Exclusive Manager Interview on Facebook

Christopher Devlin founded Selective Wealth Management in 2009, after a brief stint as a nuclear engineer. The firm's investment methodology focuses on investing in companies with six common characteristics: 1. Lasting Products of Services, 2. Durable Competitive Advantage, 3. Highly Profitable, 4. Prudent Debt Levels, 5. Skillful Reinvestment of Earnings, and 6. World Class Management.

In the almost a decade since he began managing outside money, he has substantially grown the firm. Devlin recently submitted a writeup on Facebook to SumZero. Kevin Harris from SumZero sat down with Devlin to discuss value investing, the founding of Selective WM, and Facebook.

Kevin Harris, SumZero: During your time on SumZero, you’ve topped our rankings in the Growth category. How would you describe your investment philosophy and process?

Christopher Devlin, Selective Wealth Management: Our investment philosophy focuses on purchasing the very best businesses in the world. We are very selective and only consider companies that pass our rigorous Selective Process. In order to be considered a Selective Company a business must possess all six of our “Selective Pillars”. This means the company has an enduring product or service, is highly profitable, demonstrates a sustained competitive advantage, has prudent levels of debt, skillfully reinvest earnings, and is led by world class leaders. By definition only a few businesses can be industry leaders – which forces us to be selective. We focus on our best ideas and run a concentrated portfolio – typically owning anywhere from 10 to 20 businesses at any given time with the majority of our assets in our top 5 ideas.

We believe this selective and concentrated approach that focuses on quality is tightly aligned with the investment philosophy of some of the most successful investors that have come before us.

Harris: What led you to found Selective WM? What is the story behind Selective’s launch?

Devlin: My undergraduate degree was actually in nuclear engineering. My first day on the job I sat down and built a spreadsheet to figure out how long I’d have to work before I retired. After building out that financial model I quickly realized the rate of return on my investments was by far the most dominant variable for my future net worth. This variable was more important than how money I started with, more important than my earning power, more important than my savings rate. If I could compound at above average rates of return it would really help me achieve my dreams. This sparked my interest in maximizing ROI. As I began to learn about the investment process I completely fell in love with analyzing businesses. As a professional investor you have the flexibility to be infinitely curious. Business intersects with almost all disciplines and all areas of life, which makes it the perfect profession for a life-long learner. I became so passionate about business and finance that it compelled me to become a professional investor. I think that passion is a critical ingredient to our success, because passionate people tend to produce significantly more output in almost any field when compared to those who approach their profession as a ‘job’. I really love what I do.

Harris: Selective’s 2017 Annual Letter details the extraordinary amount of cash (50%+) you held throughout the year. With the focus on high conviction ideas, why the strong concentration in cash?

Devlin: At Selective we view stock ownership as business ownership. We are of the belief that purchasing high quality companies takes time – both in the research process and waiting for an attractive price. In most years we would anticipate building on our existing collection of Selective Companies and only need to purchase between 1 and 5 companies to remain fully invested. However, in 2017, due to the change in investment structure we held nearly 100% cash in early February and would need to make an unusually high number of purchases to become fully invested.

We worked diligently at the task and made ten purchases throughout the year, but even at this above average pace we ended the year with 28% cash. It was our expectation that our cash balance would be less in 2018, but in the first half of the year our two largest holdings, MuleSoft and Syntel, were both acquired at large premiums. While this increased investment returns, it also converted wonderful businesses back into cash. Even after making new purchases in 2018 we are close to averaging 50% cash for the second year in a row.

The inadvertent large cash balance over the last two years has caused us to reflect carefully about the impact cash has on a portfolio. Somewhat paradoxically, we believe that cash is simultaneously an excellent asset and a poor long-term investment. It’s poor in the sense that it typically earns less than the rate of inflation and is certain to lose real value overtime. For this reason, we vastly prefer the ownership of Selective Companies. However, there are benefits to maintaining a cash position when the market is expensive.

The goal of our cash position is to maintain the ability to make opportunistic purchases. Some might argue that fixed income would be a good alternative, earning 1-2% more than our Treasury Money Market Fund. However, we don’t believe the slight increase in yield is worth the reduction in liquidity. When great opportunities present themselves – speed is paramount. During a market dislocation or panic the ability to move quickly can add many percentage points in an instant. The gain experienced by holding ample cash during such a panic does not appear in an annual report or on an “attribution analysis”. The gain is instead ascribed to whatever asset was purchased…however, much of the credit should go to the cash.

