What We’re Reading (Week Ending 03 October 2021)

What We’re Reading (Week Ending 03 October 2021) -

Reading helps us learn about the world and it is a really important aspect of investing. The legendary Charlie Munger even goes so far as to say that “I don’t think you can get to be a really good investor over a broad range without doing a massive amount of reading.” We (the co-founders of Compounder Fund) read widely across a range of topics, including investing, business, technology, and the world in general. We want to regularly share the best articles we’ve come across recently. Here they are (for the week ending 03 October 2021):

1. Epic Games believes the Internet is broken. This is their blueprint to fix it. – Gene Park

To Epic Games CEO Tim Sweeney, people are tired of how today’s Internet operates. He says the social media era of the Internet, a charge led by Mark Zuckerberg’s Facebook, has separated commerce from the general audience, herding users together and directing them to targets of the company’s choosing rather than allowing free exploration.

“Now we’re in a closed platform wave, and Apple and Google are surfing that wave too,” Sweeney said. “As we get out of this, everybody is going to realize, ‘Okay we spent the last decade being taken advantage of.'”

For years now, he has eyed a solution: the metaverse. And steadily, over several years, Epic has been acquiring a number of assets and making strategic moves with the goal of making Sweeney’s vision for the metaverse a reality.

The simplest way to define the metaverse is as an evolution of how users interact with brands, intellectual properties and each other on the Internet. The metaverse, to Sweeney, would be an expansive, digitized communal space where users can mingle freely with brands and one another in ways that permit self-expression and spark joy. It would be a kind of online playground where users could join friends to play a multiplayer game like Epic’s “Fortnite” one moment, watch a movie via Netflix the next and then bring their friends to test drive a new car that’s crafted exactly the same in the real world as it would be in this virtual one. It would not be, Sweeney said, the manicured, ad-laden news feed presented by platforms like Facebook.

“The metaverse isn’t going to be that,” Sweeney said. “A carmaker who wants to make a presence in the metaverse isn’t going to run ads. They’re going to drop their car into the world in real time and you’ll be able to drive it around. And they’re going to work with lots of content creators with different experiences to ensure their car is playable here and there, and that it’s receiving the attention it deserves.”…

…At the core of Epic’s metaverse vision is a change in how people socialize on the Internet. Sima Sistani, co-founder of the video chat social network Houseparty that was acquired by Epic in 2019, believes interactions will move away from “likes,” comments and posts about people’s personal lives and toward more complex interactions where users share and participate in experiences across various services.

“If the last generation is about sharing, the next generation of social is going to be about participating,” said Sistani, who has held positions at Tumblr and Yahoo before starting Houseparty. “Maybe I didn’t call it the metaverse then, but that’s what it is. It’s people, interactive experiences, coming together and moving from one experience to another, having this shareability to move beyond walled gardens.”

Sistani’s description closely resembles the innate, interactive nature of video games, which offer more ways to engage with brands and other users than simple ad-filled timelines.

“We’ve seen this happen in the past,” Sistani said. “I come from a media background, and people moved from traditional media to social media. And this new generation is moving from social media to games. That’s where they’re having these conversations. That’s where it’s beyond the ‘like,’ beyond the news feed. And that, that’s the metaverse.”

Nowhere has this been more visible in Epic’s portfolio than its flagship title, “Fortnite,” the 100-player, battle royale-style game that surged in popularity in 2018. As The Washington Post reported last year, Epic Games has become a front-runner in creating the metaverse in part thanks to the hundreds of millions of users who log into “Fortnite” every month to create, talk and, of course, shoot each other with digital guns in multiplayer arena combat. The game is a forum in which players interact in real time with intellectual properties from Marvel or Star Wars, one that both pulls from and inspires pop culture. It has even been a showcase for premium consumer goods.

2. The Mystery Man Who Made Amazon an Ad Giant – Sahil Patel and Mark Di Stefano

Paul Kotas may be the most important person in internet advertising that almost nobody in advertising has ever heard of.

