money for me to offer these services. I'm leaving the low-end open for a worse offering at a lower price. I'm leaving myself open to disruption there by a low-end provider.
What Zoom can do is they can just make it free for customers who don't want to pay. As we were discussing prior to this call, anybody can go on zoom.com, sign up and start using their conferencing services for free. There's a time limit and there's obviously a lot of features that you don't get, but it's good enough. That prevents them, in theory, from being disrupted by a low low-end attacker. Then what they can do is they can build all the features necessary and requested by their most demanding customers as well. So they can serve everyone. And in fact, that's what they're doing. Of course, Zoom is just a toy model that we're using here. This applies to a lot of these enterprise SaaS companies.
Tobias Carlisle:
There are existing conference lines, though, for example, we're using Skype, which is a competitor to Zoom. And if you search free conference calling on the internet, you can find any number of free alternatives. So why would somebody shift from Skype or from a free alternative to Zoom?
Marcelo P. Lima:
Yeah, that's a great question. It's funny because a lot of these companies have a lot of competition. It's also the case for Salesforce. The space for a customer relationship management software is incredibly crowded and yet the company continues to do very well. What I think is the case, the companies that are very well managed, they have this tremendous focus on customer pain points. What is it that our customers are looking to do? What is it that's difficult for them to do with these other offerings and how can we make it easy for them? How can we remove friction and just make the experience better and better all the time, and build additional features that promote lock-in?
Zoom has this network effect, which of course Skype does as well. But if I am a large enterprise and I have now developed all these Zoom rooms so with all the hardware there, and I have my transcription, that's built in, and I've got Zoom, if I have multiple people on the call, Zoom will automatically focus on whoever's speaking. So if you're constantly focused on building these features it becomes harder and harder for somebody to switch out. And conversely, if you're, let's say, a Skype customer you're not going to have all that functionality. So as you become a larger enterprise, you might have HIPAA compliance or whatever type of capability that you need as an enterprise customer, that you might not be suited any longer to stay with a Skype offering.
Tobias Carlisle:
Zoom does have that potential to get those network effects. And I think that that's probably been Skyped saving grace, even though the product probably hasn't advanced much. And it is reasonably difficult to use still, almost everybody has a Skype number, and you can contact almost everybody on Skype, but what about something like Netflix? Where do you see Netflix as a competitive advantage?
Marcelo P. Lima:
Netflix has this incredible focus. I think part of the competitive advantage of these companies is the fact that they are a founder-led and they have earned the right to make these very tough decisions and steer the company in certain directions. And we see that very clearly today with Facebook, for example.
By the way, for the folks who are listening to this who are familiar with Tom Russo: Tom Russo likes to talk about... Tom Russo is a famous value investor. He's been running a very successful value portfolio for, I don't know, 30 years. And he likes to talk about the capacity to suffer and how family-owned businesses have this ability to burden the income statement today in order to reap the rewards tomorrow.
When you look at SaaS or something like Netflix, it very much looks like that. Netflix is famously losing money, however you cut it from a cash flow perspective or an earnings perspective, but they're playing a different game: they're playing this game of conquering the world and garnering those customers all around the world and having the largest audience because once they have the largest audience, they can then spend more money on content.
If I spend 15 billion, which I think is what they're spending this year in content, I can spread it across many more subscribers: 140 million subscribers. Whereas my competitors who are starting now, a lot of them are starting from zero. How can you stomach that type of content spend, if you have effectively zero customers to spread that content over?
Reed Hastings saw this very early on and he played out the chess in his mind and said, look, if that's the end game, then I just have to be as aggressive as possible in acquiring customers. That was the game that they were able to play in a nutshell. Their competitive advantage I think is really their strategy and their ability to pursue that strategy, which a lot of competitors don't have.
Tobias Carlisle:
How would you contrast them to something like a Disney offering?
Marcelo P. Lima:
Now, that's a question a lot of people's minds is what's going to happen to Netflix when Disney+ starts, I think next month in November. And I tend to believe that this is going to be complimentary and not really a substitution. I think there might be some substitution, let's say a family is subscribing to Netflix and might churn out and just subscribe to Disney+. But I feel that in the overwhelming case it'll be complimentary. In other words, people will add Disney+ to their Netflix subscription and not really substitute because Netflix still has an extremely large catalog. It's got completely different types of offerings in terms of the titles that they are willing to fund. Disney's going to be more likely a more family-oriented offering.