By way of example, on August 24, 2015 the New York Stock Exchange opened sharply down due to concerns that growth was slowing in China. It was the largest single drop in market history at open. During the first 10 minutes of trading some of our long coveted Selective Companies went on sale with prices down more than 30%. We were able to act quickly and purchased several Selective Companies. As the prices recovered we earned a far higher return than what was forgone in fixed income. The speed of cash was valuable in an instant. Moments like that have occurred regularly over our investment career, but the timing of such events are unknown. Patience is required to fully appreciate the cash position. We encourage clients to view our cash position as “dry powder” – ready to buy Selective Companies at any moment.

Harris: Slightly over a year ago, you wrote up Facebook on Selective’s blog. How have your thoughts evolved on the company since?

Devlin: That’s a great question. In 2017 I wrote a blog post attempting to figure out when the break-neck pace of Facebook’s revenue growth would stop. It was a mathematical impossibility that the current growth rates could be projected out much farther into the future…the population of the planet just wouldn’t support it. We concluded in that blog post that revenue growth was going to decline substantially…and it would occur in the near future (less than 5 years). At the time the blog was published we owned no Facebook stock.

About a year later the Cambridge Analytica scandal occurred and we opportunistically purchased the business. When we made the purchase the company was compounding revenue above a 40% rate, but consistent with what we wrote a year earlier – we suspected that growth would slow in the near future. We didn’t know exactly when, but we projected growth rates of about 15% over the first 5 years of our ownership and then high single digits after that. So why did we purchase in 2018 and not 2017? The company was 50% bigger between our blog post and the purchase date… but the price hadn’t moved an inch. This dramatic increase in size for the same purchase price made it more attractive to us.

Shortly following our purchase, Facebook had an infamous quarter where the company announced growth would slow dramatically…down to 20% by year end. This certainly was a huge surprise to Wall Street as the stock experienced the largest single day decline in market history (in absolute dollars). That being said, it was still above the price available during the Cambridge Analytica scandal. Where we surprised that growth was decelerating? No. Here are a few excerpts from our research report written several months prior to the infamous conference call:

“In the last five years revenue has compounded at a rate of 50% per year from $8 billion to $40.6 billion. We anticipate that revenue will continue to climb between 5% and 25% for the next 5 years with an overall average of 15% per year. At that point in time we anticipate revenue growth will fluctuate between 3% and 12% with an average rate close to 9% for the ensuing 5 years.”

Why did we take this view? More from our report:

“The average number of users is measured by counting the number of individuals globally that access their account over a specified time period. The number of average users is typically reported as Daily Average Users (DAU) and Monthly Average Users (MAU). The company currently has 2.1 billion MAU. To put this figure in perspective, consider the following:

The earth’s population is 7.5 billion
The total number of internet users is 3.5 billion
China has a population of nearly 1.4 billion with 771 million internet users.

These individuals are currently discouraged from accessing Facebook and are blocked by a firewall (in theory).

3.5 billion - 771 million = 2.7 billion people eligible for Facebook accounts.

77% of eligible internet users worldwide access Facebook at least once per month.
This is a truly remarkable figure and several key pieces of information can be gleaned from this statistic. The first is that people like using Facebook. When more than 3/4th of the entire world’s internet connected population uses a service it is evidence that this service is resonating with a part of human existence that transcends age, race, gender, and culture. Despite the challenges that Facebook faces with security, fake news, protecting users, removing inappropriate content, etc. one thing is very apparent. The majority of people derive some type of value utilizing Facebook.

We can also infer that Facebook’s user base will grow at a much slower pace than it has historically. The limited number of people that are even eligible to open an account is relatively low compared to 5 years ago. It is a near impossibility for double digit growth rates to continue long into the future. We anticipate the number of MAU will fluctuate from -5% to +10% over the next 10 years, with an average growth rate of 5%. We believe this growth will be fueled by a growing world population, increasing internet penetration rates, and the adoption of Facebook accounts rising slowly from 77%.”

To summarize, I would say that our thoughts haven’t changed too much since our blog post.

What has changed is that Facebook’s growth is now slowing. Do we wish the slow-down occurred a bit later? Sure. Does it impact the decision we made to purchase Facebook when we did? Not all. I still believe we paid a below average price for an outstanding Selective Company that will experience growth above the average S&P 500 constituent.