Mention Kotas—the leader of Amazon’s burgeoning, multibillion-dollar ad businesses—around ad agencies, as The Information did to more than a half-dozen senior ad executives, and you’ll get blank stares. One of those executives, whose agency will spend between $100 million and $150 million on streaming video ads alone this year with Amazon, Google-stalked Kotas in the middle of a phone interview to see if he could recognize Kotas. He couldn’t…

…Kotas himself seems more than happy to remain anonymous. At least twice in the past, Kotas has made a curious request of his Amazon colleagues before meetings with ad executives: He didn’t want his team to introduce him by his actual title, which is senior vice president, or reveal who he was. Instead, he asked colleagues to tell the clients that he was involved in “product,” according to people who heard those requests.

One explanation for Kotas’ stealthiness is that Amazon, at least in the past, wanted to avoid drawing unnecessary attention to its ad business for as long as possible, according to current and former Amazon executives interviewed by The Information. If competitors like Google grasped how aggressively it was going after the ad business, Amazon executives worried, those rivals might return the favor by pushing harder into Amazon’s core commerce business.

Another person familiar with the matter said Kotas made the requests so he could hear unbiased feedback from ad agencies without his title influencing what they said…

…As Amazon’s ads business grew, so did Kotas’ stature at the company. Initially, he was in charge of product and engineering for advertising at Amazon, with Jeff Blackburn, another longtime executive at the company, overseeing the sales side of the business. But in early 2014, Bezos put Kotas in charge of the entire advertising group. Kotas had been part of Amazon’s S-Team, a group of senior leaders who plot long-term strategy for the company, since 2011. He was elevated to the rank of senior vice president in 2014.

As an engineer, Kotas seems to have a preference for the technical side of digital advertising. At a gathering of Amazon executives in 2017, Kotas was asked what he found the biggest challenge in the ad business. His answer, according to a former Amazon executive who heard the remarks, was to “turn a relationship business into an automation business.”

Around the same time, though, Amazon ad sales executives realized they needed to invest more, not less, in the relationship side of their business. This required assigning someone to build out a team focused on working with and interacting with ad agencies, which control many big marketers’ ad budgets.

Seth Dallaire, Amazon’s vice president of global ad sales at the time, appointed Ryan Mayward to the task of starting an agency-partnership program and team. While Kotas signed off on Mayward’s appointment, he remained on the fence about the initiative until Mayward made a more comprehensive proposal for why it required such a large team, said a person familiar with the Amazon ad team’s discussions at the time. The reason for the hesitation: Kotas and Amazon’s ad team preferred to work directly with brands whenever possible, and they required convincing that the approach needed to change to keep ad revenue growing.

Eventually, Kotas came around to the plan for the ad agency team.

Over time, the company’s ad business grew into one of its most lucrative new efforts in years. In 2015, Amazon’s “other” segment had just over $1 billion in revenue. Last year, it brought in more than $21 billion.

3. The Tesla ‘Bubble Or Not’ Debate – Tom Lauricella, Catherine Wood, Daniel, Needham, and Rob Arnott

Needham: You made a very strong case for electric vehicles. Why will Tesla be the one that benefits from that? Why won’t the more traditional autos or the many other electric vehicle manufacturers capture that trend?

Wood: The traditional auto manufacturers had to make or have to make a major leap.  The vast majority of their sales today are gas-powered vehicles. They need to transition to electric. Tesla’s already started electric and has four major barriers to entry–has created four major barriers to entry. One, battery costs. It built its cars on cylindrical batteries. Most other auto manufacturers base their cars on lithium-ion pouch batteries. The costs of lithium-ion pouch are much higher today–I think roughly 15%, 20%–than the cylindrical batteries that Tesla uses.

The second barrier to entry is the artificial-intelligence chip that Tesla designed. Now, Tesla is taking a leaf from Apple’s book. As you will remember, Apple created the concept of a smartphone. It believed that we would have a computer in our pocket. Nokia, Motorola, and Ericsson did not believe that. They did not design their own chips.  And you know where they are today.

The other barrier to entry is the number of real-world miles driven that Tesla has collected. It has more than a million robots out there collecting data and sending it back every day.  My car is one of them. Therefore, it is able to discern corner cases and design its full self-driving system to incorporate these corner cases in a way that other auto manufacturers cannot.

And then the fourth barrier to entry–and it surprised me this one lasted as long, but I guess the dealer system was the reason–Tesla is still the only car doing over-the-air software updates to improve performance and prevent breakdowns.

Those four barriers to entry we believe have put Tesla ahead, and we think the distance actually is increasing.

Needham: Rob? You’ve got some opinions on electric vehicles and also Tesla.

