The Search for the Next Tencent

(Boyu Hu, December 2017)

Hello everyone! Thanks again for coming to the first annual conference of XVC. Leo and Rickey just gave you a glance at our approach for “Marketplaces” and “Social Networks.” Next, I’ll talk about our investment strategies and overall approach to early-stage investment.

In 2016, I published a blog post titled “A Venture Capitalist's Personal Stock Investments.” A chart in that article shows the performance of my personal stock account from early 2012 to April 15, 2016.

The red line represents net gain, and the blue line represents principal. In the end, the net gain became 4x the size of the principal, so I made 5x on the initial investment.

Looks great. However, there is a small regret. Why? As you can see, when the profit (the red line) was close to the size of the principal, I needed money and withdrew the initial investment, leaving only the profit in my account. After that, I withdrew another two times, a total amount that equals 3x of the principal. As you can imagine, if I had kept the money in the account, the red line would become steeper, and the net gain could have doubled. Therefore, I paid a big price for cashing out too early.

Some of you might wonder how the portfolio would perform in the following one-and-a-half years.

If the portfolio had remained unchanged, it would have been worth 10.8x of the initial principal as of last Friday (December 15, 2017). Unfortunately, in the few weeks after April 15, 2016, I sold all the stocks because my family needed the money. So, nothing actually happened after the “circle” area.

This is my first takeaway: to become an investor, you need to have capital.

Without your trust and support, we would not have the money to invest. I would like to say thanks to you all.

We have only made a few investments, so we thought that maybe we should skip the annual investor meeting this year and do a conference call instead. But then I realized that it’s a chance to get together and exchange ideas. So, we decided to arrange this simple gathering on our little island, in our own meeting room.

Today, I want to talk about how we make VC investments using the approach that I trade stocks with.

Let’s take a look at this chart first. This is a company I had been holding for a long time.

Some people might find it familiar. It’s Tencent. Since its IPO in 2004, Tencent’s market cap has grown 590x to $473 billion, implying an IRR of 63%, before taking the dividends into account.

The way I trade stocks can be concluded in one sentence: “Find a small number of great companies like Tencent and hold them for a very long time.”

Before we turn to where and how to find great companies like Tencent, let’s talk about the overall strategy first. I did not spend much time trading stocks. For most of the time, I only held four to five stocks, and I only change one or two each year, leaving the portfolio alone most of the other time.

The first mindset behind the approach is “long-term value driven.”

I believe “long-term value driven” is a structural advantage. Why?

Taking a long-term view, stock prices always reflect the value of the companies, but in the short term, the stock market is like a “voting machine,” and the changes in stock prices have little to do with the intrinsic value of the companies and merely reflect the results of the votes that the buyers and sellers made with their money.

Let’s look at the stock price chart of Tencent. Overall, it looks like a perfect exponential function. But if you cut it into small periods, you’ll see ups and downs in every period.

I did a calculation: out of the 3,353 trading days after Tencent went public, only 1,684 days saw its stock price rise – roughly 50% of the days. So, for every single day, the ups and downs of stock prices are quite random, and, if you try to predict its price movements, it is quite hard to make money. But, if you ignore the movement and hold the stock for a long time, you can make a lot of money.

We know that Tencent had two series A investors who each invested $1.1 million, with each owning 20% of the company. One year later, someone offered to purchase their shares at 10x the original price, and these two investors sold their shares over time. The sale of the Tencent shares might be their fastest money-making deal, but it might also be their most regretful decision.

To focus on the long-term value of the companies is a simple idea, but why can’t most market participants resist the interference of short-term fluctuations?

I have discussed this topic in my blog post “A Venture Capitalist's Personal Stock Investments.” It’s in our handouts, and I believe most of you have read it before, so I will not discuss the details.

At first, I only had Tencent in my portfolio. Later, to “diversify,” I lowered my position in Tencent and bought stocks like Tesla, QIWI, TAL and JD, which made a 2-3x return on average. In retrospect, if I had only held Tencent, the total return will be much higher. So, I paid a huge price for diversification.