Harris: Your FB idea mentioned that you were surprised at FB’s ARPU growth. Net of under-monetised Whatsapp, where do you think that FB /Instagram ARPU will plateau?

Devlin: Another excellent question. I’m going to start by saying right out of the gate “I don’t know.” It seems odd to purchase a business where such a critical metric, such as when ARPU plateau’s, is unknown. However, in this instance I don’t think you need to know the answer to that specific question. What’s very important is that we are directionally correct about ARPU over the next 10 years. We are confident that ARPU will go up globally. It likely will compound at a much slower rate in North America and significantly higher rates in the APAC regions. When we estimated ARPU growth we used very modest assumptions (single digit assumptions in North America and slightly higher elsewhere) and still believe that Facebook would grow faster than an average business. For this reason we felt comfortable making a purchase without knowing specifically when ARPU would plateau. The word ‘plateau’ is also a bit misleading. In this instance I think the ‘plateau’ for ARPU is still compounding annually at a single digit rate.

This could be accomplished when monetization is mature across the entire eco-system (Facebook, Instagram, Whatapp) due to several tailwinds – such as increasing bids for similar ads each year due to inflation, better conversion rates, superior targeting, etc. Engagement could even go down and ARPU continue to rise.

Big Picture: I think we are directionally correct on our ARPU estimates over long periods of time.

Harris: Facebook core seems to be running into a natural growth ceiling based on their saturation of 75%+ of the global internet user population (1.47B DAU’s, 2.23B MAU’s). Will the inevitable slowing of user growth from double digit year on year impair revenue growth?

Devlin: Of course. Revenue growth is a product of user growth and ARPU. If user growth slows the compounding decelerates.

Harris: A recent New Yorker profile of Mark Zuckerberg aptly summarized the change in public opinion towards him and other billionaire tech founders as: “formerly valorized as the vanguard of American ingenuity and the astronauts of our time… [are now] being compared to Standard Oil and other monopolists of the Gilded Age”. Could this global shift in public opinion catalyze punitive regulations towards future acquisitions or content monitoring policies?

Devlin: Before getting to the question I think it’s important put some perspective on public opinion here. The comparison between our tech monopolists today and the robber barons of the Gilded Age seems quite unfair. During the Gilded Age nine percent of Carnegie's workforce died annually working in his steel mills. Monopolies created terrible working conditions, underpaid individuals, and had the capacity to overcharge customers. Compare that to our tech monopolies today.

They create some of the highest paying jobs in the world and are consistently ranked as the best places to work year-in and year-out. The services they (Facebook / Google) provide have amazing utility and are free for consumers. This is a very important difference. I believe many of these tech titans have been able to reach their monopolistic scale and global dominance because of that disconnect between those who consume the service (internet users) and those who pay for it (advertisers). It has allowed them to scale to unpresented levels without so much as a peep from consumers. Recently however, the complacent consumers are starting to voice concern.

The shift in public opinion over the last few years has been dramatic, both here in the US and abroad. It will definitely catalyze punitive regulations in many areas surrounding Facebook. Do I believe it is possible to forecast what those new regulations will be? No. Do I think they’ll have a negative effect on Facebook’s business model? Probably. But I also believe those impacts will be short-term. Facebook will find discover how to improve monetization their global platform in a way that users enjoy and appreciate overtime. Sometimes Facebook will mis-step and upset people, but that is bound to happen with nearly 3 billion people utilizing your platform daily. It’s not easy to keep 3 billion people happy. If you make a mistake that impacts 3% of the user base you have 90 million angry people on your hands (I believe the number impacted by Analytic was less than this). Normally a 97% is a wonderful accomplishment that allows you to graduate with honors…but with a business the size of Facebook a 97% pisses off a lot of people.

Do I think it’s possible to run Facebook perfectly? No. They’re going to make mistakes – like mishandling sensitive information, not catching fake news, not monitoring ad sponsors close enough at election time, etc. When you run a business as big and as important as Facebook it’s not possible to see around every corner. Do I think they’ll do better in the future? Yeah, probably. Governments around the world certainly got their attention this year and major changes are coming to monitor how Facebook impacts the world globally. At Selective we intend to own Facebook for decades and we’ll have good years and bad years, but overall we expect the company will be very successful. We believe they are going to survive and thrive in the face of all the new regulatory rules.