Arnott: I certainly do. We wrote a paper earlier this year called “The Big Market Delusion,” which looked at industries that are up and coming that are disruptive. Kudos to Cathie on looking for disruptors. They’re very, very important. But disruptors get disrupted, and I’ll come back to that in a minute.

The thing that we found very interesting is you find these cases in the Internet bubble, in the supercomputer bubble in the early ’80s–the list goes on and on–where every company in the industry is priced at lofty multiples, as if they’re all going to succeed.  Yet they’re competing against one another, so there will be winners and losers. And the market’s pricing things as if they’re all going to be winners.

I mentioned disruptors get disrupted. Palm was spun off from 3Com back in the year 2000 and had an initial value that was more than 3Com was valued at before the spin-off, and within a day or two was worth more than General Motors. Palm was disrupted. BlackBerry came along with a better product. BlackBerry was disrupted. Apple came along with a better product.

So, what we find is again and again: Disruptors are massively important to the economy and to economic growth. But you have to look at (a) how disruptive are they, (b) how much of a premium are you paying for that disruption, and (c) are they vulnerable to being disrupted themselves?…

Needham: So there’s upside there. Maybe, Rob, just on to the fundamentals, I’m going to use a quote from a very well-known value investor, Warren Buffett. He said, “Beware of past performance proofs in finance. If history books were the key to riches, the Forbes 400 list would be full of librarians.”

Your approach is very much geared in looking at historical fundamentals and relying on some of those relationships to hold. So, how do you think about some of these disruptive elements when you’re building a strategy based off historical fundamentals or making assumptions about fundamentals?

Arnott: A lot of our work is based on mean reversion. Cathie alluded to mean reversion valuation multiples for the disruptors. Mean reversion is the most powerful factor at work in the capital markets. It shows up on earnings growth. When you have very rapid earnings growth, it tends to mean-revert down. When you have tanking earnings, it tends to mean-revert up. Not in all cases. There are value traps.

So when you’re looking at a whole spectrum of disruptive companies, there will be some that turn out to be spectacular. Go back to the first tech bubble. How many of the 10 largest market-cap tech stocks in the market in the year 2000 outperformed the market over the next 10 years? Zero. Not one. How many outperformed over the next 20 years? One, Microsoft. What about Amazon and Apple? They weren’t anywhere near the top 10. They were bubbling up from underneath, and in the case of Apple, was perceived to be poised of the brink of ruin.

So what you find is that when you have bubbles, and bubbles can appear anywhere–I’ll come back to a definition for them in just a moment–when you have bubbles, they tend to burst. Our definition for a bubble is a very simple one. If you’re using a discounted cash flow model or some other valuation model, you’d have to use implausible assumptions to justify today’s price. We plugged in 50% growth for 10 years for Tesla, assumed profitability matching the best of any automaker–and that may be the wrong choice, but the best of any automaker of any single year of the last 10 years–and we came up with a net present value of 430 bucks. I view 50% growth as implausible. Cathie does not. So I view Tesla as a bubble. Cathie does not.

But two things are interesting about bubbles. One, they can go much further and last much longer than any skeptic would expect. So be very careful about short-selling bubbles. You can make a ton of money if you have a good exit strategy.

The second observation about bubbles is that implausible growth assumptions doesn’t mean impossible. Amazon in the year 2000 would have qualified for my definition of a bubble, because you’d have to use extreme growth to justify the then-current price. Amazon was a terrible investment in the 2000s, got it all back with room to spare in the 2010s. And in the 2010s, it grew 26%, 27% per annum, which was enough to make it 11 times as large as it was 10 years previous–11-fold growth.

Now, to justify Tesla’s current price, you’d have to assume roughly 50-fold growth over the next 10 years. Is that impossible? No, anything is possible. Do you believe it’s plausible? I don’t. So I view it as a bubble. And as with Amazon in the year 2000, I could be proven wrong. But as with Amazon in the year 2000, you might have to wait a while for the market to catch up to the actual growth opportunities if they are as extravagant as Cathie says.