I think it's brilliant for Disney. It's about time that they did something like this. I've already signed up. I got the deal where you can pay a huge discount for three years. So I'm very excited for that, but personally, I won’t be canceling my Netflix subscription because I feel that it's something that I'll need to have as well.
Tobias Carlisle:
In order to compete in this new world, it sounds like you need to be a content producer, and it's not quite like being a cable company, which had to punch into your house and had the connection there. And If you're in various parts of the states, you have one cable company in your area and they basically charge you whatever they want. And there's nothing you can do about it. Whereas for Netflix, it's just a tile on your Apple TV. And if you look through the tiles that are available there are a lot there. And basically, I think that what people tend to do is they have a show that they like. So Game of Thrones, for example, means that you need an HBO subscription. Once Game of Thrones goes away, HBO’s offering looks a lot weaker, and I'm sure that there are a lot of people thinking about churning off. Netflix needs to have that killer show, doesn't it? Or is it always going to be the bargain bin of you just going to scroll for 45 minutes? Maybe you find something you like, maybe you don't, what do you think?
Marcelo P. Lima:
Yeah, look, and I want to be clear. We do own Netflix. It's a quite small position for us. So it's not something that I am pounding the table on, et cetera. I'm a big believer in the company long-term. And if you do the math I believe that the company can be extremely profitable in the future. Having said that, there are challenges, which you just outlined, what I think is the case is that Netflix has to have two things. One, it has to have filler content, which is the content that you described as just flipping through the catalog and being on a couch and vegging out and watching Friends or Seinfeld or whatever the case may be. But then it also needs to have that big attraction.
Stranger Things, or Game of Thrones in the case of HBO, something that will give people that must-watch content, which is the strategy that HBO or Showtime have been very successful with. We'll see how that plays out. I think that Netflix is doing a very good job with its local content. Things in India and in Brazil and Germany have been extremely successful in other countries.
A lot of that content travels very well. There's Spanish show, Money Heist, which was very popular around the world. That shows that this strategy of amortizing content costs across a huge base of subscribers works. We'll see if they can continue with their magic touch.
Tobias Carlisle:
It becomes a slightly more difficult proposition though, doesn't it? Because it becomes you need to have that killer content. You need to have that thing that people will go and see. It makes you a little bit more like a movie studio, which if you miss in any given year, then that could hurt you pretty significantly.
Marcelo P. Lima:
Right. But let me take the other side of that argument. If I have the largest content budget and the largest audience, then I'm going to attract the best producers, the best actors, the best talent. And so it becomes a lot easier to actually produce winning content. If I'm Netflix and I have a roster of phenomenal writers, producers, and actors, I'm going to have my share of hits. And my share is going to be bigger than the next guy's share because the next guy is spending a fraction of what I am in content has a fraction of my audience.
So you could, I think, very persuasively make this argument of increasing returns where the big gets bigger by virtue of already being bigger than everybody else. And having that ability to attract all that talent.
Tobias Carlisle:
Just changing gears a little bit. We leaped ahead, you sent me this wonderful document, a white writing on a black background and there were some very interesting points in that early on that I want to make sure that we address. You talked about some of the the problems with value investing. So let's go back to that. I think it was largely driven by a study that Irving Kahn did on GEICO. So would you please take us through that?
Marcelo P. Lima:
Yeah, sure. You noticed I like to use dark mode in Word.
Tobias Carlisle:
In everything.
Marcelo P. Lima:
In everything. We were talking about the drawbacks of this liquidation type of investing. How it becomes a smaller part of your portfolio as it goes along. And then you also have to pay taxes when the thing finally liquidates two, three years from now. Again, going back to Tom Russo, he talks about having all these unrealized gains in something like Heineken. That he's owned for 30 years. That's pretty attractive, never having to pay taxes on it, if you don't sell it. In our case, we were also in foreign markets. So we had to hedge the currency, which is fine, but hedging the currency has a drag on your portfolio because it costs money to hedge.