Harris: Facebook’s recent 10Q mentions that “we anticipate that future advertising revenue growth will be driven by a combination of price and the number of ads displayed”. Will this represent a dilution of the user experience? How can Facebook preempt the precipitous, snowballing decline that many other social platforms have experienced?

Devlin: I doubt it. It’s easy to stay on top of the user experience when you are the size and scale of Facebook. You can rollout A/B tests of different layouts, new feed ad densities, types of ads, etc. and measure user behavior. If engagement drops as you fill the news feed in one test cohort you can adjust it. I believe that Facebook’s social network is large enough, durable enough, and important enough to withstand a myriad of planning mistakes with the user interface over the years.

I think it would be unwise to compare the snowballing decline of any other social media platform to Facebook. In my view, there are only two social media platforms on earth that have achieved scale – Tencent and Facebook. It wouldn’t surprise me at all if in 10 years we talk about how people use to tweet or snap, etc. But Facebook is different. The history of planet earth is being uploaded to their datacenters to the tune of 1 billion photos a day. And these photos matter. Consider the following story:

When I was a teenager I borrowed my Dad’s car to go swimming with some friends. We parked in a remote location off the main road and hiked through the woods to our swimming spot. While we were away my Dad’s car was broken into. A rock was thrown through the window smashing the glass and the inside was ransacked. Upon returning to the car I called my Dad to inform him about the unfortunate event and I clearly remember his response when I told him everything had been stolen from the car.

“Is the picture of your Mom and I from high school still there?”

Of all the items stolen from the car, this was the one most precious to my Dad. Much to his dismay the picture had been stolen. Unlike the other items that were missing – this item could not be purchased with money. No technological breakthrough, no amount of effort, and no amount of money would bring back the photo. It was a memory and it was precious to my father. The memories of our lives are important to us, valuable, precious, and impossible to replicate.

Now back to Facebook. The site was launched while I was in college and as a student I was one of the first to join the network. Once Facebook emerged as the clear winner against MySpace a unique phenomenon began to occur. The history of my life was almost exclusively chronicled on Facebook. There is an average of one picture for every three days of my life – and believe me, I am not an active Facebook user (most are posted by people I’m with and I’m tagged). The more time passes the more nostalgic I get looking back on old photos. When I am 80 I am going to be so grateful for my Facebook account.

In addition to the photos there are comments from friends. I recently went back and had a chance read comments from various trips I took over the years. The memories came flooding back. It was a wonderfully positive experience. The entire world is uploading the history of their lives at a pace of more than 1 billion photos per day. The same part of human existence that touched my father when reflecting on the photo of my mom is being uploaded to Facebook. If done correctly – it’s a wonderful thing. And we believe these memories cannot be duplicated with time, energy, effort, or technology.

Imagine 100 years from today. My great great grandkids will have the ability to see who I was, what I was like, who I spent time with (if I give permission to Facebook to share my account prior to my death…?). This is a wonderful service for future generations. How could a company replicate such a wonderful service? All the photos, memories, comments, stories, and effort that we’ve put into the platform for the last 14 years has created a network and a legacy that I don’t believe will be easy to move or replace. We believe the moat around Facebook is getting wider everyday.

That being said, short-term data shows declines in user numbers for the youngest cohorts. This should be expected. Facebook becomes more interesting for people as they get older. As you age and mature you are posting pictures of your wedding day, your first child, your parents holding grandchildren, etc. You spend time staying in touch and looking at the lives of people that use to be very important in your life, like your brothers or friends from college, old work colleuges, etc. When you’re in high school you live with your family, you don’t have many friends that are scattered across the world, and you’re too cool to stay in touch with Mom and Dad. SnapChat makes way more sense for this young cohort. You can send inappropriate and temporary images as you discover who you are. I wouldn’t expect Facebook to ever really dominate the youngest cohorts, but I do expect that as this cohort matures many of them will spend less time on SnapChat and more time on Facebook. Priorities change overtime and Facebook definitely plays a critical and positive role in the world today.

Harris: What is your view on the Q4’17 quarter-over-quarter decline in DAU’s in North America? Could this be a canary in the coalmine regarding a broader global decline in Facebook’s mature markets?