4. Evergrande’s Fall Shows How Xi Has Created a China Crisis – Niall Ferguson

A major mistake of the Cold War was the tendency of Western observers to overestimate the Soviet Union. I have often wondered if the same mistake is being repeated with the People’s Republic of China. Then again, for every article over the last 10 years that predicted China’s economy would overtake that of the U.S., there were at least two prophesying a “China crisis.”…

…Will China surpass America? No, I don’t think so. Nearly three years ago, in the heat of a lively debate in Seoul, I bet the Chinese economist Justin Yifu Lin 20,000 yuan (roughly $3,000) that China’s economy — defined as GDP in current dollars — would not overtake that of the U.S. in the next 20 years. I am sticking with that bet, even if the Lehman Moment for the Chinese financial system never comes. Here’s why.

Let’s begin by recalling how many experts believed the Soviets would overtake America. In successive editions, the economist Paul Samuelson’s hugely influential economics textbook carried a chart projecting that the gross national product of the Soviet Union would exceed that of the U.S. at some point between 1984 and 1997. The 1967 edition suggested that the great overtaking could happen as early as 1977. By the 1980 edition, the time frame had been moved forward to 2002-2012. The graph was quietly dropped after that.

Samuelson was by no means the only American scholar to make this mistake. A late as 1984, Harvard’s liberal guru John Kenneth Galbraith could still insist that “the Russian system succeeds because, in contrast with the Western industrial economies, it makes full use of its manpower.” Economists who discerned the miserable realities of the planned economy, such as G. Warren Nutter of the University of Virginia, were few and far between — almost as rare as historians, such as Robert Conquest, who grasped the enormity of the Soviet system’s crimes against its own citizens.

We know now how wrong Samuelson, Galbraith et al. were. After 1945, according to the late Angus Maddison’s estimates, the Soviet economy was never more than 44% the size of that of the U.S. By 1991, Soviet GDP was less than a third of U.S. GDP.

China has of course learned lessons from the Soviet experience. Beginning in the late 1970s with Deng Xiaoping, China’s leaders understood that the Communist Party could harness market forces for the perpetuation of their own power, but they must never relax the party’s political grip. If there is one thing the CCP can be relied on never to produce, it is a Chinese Mikhail Gorbachev.

In the same way, the Chinese have learned from the American experience. I remember vividly how, in the wake of the 2008 collapse of Lehman Brothers, eminent Chinese economists visited Harvard (where I taught at the time) and doubtless many other institutions to research the causes of the global financial crisis. Somewhere in President Xi Jinping’s office there must be a copy of the report they subsequently wrote. If there is another thing the CCP can be relied on never to produce, it is a Chinese Lehman Moment.

Yet, as the great English historian A.J.P. Taylor once observed of the French Emperor Napoleon III, he “learned from the mistakes of the past how to make new ones.” As I contemplate Xi, I find myself wondering if the Communist Party has inadvertently produced a Chinese version of Napoleon III, whose reign was also marked by rampant real estate development. (The Paris you see today was in large measure the achievement of his prefect of the Seine, Georges-Eugene Haussmann.)…

…The People’s Bank of China has already taken action. On Thursday, it sought to alleviate the financial stress with the equivalent of $17 billion in the form of seven- and 14-day reverse repurchase agreements, its largest open-market operation since January. Evergrande shares in Hong Kong duly rallied. Crisis over. Stand down the plunge protection team.

All this goes to show that a Lehman Moment was never in the cards. China’s state-controlled financial system has state-controlled crises, which are targeted at particular firms “pour encourager les autres”— not to trigger the kind of generalized bank run that drove the global financial system to the point of collapse in the winter of 2008-2009.

Nevertheless, it is possible to avoid financial contagion without necessarily avoiding a more insidious macroeconomic contagion. As the Harvard economist Ken Rogoff showed last year in a paper co-authored with Yuanchen Yang of Beijing’s Tsinghua University, real estate plays an even bigger role in China’s economy today than it did in the U.S. economy on the eve of the financial crisis. The impact of real estate-related activities amounted to 18.9% of U.S. GDP in 2005, its pre-crisis peak. The equivalent figure for China in 2016 was 28.7%. None of the 10 other countries in their sample come close, except Spain on the eve of the financial crisis (28.7% in 2006).

The detail is eye-popping. In all, around 27% of Chinese bank loans come from the real estate sector. Real estate is the main form of collateral for loan securitization. In 2017, almost 18% of the urban labor force was employed in real estate and related industries. In 2018, the sale of land by local governments accounted for as much as 35% of their revenues.