Then I think the biggest one is opportunity costs. What I realized is that while I was grinding this out and going on these one-off situations and putting all this work into figuring out what these liquidations were going to look like, there were these phenomenal businesses that were just compounding, and I wasn't even looking at them. When I opened my eyes to this, I went back and rediscovered this part of Ben Graham, which I had read many years prior.
This epiphany was Ben Graham in the appendix to Intelligent Investor. He writes almost matter-of-factly. He says, oh it's ironic that we put 20% of our fund in this one company. He doesn't even name the company. We know it's GEICO, but he doesn't even name it. We put 20% of our fund in his company in 1948. And the thing was a 145-bagger. Actually, between the time he bought and the peak, it was a 600-bagger, but the way he describes it in the appendix, it makes it sound like it's a 145-bagger over 24 years.
He says that one stock made more money… He says, “Ironically enough, the aggregate of profits occurring from the single investment decision far exceeded the sum of all the others realized over the 20 years of wide-ranging operations in the partners’ specialized fields and involving much investigation, endless pondering and countless individual decisions.”
Everybody knows the famous stories of Northern Pipeline and the Guggenheim liquidation and all that. Those things were a lot of work. He had to solicit proxies and wait for the next annual meeting and build his case and all that these things were illiquid.
Even back then it was a ton of work, and the profits were capped, and then you have to move on to something else. And that one investment in GEICO made more profits than all that other activity. So, I thought: could you have put on a 21st century analyst’s hat, and analyzed GEICO with all the things that we know today in terms of disruption theory and the way businesses evolve, and the total addressable market, and all that? I went back and I said, look what is GEICO?
GEICO was this company that Graham invested in 1948. It was founded in 1936. What could we have figured out right before Graham invested? Let's go through the list here. In 1948, if you looked at car registrations, this is public government information, car registrations had compounded at 20% a year over the previous 47 years. A lot of people are buying cars. If you look at cars per capita there was only one car or truck per capita in the U.S. [Editor’s note: I misspoke here: what I meant to say is that there was one quarter of a car per capita, that is, one car for every four inhabitants in the US, at that time, as I explain in this blog post.]
Maybe the population is growing a lot. There's going to be baby boomers. Now, maybe that wasn't a known fact, but then maybe there's room for more cars. I found this article in the New York Times from February 1947. A year before Graham invested and the article quoted a car dealership saying, look, if you try to buy a car, you're going to find a waitlist that's going to take you months. And it might even take you into the next year. We just don't have enough cars to supply to people. And part of that was because there were steel shortages after the second World War, but clearly there was pent-up demand. In fact, I found this funny thing, which is kind of crazy to imagine today. People were vacationing in Europe, and they were taking their cars in the ships to go to vacation in Europe, because you couldn’t land in Europe and go to Hertz and rent a car. It didn't exist.
Tobias Carlisle:
That's amazing.
Marcelo P. Lima:
And then finally the interstate highway system had been talked about since 1938 and there was a series of documents that had been published with the proposal for the interstate highway system. So, okay. You're Ben Graham and you're like, okay, well, there's going to be an explosion of miles driven. And if you think of Brian Arthur's idea of increasing returns, when you have a new technology, like the car, you're going to have a lot of things that are enabled by the car. So you're going to have gas stations and you're going to have probably a lot of hotels and destinations that people can actually take their cars to go see and things to do.
That will create this flywheel effect where the more things there are to go and do the more cars get sold and vice versa. And in fact, Walmart was founded a couple of decades later in the sixties as a business uniquely enabled by the car. Finally, GEICO was a disruptive business because it was direct to consumer. It didn't have the sales agent and the additional layer of costs associated with that. They had something like half the cost structure of a traditional insurer. This is also in the Irving Kahn study. I think it was a 14% cost ratio compared to 28 percent cost ratio for a traditional insurer.