Looking back ten years from now, investors would have been better off if they had simply put all their money in Tencent’s stock and held it instead of investing in any funds. Tencent has set a really high benchmark for all fund managers.

Seriously, how many investment portfolios can beat Tencent on long-term yield? Not many, I would guess. However, I happen to have found one.

Yale Endowment published the 20-year IRR of its VC portfolio. It surprised me with the number 77.4%. I had a chat with them, and I found that their VC portfolio is highly concentrated and stable – their philosophy is to build very long-term relationships with a very small group (mid-teens) of talented GPs around the world.

Why do they invest in this way? Let’s take a look at two sets of data.

The left chart shows what Cambridge Associates found through tracking the VC industry over the past 30 years – on average, VCs’ returns have been similar to the S&P 500, but “Top VCs” performed much better than the S&P 500 during most of this time.

The right chart is from Wealthfront. They studied 1,000 VCs and found that the top 2% made 95% of the money. So, top VCs generate good returns, and the 2% super top VCs generate extremely good returns.

Now let’s look at the exit data of a few “super Top VCs.” Sequoia has invested in more than 600 companies in the past 20 years, of which, 8% had IPOs and 20% got acquired, but most got nowhere. For Accel Partners, the numbers are 6% IPO, 32% M&A, and 62% had no exits. Benchmark Capital, another exceptional VC, had 10% of its investments in last 20 years undertake an IPO, 35% were acquired, and 55% went nowhere.

This is a chart I found on the website of Andreesen Horowitz. It shows that funds with the highest returns also have a higher “loss ratio” than average funds with normal returns.

From these data, we can get a couple takeaways. The first is that, instead of investing in a large group of diversified VC funds, investors should find a small group of really good ones and make long-term, concentrated investments. Some of our investors have been in VCs for 40 years - top university endowments, family offices and fund of funds in the U.S., Europe, and Asia. Many of them are significant and long-term LPs of Sequoia, Benchmark and Accel. They embrace a similar investment philosophy: finding a few top fund managers around the globe and investing in a concentrated, long-term way.

Another takeaway is that, even for these top VCs, there is great uncertainty in each deal.

Only 8% of Sequoia’s portfolio companies went public in the last 20 years. With XVC’s first RMB fund, we plan to make concentrated investments in 10-15 early-stage companies. If we can perform as well as Sequoia, one of the companies will go public. Does this sound scary?

Frankly speaking, I was once quite anxious about starting my own fund. I have had very good luck in my career. The first company I co-founded went public. The ten early-stage companies I have invested in now have an average market value of 2.4 billion USD. I worried, what if one day I lose the golden goose?

I spent a long time thinking about this question – are there scientific ways to manage “luck?”

I will share with you my thoughts.

Basically, most of the VC funds will see only 1/3 of their portfolios become successful, even if due diligence is conducted carefully and investment decisions are made seriously. Another 1/3 of the investments can return the cost, and another 1/3 will be written off. We work very hard, but it’s difficult to always beat the market on success rate due to the nature of early stage investment.

But we can choose our own battlefield. What we can do is make sure that when a company succeeds it will be a huge success. In other words, we should search for opportunities in areas and models with no glass ceiling and look for companies like Tencent.

Why is Tencent “a company without a glass ceiling?” The nature of its core product is unique – a messaging service that improves communication, a fundamental human need. Every day, 700-800 million people use its apps to send and receive messages, to obtain information, to kill time and to buy products and services. In this “virtual world,” it can not only issue policies and collect taxes like a government but also change the laws of physics like a god.

Next, I will use three companies I invested before as examples.

The first company is a social network service. It only had eight employees working in a very old and shabby apartment when I found it in 2014. It had no income nor any tangible assets, and its competitor is getting 10x the number of new users per day. It’s difficult to calculate how big it can become, but I found that its core users were “normal people,” and the need it serves was “people entertaining each other.” This is a very fundamental and common need, so if it succeeds, it has a chance to become a huge company. Now, every day over 100 million people are using its app, spending 60 minutes on average. Its monthly revenue is over one billion RMB.