Much as happened in Japan in the housing bubble of the late 1980s, the market value of China’s housing stock is now more than double that of the U.S. and triple that of Europe. This means that housing wealth forms a significantly larger share of overall assets in China (78%) than it does in the U.S. (35%). Rogoff and Yang conclude that Chinese households’ consumption is therefore “significantly more sensitive to a decline in housing prices” than that of their American and Japanese counterparts. A “20% fall in real estate activity could lead to a 5-10% fall in GDP, even without amplification from a banking crisis, or accounting for the importance of real estate as collateral.”

To put it simply, China’s growth has been boosted for many years by the construction of an excess supply of housing units. This has been financed by an unsustainable mountain of debt. As the Beijing-based economist Michael Pettis noted last week, “China’s official debt-to-GDP ratio has soared by nearly 45 percentage points in the past five years, leaving it with among the highest debt ratios for any developing country in history.”

5. Dangerous Feelings – Morgan Housel

The feeling of mastering a topic, particularly if that topic adapts and evolves.

The first law of hard work is that you expect there to be a payoff. How could it be any other way?

But a dangerous feeling occurs when you want the payoff of years of hard work to be an assumption that you’ve mastered a topic. Or that you don’t need to update your views because you already spent years of hard work learning those views.

You see it all the time in so many industries. Veterans fall behind the younger generation because if veterans admitted that they had to adapt to what the younger generation is doing they’d feel like the hard work they put over their career was for nothing.

Even if you know your field evolves, the idea that what you learned in the past may no longer be relevant is so painful that it’s easy to reject. The longer you’ve been in a field the truer that becomes. It’s hard for a 50-year veteran to admit that a rookie might know as much as he does. But if what the veteran learned 30 or 40 years ago is no longer relevant, it can be true. And the rookie may be more aware of what he doesn’t know, while the veteran is iron-clad sure of his beliefs because he’s worked hard and expects a payoff.

Some things never change, and learning them in one era can help you in the next. But the more your field evolves – the more it involves people’s decisions – the smaller that set of learnings is, and the more you need to fight the urge to think that your long-term experience means you now permanently understand how the field works.

6. Axie Infinity faces big test as player earnings fall – Derek Lim

Lately, things have not been great for many Axie Infinity players, most of whom play the game solely to make money.

The value of small love potions or SLP, the in-game currency that players exchange for cash, has plunged from US$0.35 in mid-July to the current price of US$0.059. Prices have not recovered despite recent tweaks to the game’s economy.

“Earning US$100 every 15 days is not that substantial at all,” says Peter Villagracia, an independent Filipino Axie Infinity “scholar” who used to earn more than twice that figure in the same amount of time. Scholars are players who can’t afford to own axies – the digital monsters in the game – so they rent them from “managers” under a profit-sharing model.

For Althea Torres, another independent Filipino scholar and a single mother of three children who relies on the game to support her family, the change is more drastic.

She began playing Axie Infinity full time at the start of May 2021 because it allowed her to bring in more money while spending more time with her children. Before that, she was working at a small roadside vegetable and fruit stall, earning between US$5 to US$7 daily for a hard day’s work. At the game’s peak, she could make between US$15 US$20 per day, but now she only gets around US$6 a day.

“I didn’t realize it, but this game became such a huge part of my life. In fact, it became my only source of income because other jobs just couldn’t match what I was earning while playing Axie,” she tells Tech in Asia.

Torres adds: “When the price of SLP fell, it became really hard for me to survive because my earning power dropped by so much. I have to pay rent, feed four people every day, and buy other necessities that we have to use in our daily lives. It’s scary because I don’t know how I am going to keep providing all these for them.”..

…As we discussed before, the health of the game really hinges on balancing supply and demand for SLP.

When the supply of SLP outstrips the demand, the token will likely lose its value, causing a downward spiral as players are no longer as motivated to play the game.

It seems that this scenario is playing itself out right now, with far more SLPs being minted than burned. “Burning” refers to the act of spending the tokens, which then results in the tokens being deleted forever.

“Because of the fact that breeding has always been so profitable, managers will simply keep breeding axies to maximize their profits, before allocating bred axies to scholars who will then mint even more SLP with them,” a manager who wanted to be called by the moniker Precision tells Tech in Asia. “This will really cause the supply of SLP to grow exponentially because almost every manager will be doing this.”