Finally, it had a better risk targeting model. This is before big data, before FICO Scores. How do you figure out the low-risk cohort to target? Government Employees Insurance Company. It's the name of the company. They slowly started spreading and in going to veterans and teachers and people with bachelor's degrees, which they were all low-risk customers. So it ticks all the boxes. You have this huge addressable market; you have a disruptive innovator; and you have this huge secular tailwind.
Now, could you apply that type of thinking to investing today? And if so, could you just build a portfolio of GEICOs and just get rid of all those cigar butts?
Tobias Carlisle:
And so that's what you sought to do with Heller House. You're trying to identify this company. So you've talked about this is the three components of disruptive innovation, just to, I think I'm just summarizing what you've just said, the enabling technology, the innovative business model and a coherent value network. And that's the, I missed the gentleman's name, but that's the guideline that you looking for when you're seeking these businesses?
Marcelo P. Lima:
Right. So this is textbook Clay Christensen. And you could argue whether these things have evolved, but I think the core of the theory is completely spot on. And the enabling technology is something like internet distribution. The innovative business model is something like Netflix mailing DVDs, and then streaming or software as a service, a new way to deliver the product or service. Then the coherent value network is really about having, in the case of the automobile, the value network was the creation of gas stations and roads, and everything that enabled that business to develop. The example that Clay Christensen gives him the book, which is super interesting, is steam ships versus sailing ships from the 19th to the 20th century. Steam ships were about to take over. But before that, they were worse in just about every way.
They were slower than sailing ships. They were more expensive than sailing ships. They were very suited to inland waterways where you didn't have wind or perhaps you had to travel against the wind and performance was measured in a very different way. The customers of the sailing ships who needed to go across the Atlantic or across the ocean had really no use for an inferior technology, but that inferior technology kept getting better and better and better. And over time it took over the business of the sailing ships. By then the manufacturers of the sailing ships had no expertise in building high technology steam ships and all of them failed at transitioning to this new technology.
Tobias Carlisle:
One of the very interesting points that you've raised, you pose this question. What if every company in the S&P 500 became a software as a service? What impact does that have?
Marcelo P. Lima:
Yeah, so Marc Andreessen has this thesis. That software is eating the world, and we can clearly see that happening, but then he goes a step further and he says, look if that's the case, then that means that every company must become a software company. We're already seeing that. I go to a lot of these conferences, and you see old school companies saying “We're hiring a bunch of developers. We're becoming a software company. We're hip, come work for us.” And it's true to a big extent, and they have to do it because otherwise they'll be left behind. But then Marc Andreessen says, well, let's take it a step further. If this is all true, then it turns out that the largest businesses in the world, the most successful businesses in the world will eventually be software businesses.
We can already see that happening in many cases with things like Facebook and Alphabet (Google). These are some of the most valuable companies in the world, and they're almost purely software based. It's an interesting thought experiment because software as a service has this very interesting characteristic. Let's go back to Walmart, for example. Walmart for many, many years after its inception in the sixties, it was not generating free cashflow. It was a "money-losing business" because they were building a lot of stores. They were opening a lot of distribution centers and they were expanding across the country. But the unit economics were very attractive. They would build a store and they had a very high ROI on those stores.
The same thing is true for the most successful software as a service companies. They are spending a lot of money on research and development today to build the product and maintain it. They're also spending a lot of money on sales and marketing because back to our previous point, it makes a lot of sense to expand and try to capture as many customers as possible because then you can deliver services to them at a zero marginal cost.
That makes the income statement look awful. It makes it look like—which it is—a money losing business. But if the unit economics are there and the way we measure unit economics in the SaaS world is this lifetime value of the customer. How long are you going to be my customer? How much are you going to pay me? And so that's the value of you as a customer divided by the customer acquisition costs, which is how much am I spending to acquire customers in the first place.
If that's a very high ratio, then it's a good return on investment. Therefore, I should be pouring money into this opportunity. In fact, the canonical example is that if you're a successful SaaS company, the faster you grow, the more money you lose and now obviously everything that's good gets taken to an extreme. You also have bad SaaS companies that are growing like this, and you have to distinguish between the good ones and the bad ones.
By virtue of this, these companies trade at very high P/E multiples or sometimes inexistent P/E multiples if they're losing money. Just as a thought experiment would be funny if the most meaningful S&P 500 companies or the large, let's say the entire index is composed of these companies, because software is leading the world and that would just be bonkers.