Devlin: The Q4 2017 decline was less than 1%. This market is very mature and almost all internet users of age have Facebook accounts. This makes growth nearly impossible. In a market as mature as North America I’d focus less on year-over-year growth rates and more on account penetration across all internet users. After all, if the population declines users will drop, but there is nothing wrong with the business model. Looking at it this way penetration is largely unchanged over the last 3 quarters and users have maxed out at about 185 million for the continent. It was briefly 184 million in Q4 2017, but is back up to 185 million. We believe it’s silly to fret about this when they have fully penetrated the market and the 185 million is an estimate anyway. It’s probably not even within 1% accurate given the number of fake accounts that Facebook battles daily. Despite all the drama that unfolded in the media, the Q4 2017 user numbers for North America should be considered noise.

Harris: How has Facebook’s last twelve months most surprised you? Where have you been most right and most wrong?

Devlin: The biggest surprise in the last twelve months was not that user growth is slowing or revenue growth will decelerate. These were largely known based on our research, we just didn’t know when it would occur. What did surprise us is that Zuckerberg instead to greatly increased spending by hiring additional personnel. For the last 5 years Facebook has demonstrated exceptional operating leverage and we expect this to continue with additional margin expansion as revenue climbed. I think the congressional hearings really shook Zuckerberg and now he intends to hire for security, data privacy, content management, etc.

While I am a huge fan of striving for the best possible user experience I’m very curious if arbitrarily increasing headcount and expenses will achieve superior results. I’m not sure if 10,000 programmers dedicated to improving privacy and security perform better than 5,000. Software solutions often scale very well and the largest contributions come from your very best programmers. It surprised me that Zuckerberg seemed to arbitrarily pick new operating margin targets over the next 3 years. If this improves the quality of the network – wonderful. If not, we should really consider if it’s a good idea. This development on the cost side was our biggest surprise and it looks like margins won’t expand as much as we anticipated. That being said, I think we had a healthy margin of safety when we made our purchase and the operating results of the business will still justify our purchase price over the next 5 years, but only time will tell.

Harris: What role do online research tools like SumZero play in your research process?

Devlin: SumZero plays a central role in our research process here at Selective. Over the years we’ve utilized the research database to source new ideas and to look for dissenting views on our prospective long ideas. One of the first things I do when considering the purchase of a new business is look for short-thesis on the company on SumZero. Having access to this network of professionals helps us get up to speed on a specific name significantly faster than if we started from a blank slate. It’s a wonderful compliment to our in-house bottom up approach.

Harris: What books or investors have most impacted your investment and valuation philosophies?

Devlin: I’d say most of the usual suspects played a role in shaping my investment philosophy – which include Graham & Dodd, Buffet, Munger, Klarman, Lynch, Greenblatt, and Fisher. Then some non-investors like Jim Collins and William Thorndike. But if I had to name one that really impacted me I would say Nassim Taleb.

In 'Fooled By Randomness' Taleb introduced this idea of alternative histories which hit me like a ton of bricks. Alternative histories was this idea that the reality which plays out in our lives is not the only possible outcome. There are substitute courses of events that also could have occurred, but didn’t – simply by chance. It made me think about how if someone repeatedly made the same decisions over and over; what possible outcomes could have occurred? What would their lives look like in all alternative histories?

What’s interesting about this intellectual framework is it forces you to not look at how things are, but also how they could have been. For example, I became obsessed with analyzing the track record of investment legends to determine if they were lucky or good. What happens to Soros track record if the British Pound doesn’t collapse? What happens to David Einhorn if he didn’t earn 100x on a small investment in an internet stock just prior to the tech bubble? What happens to Bill Ackman’s record without General Growth Properties? Where these good ideas… or just good outcomes? Did they take on too much risk and just barely avoid disasters? Perhaps some legends aren’t as good as we think and just were fortunate that the path that was actualized was in their favor. If they made the same decisions over and over across 1,000 different lifetimes…what would their lives have looked like across all 1,000? In this life they may be famous, but if they behaved the same way perhaps 999 times out of 1,000 they don’t succeed. Or, conversely, who produced average due to bad luck, but had world class decision making? Overall, I think this framework really made me think critically about decision making.

Additionally, this shifted my priority from wanting to have the best performance to being the best investor. These are very different things. The best performance will be whoever is fortunate enough to have a path actualized that combined both risk and luck. The best investor would be whoever makes the best decisions considering all possible alternative histories. Maximizing the results across all possible alternative histories is something that is very important to me. And I also very carefully think about my own investment performance. Was the decision making that led to the results well founded? What other histories could have been actualized? Did you succeed, but did something stupid? Are you lucky or good? These are really tough questions – and Taleb got me to think about these types of things constantly.