It seems that this delicate balance between supply and demand was not achieved. As Precision observes, “The value of battle gameplay here is eroded through a lack of burn channels, as well as a flawed game design that doesn’t create enough demand for SLP.”

The manager adds: “The game’s initial failure at preventing bots and whales from farming SLP at an incredible rate is also a factor in my opinion, because this caused a huge pump of the supply of SLP.”

Demand for SLP is created simply by giving players more ways to spend or burn them.

“I think the main problem really lies in the fact that there has been no expansion or extension of the current gameplay to give SLP more intrinsic value. Right now, it only really has one use case, which is to breed axies, so the whole ecosystem is fragile,” notes Coby Lim, co-founder of crypto startup Fincade who’s also an Axie manager.

“Sure, everything takes time, but I think this should have been factored in and prioritized by the team from the start. They must have seen it coming,” he says…

…Axie Infinity is a double-edged sword for many of its Filipino players, who make up a huge proportion of the metaverse. On the one hand, it provides them with an alternative income. On the other, this may create an unhealthy dependence that puts them at the mercy of the game’s developers.

Because managers are incentivized to bring in as many scholars as possible, scholars may not be aware that the income they’ve earned in the past may not hold steady in the future.

Furthermore, while managers can simply write off their losses and invest in something else, scholars rely heavily on the value of SLP to survive on a day-to-day basis.

Axie manager Chew argues that the long-term viability of the game’s model needs to be scrutinized.

“Yes, it is admirable that the founders have [changed] the lives of many by [helping them] bring food to the table. But I feel that the main question that they should be asking themselves right now should be how and whether this model can be made sustainable in the long term,” he says.

“They may be trying to do that, but I think for many of us who are watching keenly, it doesn’t seem to be going down that road, and that spells trouble for these people who really need the game to be [sustainable].”

7. Scaling to $100 Million – Mary D’Onofrio and Ethan Ding

When it comes to building and scaling a cloud business, founders, CEOs, CFOs, and board members alike want to know what “typical” and “best-in-class” look like. Leaders, like you, want to model their businesses around these benchmarks to achieve their goals.

There is a problem, though. Private market financial benchmarks are some of the most elusive financial data points in the world. They are also some of the most helpful. If you’re a cloud startup seeking to emulate the success of companies like Shopify, Procore, and Twilio, understanding how your predecessors grew and achieved key milestones is a critical part of the equation. But not everyone has access to this type of information. Private companies lack reporting requirements that would make their benchmarks known, and backers of private companies hold their portfolio company information close to the chest. Considering these factors, only the highest-flying, venture-backed companies have the opportunity to learn from the stories of the past, leaving other startups at an inherent disadvantage—until now!

We’re releasing “Scaling to $100 Million” as the industry’s definitive benchmarking report for cloud companies looking to scale to new heights. For more than a decade, Bessemer has made over 200 cloud investments and has one of the largest cloud portfolios of any venture firm in the world.* As we share this information with leaders like you, we hope this body of analysis proves to be a valuable resource for what growing your cloud business looks like at every stage…

…Examining Bessemer’s cloud portfolio over the last decade, we find that the expected growth rate for companies decreases over time, as it is easier to grow at a higher rate on a smaller base of revenue and the marginal dollar is always harder to acquire. The average growth rate for companies between $1-10MM of ARR was nearly 200%, and this average decreases to 60% for companies over $50MM of ARR. We also find that the middle 50% of cloud companies have a tighter and tighter band of growth rates as ARR scale increases: the middle 50% of companies from $1-10MM of ARR are growing from 100-230% while the middle 50% of companies from $50MM+ of ARR are growing from 35-80%.

While there is some selection bias for companies that are at the higher ends of the ARR range (the companies that make it to that scale are the most successful ones), an important note is that average growth rates continue at high rates, even at scale. We find that this tends to happen because of two reasons.

First, by $50MM or $100MM of ARR, the Cloud Giants are crowned. Given the virtuous cycle of market leadership, the leaders that emerge are able to further consolidate their markets, accelerating growth. For example, when Bessemer first funded PagerDuty in its Series B in 2014, it was at $12MM of ARR and had material competition from VictorOps, OpsGenie, and xMatters. By the time PagerDuty crossed $100MM of ARR in 2018, all of these competitors had either been acquired or fell behind, leaving PagerDuty as the only true standalone company in the incidence response category and allowing it to capture more of the pie.