It would throw off a lot of people's models of the world because people are relying on the S&P trading at a several times P/E multiple, the 30-year or the 10-year treasury is at two and a half percent. Sort of makes sense. But all of a sudden if your P/Es are a 100x… let's say that they're fully justified, because these companies are growing and let's say a 100x is the correct number. It would be this crazy world. I just think it's interesting to think that way, as a forcing function, to adapt as an investor and how you analyze things.
Tobias Carlisle:
So say that may manifest as a bubble or it might appear as a bubble, but it wouldn't in fact be a bubble because they may be undervalued at that level?
Marcelo P. Lima:
Right. I think that if we could snap our fingers and that world existed all of a sudden I think a lot of people would be saying we're in a bubble. The same way a lot of people are saying that SaaS companies are in a bubble right now. And I think some of them are, but I think some of them are not. So you really have to distinguish.
Tobias Carlisle:
Do you view that as a transition state or as an instant?
Marcelo P. Lima:
Well, I'm not really saying that this is going to happen. I just think it's an interesting thought experiment. I think the most likely scenario is that we're going to have a lot of very valuable software-based companies, but by the time they become big enough to be a big part of the index, they will be already very profitable and therefore trading at a much more normal multiple. Facebook, as an example, trades today at a very low multiple relative to the quality of the business. It's obviously not an enterprise SaaS company, but it's a "software as a service company" in the sense that it delivers software over the internet at next to zero marginal costs.
So very different. And I know, I don't want people to get angry at me not making this distinction because they don't have monthly subscribers and churn numbers and all that. But effectively it's a software company.
Tobias Carlisle:
So you said that the strategy in this marketplace is to land and expand. So can you just describe what that is?
Marcelo P. Lima:
Yeah. I heard this phrase recently, and I don't remember who said it, but you cannot harvest off land that you haven't you haven't conquered yet. first you must conquer the land and then you must plant the seed and harvest the land. Let's play this out. Because there's a zero marginal cost to conducting business over the internet, if I am not the aggressive guy who's going to build, whatever product it is, let's say whether it's Zoom or Slack or Netflix or whatever it is, if I'm not going to be the aggressive guy who's going to build this product and conquer the world first and then monetize my user base, somebody else will.
Let's say I launch Zoom and I'm charging 50 bucks a month; somebody will undercut me and offer a freemium version where it's free to use, and then you can upgrade as you need it. And I will be disrupted if I'm charging 50 bucks a month to everyone; that'll prevent me from landing the most customers, therefore the better strategy is to offer that freemium version and get the most customers possible. Invest heavily in sales and marketing to then try to upsell those customers into a paid tier.
You can see this clearly with a company like Slack. Slack offers you this free tier where several users can use the product. And they'll say, guess what, if you want to search past 10,000 messages, you must pay us because we're not going to store your messages if you don't pay us.
That becomes a choice, whether you want to upgrade or not, but to get those better features you must. They already have you as a customer, you're already used to using them. You have all your channels set up and you're like what? Seven bucks a month? Fine. Then you become a customer. I think it's a very interesting strategy. If you play the game theory on this, you can see how it's a strategy that makes a lot of sense.
Tobias Carlisle:
You said that the incumbents have two problems when they're confronted with challenges like this.
Marcelo P. Lima:
Right. The first problem is, and this is something that Clay Christensen describes a lot, and very clearly is, this idea that if you're going to... Because disruptors start out at the low-end, and because they start out with these relatively unprofitable customers, and a lot of times they start in markets that are inexistent or very small because these are not markets served by incumbents, by definition. That becomes a very, very poor proposition. If I am a large incumbent and I have $5 billion in sales, why on earth am I going to spend my time chasing an opportunity that may or may not pan out in a market that may or may not exist. At a lower margin that I have right now? I'm never going to do that. I'm going to dedicate all of my resources to serving my most profitable customers. And trying to sell them more things. So that's a huge problem.