The second one is a mobile e-commerce company serving restaurants. When I found it three years ago, its monthly revenue was a little over 200,000 RMB, and it only fulfilled 90 orders per day, each of which made a loss. But this was an industry with almost no glass ceiling. The restaurant industry is making 3-4 trillion RMB annually and growing, a third of which is spent on procurement. In the second year, it achieved a GMV of over 1 billion RMB, and the third year 4-5 billion. Now the company’s cash flow is very healthy, but it’s still far, far away from its “glass ceiling.”

The third one is a used-car auction platform. The used-car market is huge. In the U.S., 10 million new cars are sold every year, but 40 million used cars are traded. This means that every car will be traded 4-5 times during its life cycle. In China, the number of new cars sold is already 1.5x more than in the U.S., but used-car sales are only a fraction of new car sales. So, the glass ceiling of this company is also very high. Its GMV is now in the tens of billions and still growing fast.

Now, let’s take a look at the areas XVC has invested in.

l  Tutoring - this is an industry with no glass ceiling. As long as you offer a good service that can help kids achieve competitive advantages, parents are willing to spend their life savings. The two leading companies, TAL and EDU, have a combined market cap of more than $30 billion, but they own less than 3% of the total market.

l  Auto parts - this is a market even better than used-car auctions. In the U.S., this is a $150 billion market, and there are five companies with market caps over $10-20 billion.

l  Home renovation, building materials and furniture are each a trillion RMB market.

l  Social networking and transportation are also huge markets.

These companies have no glass ceilings. I cannot say they will all succeed, but if one of them succeeds, it will be a “huge success.”

My second approach to “catch the good luck” is to only partner with great founders.

Why only great founders? Because:

l  Even if you are in a very promising area, if you lose, you get nothing. Great founders are good at winning.

l  In bad situations, great founders can turn things around.

l  When the companies are successful, great founders won’t stop and rest, but will continue to learn, work hard and take the companies to a higher level.

Let’s use Pony Ma as an example. Why do I think he is a great founder?

First, he is an excellent product manager and has very good thinking habits. In the early days of QQ, he would spend hours every day chatting with his users to gain first-hand user insights. This shows that he is a fact-driven and user-centric person.

In addition, he is an excellent leader. He is not only good at recruiting and retaining extraordinary talents such as Martin Lau and Allen Zhang but also good at building a system and culture that enhance efficiency and foster innovation through internal competition and collaboration.

Let me give you a few more examples. These are three companies I missed.

The first one is Toutiao. When I got to know the company, it was raising its series B. No one in the VC industry believed its story. In the end, a Russian guy Yuri Milner personally invested. I didn’t understand the model, so I didn’t even care enough to contact the team. The first time I met Yiming Zhang was in 2016, when I was trying to raise money from him for XVC.

The second one is Pinduoduo (PDD). When I met the founder, Colin Huang, he was raising a Series A. I seriously conducted due diligence, and I seriously turned him down. It is said that their GMV per month reaches 3 billion USD, and the company is now valued at 10 billion USD.

The third is Qudian. I met with Min Luo in Series A, B and C, but could not build enough conviction each time. They recently went public on the Nasdaq, and their latest quarterly net income was 650 million RMB.

These three companies have one thing in common besides being very lucky – a great founder.

Yiming Zhang is a great product manager. He always sees the essence behind the phenomenon, so his mind has no boundaries. When other news apps were claiming to offer “news with attitudes,” he already realized that gossip is no different than news, and he insisted on replacing human editors with machine recommendations. When others worried about user perceptions, he threw away these useless theories and brought on video and newsfeed ads. When others were worrying about “being focused,” he quietly launched several new products with 10 million+ DAU.

Colin Huang is an excellent product manager as well. Although PDD rode on the “wave of WeChat traffic,” it experienced huge transformations, changing its business model from B2C fresh-food retailing to a third-party marketplace and bringing its users from WeChat groups into its app. Besides, it is hard for a supply chain to keep up with such exponential growth. To build a team that can support such a rate of growth, the founder has to be a really good leader.

The founder of Qudian is also good at handling big transformations and surviving extreme difficulties. Their initial business, after growing rapidly for two years, was suddenly shut down by the government. They had to completely abandon the old business and start from zero. It took them only one year to stage the comeback and make a quarterly profit of 650 million RMB.