Second, market leaders tend to accelerate their growth and expand their total addressable markets (TAM) by adding “Second Act” products, so even if there is growth decay in the core product, there are constant second, third, and even fourth winds behind company growth as a whole. Cloud leaders tend to be multi-product companies. For example, our portfolio company Toast has successfully layered Payments and Capital onto its already-large point of sale (POS) business.

Examining our cloud company data, we also note that it is very rare to see a best-in-class growth rate company quickly devolve into a laggard. Similarly, it is very rare to see a mediocre grower evolve into a high-grower…

…Cutting the data by industry rather than ARR range, we find that gross retention largely hovers in a similar range but net retention varies much more across industries. Developer tools have the greatest average and median net retention rates across our portfolio, in line with what we would expect from a bottoms-up sales strategy that expands seat count and usage as it permeates an org. Collaboration software shows a similar dynamic. Though there are exceptions, industries such as sales and marketing software, customer experience software, and finance / legal tech tend to have lower net retention, likely because land ACVs are higher and expansion over time is lower (often these tools sell a complete platform, rather than individual seats or usage tiers)…

…The beauty of software is that there is practically $0 marginal cost to replicate and distribute it. Gross margin, a company’s revenue after the cost of goods sold (or gross profit) divided by revenue, is an incredibly important metric for cloud companies because it measures the effectiveness with which companies can deliver their software to their customers. The aim is to make it as high as possible, reflecting the lowest marginal cost. A high gross margin means that a cloud company can invest more into operating expenses rather than product delivery, leading to more selling, product iteration, and ultimately, growth. Typical expenses that you will find in COGS for cloud companies are hosting costs, software implementation costs, and services costs, including customer success—these are all variable costs.

Given that the marginal cost for delivering software should be very low, investors expect gross margins for cloud companies to stay within a fairly tight band. It is perhaps the only operating or cost metric that has very little wiggle room—the average gross margin for a cloud business regardless of maturity is 65-70%, and the distribution of the middle 50% stays within ~60-80%.

That said, some of the strongest cloud companies in our portfolio have been ones with gross margins below that range. For example, throughout much of its life in the Bessemer portfolio since the seed round in 2009, Twilio’s gross margin was ~50%, which accounted for the fact that it had to pay telecom service providers in its COGS. Twilio continues to be one of the strongest BVP Nasdaq Emerging Cloud Index performers today with a market capitalization of over $60 billion…

…When looking at burn for a cloud business, we want to consider it in the context of growth. Burning $100MM a year sounds high, but what if a company burned $100MM and added $1 billion of net new ARR? In that context, it doesn’t sound so bad. As this hypothetical suggests, investors look at the cash consumption of a business relative to the revenue that it generates, which is why the efficiency score becomes a helpful metric. Efficiency score equals FCF margin of ARR plus ARR year-over-year growth rate—as such it helps to show the tradeoffs between growth and profitability, but it is generally only applicable after achieving $25MM+ of ARR (before which revenue bases are too small for it to be meaningful). We encourage Bessemer portfolio companies to target 70% efficiency scores between $25-50MM of ARR, and a slightly lower 50% at $100MM+ as YoY growth rate drops off dramatically and companies find the right balance of profitability against a “grow-at-all-costs” model.

Efficiency score = FCF margin of ARR + ARR YoY Growth Rate

Younger companies tend to have higher growth rates and higher burn rates, and companies at maturity have lower growth rates and lower burn (and sometimes cash flow positivity). The “Rule of 40” is often referenced—that companies should have efficiency scores of 40%+ – but the average BVP Nasdaq Emerging Cloud Index efficiency score is actually closer to 50%, anchored up by the likes of Zoom, Shopify, Datadog, Crowdstrike, and other high performers. For example, even at over $3.8 billion of LTM revenue, Shopify is still growing ~60% YoY with ~10% FCF margins for an efficiency score of close to 70%.


Disclaimer: None of the information or analysis presented is intended to form the basis for any offer or recommendation. Of all the companies mentioned, we currently have a vested interest in Alphabet (parent of Google), Amazon, Apple, Datadog, Facebook, Shopify, Tesla, Twilio, Zoom. Holdings are subject to change at any time.

Ser Jing & Jeremy
thegoodinvestors@gmail.com