And then the second problem is they have this skill in controlling distribution. In the pre-internet world, controlling distribution, back to the Proctor & Gamble example. I have the slots in the supermarket and the end caps and the trucks to the supermarket. And I also have this skill in segmenting my customers. So I'm going to sell... I have the premium detergent and the medium and the low-end detergent. I do not have the skill, however, in going direct to consumer, I do not have the skill in internet advertising. I do not have the skill in aggregating customers at scale in this land and expand motion.
So it's very difficult for these companies to do these two things obviously to disrupt themselves and to have this internet mentality. By the way, it's fascinating because if you look at the history of Amazon, Jeff Bezos read this book Innovator's Dilemma and he makes all of his top executives read the book. It's fascinating to me because as you read the book, you're like, wow, Amazon has engineered itself to be the antidote to this. You see a problem in the book, and you're like, wow, that's clearly why Amazon does X or Y.
I'll give you an example. Small markets or inexistent markets only really matter if you're a small team. If you're a huge company, you're not going to look at it. So what does Amazon have? Amazon has these two-pizza teams. Because for a two-pizza team, a market that's $10 million in sales or a hundred million dollars in sales is going to be a meaningful market. So that takes care of that problem. It's very interesting. There are endless examples of this in the book where you're reading it and you're like, well, Amazon operates this way to counteract this problem.
Tobias Carlisle:
So I love the description that you've given. How does it show up in your portfolio? What do you hold and how do you, a further question, how do you manage those positions? How many do you hold? How many do you trim when they grow? How do you look after the portfolio?
Marcelo P. Lima:
Yeah. That's, that's a great question. So I really want to get to the trimming question, because I think it's a very interesting point. As you study these companies you start seeing all these secular trends and secular trends are displayed in these technology adoption curves. If you go back and look at the history of technology since the industrial revolution every technology has had this adoption curve, everything from electricity to indoor plumbing, the automobile, dishwashers, radio, TV, internet, social media et cetera. And it starts out obviously with zero penetration, and then it goes how many households have the access to this product or service? And then you can see, okay, well, what are the businesses that are taking advantage of this adoption curve? And what are the businesses that have the wide moats or the growing moats that are riding this wave?
Apple is this canonical example that... We don't own Apple by the way, but it's a canonical example of a company that created the iPhone. And it was a very successful new category, this computer in your pocket with a touch screen, et cetera. It became this enormous adoption curve of more and more people around the world adopting this product. They rode that to become one of the world's most valuable businesses. Facebook, with the adoption of social media around the world; Google with search, and then Android. Where on these adoption curves are we today? What are the adoption curves of the future and how can we position ourselves to take advantage of that?
That's how I think about the portfolio. One big adoption curve that I'm very excited about is cloud computing. The analogy there is back in the day in during the industrial revolution, every factory was electrifying, electricity became a thing in the late 1800s. People were electrifying their factories and replacing water wheels or steam power with electricity. It was very burdensome. You had to build your own power plant and you had your own engineers and all that. This entrepreneur Sam Insull came and said, look, I'll build a centralized power station and I'll run a wire to your factory. Now you can get rid of all this machinery and all your engineers and all that.
Well, that's exactly what's happening with cloud computing. Get rid of all your servers and your database administrators and a lot of your IT folks dedicated to just maintaining this hardware. We'll keep it all for you in the cloud in a central secure data center. We'll run a fiber optic cable to your office, and we'll give you incredible functionality through all these pieces of software, these APIs and all these things that we're developing and constantly updating and patching.
So that's exploding. And every company around the world is in the process of moving to the cloud or is already on the cloud. The largest one is Amazon Web Services. Those companies, Amazon Web Services, number one, Microsoft Azure number two, and Google Cloud Platform number three; I'm ignoring the Chinese clouds, Alibaba and Tencent, because I have a hard time believing that they will be successful in the west. Which Chief Technology Officer is going to want to put data in a Chinese cloud unless you're serving Chinese customers and you're required to? I don't think it's going to be that big of a business in the west.
Those companies are growing very fast. You can see the numbers AWS discloses, its sales and operating income. Then you have estimates of what the size of the market is. Where are we on that adoption curve? That's very exciting, but then this is an enabling technology. The cloud enables new business models to be built on top of it. And of course that's software as a service. And so what are the platforms being built on top of AWS and Azure and Google Cloud? That's another layer that we're looking at, obviously to see this adoption, these adoption curves taking place. That's a very big theme in the portfolio.
Tobias Carlisle:
How do you value something like that? You're estimating the size of the market, the total addressable market. You're trying to work out what that's worth at close to a steady state, and then you're discounting that back. I mean, I understand the working out the lifetime value of a customer, working out the cost to acquire that customer. So how do you ultimately value something like that?
Marcelo P. Lima:
Yeah, great question. And I want to get to your question about trimming, which I think is a very interesting question, but the value of anything is just the present value of its future free cash flows. I do believe there is strategic value. We have all these customers, we haven't monetized them yet, and we have... But I don't value things based on strategic value. It's really old school free cash flow. The way I think about it is these companies all have a certain margin structure and the margin structure today is probably very different from the margin structure at a maturity. And by maturity, I just define it as the company is large enough now to have a more rational cost structure in terms of sales and marketing expenses to its revenues and R&D to revenues, et cetera.
We can look at comparables and look at what these companies should have in terms of margin, like Salesforce when it was just starting out versus Salesforce today. And there's plenty of examples. So a company like Slack is a good one to pick on because Slack has 85% gross margins. And that's a very good starting point. If you're selling a product that has 85% gross margins then you have a lot of play there as far as your operating expenses.
It's quite likely that Slack is going to have very healthy free cash flow margins down the road. And you can look at Atlassian as a potential comparable although Atlassian is the gold standard of very low sales and marketing expenses, but it's possible that Slack is going to have a 25 to 30% free cashflow margins which is in fact what the company has guided to.
We always build a model for these companies and we figure out, well, how many customers can they have, realistically? How much can their top line grow and how will the margins evolve? Is there going to be any free cash flow along the way that gets piled up on the balance sheet or is used for mergers and acquisitions? We try to come up with a realistic value for the business several years from now. And then we figure out what the internal rate of return is between today's price and the future price of the business.
And this is of course... By the way, I always tell everyone, I know that my models are wrong. I just don't know the magnitude and the direction, because these things are a lot of art and a little bit of science, but it's better to have a rough idea of what the market is imputing. So I'll tell you back in the day when I was investing in things at 10x earnings, for example. The way that I would do this, I remember buying eBay, for example, eBay in 2008 was trading at something like six times free cash flow at the bottom, or maybe it was at the bottom in 2009. And I remember buying this thing and making a case that look anything below 10x earnings, if I'm using a 10% discount rate, if you plug it into the perpetuity formula it's implying perpetual decline. Saying this company is never going to grow again.
And I just think that's absurd. I think eBay will grow. So that is a very interesting way to bracket your margin of safety. And you can do the same thing with high growth companies. You can say, well, how low of a rate of growth do I have to put into my model, such that I get the current enterprise value between now and judgment day. And sometimes it's just this incredibly low rate of growth. And you're like, well, I clearly do not believe that this is going to be the state of the world.
I think that there's a very low risk that I'm going to lose money on this, but I have huge optionality to the upside, if the company grows the way I think it'll grow. Back to our discussion on Microsoft versus Coca-Cola, obviously not all of these companies are going to become big like Microsoft. But they have a lot of flexibility as far as building products and discovering customer pain points because they are software-based companies; they all have visibility. They know what you're clicking on in their application, what features you're using, what features are frustrating their users, and they can then build things to improve that customer experience. It's this very tight flywheel that has really never existed in the history of business. So it's very interesting.
Tobias Carlisle:
Marcelo, it's absolutely fascinating, and I could keep on speaking to you for another hour. I'm very grateful for you coming on and educating me and everybody who listens. If folks want to get in touch with you, if they want to learn more about you, how do they go about doing that?
Marcelo P. Lima:
Well, thank you, Tobias. And I'm on Twitter, @MarceloPLima and the website is hellerhs.com.
Tobias Carlisle:
And I'll make sure that those links are in the show notes. Marcelo, thank you very much.
Marcelo P. Lima:
My pleasure, Tobias.