To summarize, when big opportunities emerge, only the most exceptional founder can catch them. Our job is to find this kind of founder and invest in them.

The third approach to catch “good luck” is to find business models that can monopolize key and scarce resources.

Let’s see what key and scarce resources Tencent has monopolized.

Let me ask you one question – what would you feel if you forgot to bring your cell phone with you one day? A little anxious? I would feel anxious even if I forgot to take it with me to the restroom. What really created the anxiety is not cell phones, but WeChat. If you don’t believe me, try uninstalling WeChat and see if you feel the same anxiety.

In the ancient age of hunting and gathering, if someone loses contact with the tribe, her survival rate would decrease, so in the long history of evolution, our genes learned to apply substances such as dopamine and endorphin to build a reward and punishment system, keeping us in contact with our families and friends. If we stay in contact, we feel good; if we lose contact, we feel bad. Social network systems such as Facebook and WeChat greatly raised people’s threshold of “staying in contact.” People will feel uncomfortable even when they have only lost contact for a little while. The mechanism of the addiction formation is similar to drugs.

WeChat, through this mechanism, monopolized the “communication protocol.” Because the people you chat with will only frequently check and reply to messages on WeChat, the reply rate and speed on WeChat will be much better than other instant messenger tools, including text messages. On the other hand, when other people try to reach you, they will prefer WeChat.

People never signed any protocol to use WeChat together, but they abide by that protocol. Anyone who does not follow this rule will be punished. That is a natural monopoly.

Why invest in this type of company? Why not simply look for a high growth and high margin company?

What we invest in are all early-stage companies. The high margin of today does not prove a high margin in the future, and vice versa.

The real important thing is whether you have a long-term pricing power.

A company that can monopolize core and scarce resources has strong pricing power. Because it monopolizes core and scarce resources, its suppliers and customers do not have alternatives. As a result, it can use its pricing power to protect profit margin and create value for shareholders.

What’s more, this type of company can overcome “anti-scalability gravities” and become big.

If you have worked at a big company, you would know that the bigger the company, the more inefficient and costly its management will be. Tencent is not an exception. Every additional layer of management between the entry level employees and shareholders of the company will reduce the entry level employees’ sense of responsibility.

Owners of sole proprietary companies have the highest sense of ownership. They are extremely hard working and thrifty; they enjoy a lower tax rate and they have a lot of low-cost ways to acquire customers and deliver products and services. These are all “anti-scalability gravities,” and only the companies that monopolize core and scarce resources can overcome the gravity to continue to grow and maintain their competitive advantage as they gradually lose their efficiency and flexibility as small firms.

In addition, this kind of company can decouple from the macro environment. When inflation occurs, they usually raise prices first; in the deleveraging cycle of economies, they can expand their market share. So, you can sleep well while you hold them for a very long time.

Speaking of this, I haven’t defined “core and scarce resources” yet.

What are core and scarce resources? Some resources are scarce, but they are not core. The core and scarce resources are what can significantly influence customers’ purchasing decisions. For retailing industries, a good location is a core and scarce resource. For consumer foods, a good shelf space is a core and scarce resource.

For the areas we focus on, we have defined some patterns of core and scarce resources.

• Networks and platforms that enjoy quality of scale or network effects
• Applications that can generate data and use the data to improve their quality
• Ecosystems that can corner participants
• Self-reinforcing brands

We just talked about investment thesis. Now, I want to talk about execution. We believe the keys to execution are “Focus, Patience and Sincerity.”

Before I go into details, I would like to first mention the concept of a “decision engine.” We can think of every investment firm as a decision engine like Alpha go – inputting information from one end and outputting decisions from the other end. A hedge fund is a decision engine, and a VC is also a decision engine.

This decision engine has two limits:

1)    Each decision engine can only have one set of algorithms, and it cannot be too complicated.

2)    Each decision engine can only handle a limited amount of information.

We must admit that we have limitations. As an investment firm, our capabilities have boundaries. Algorithms cannot be too complicated, and there are conflicts between different algorithms. If we choose one of them, we will have to abandon the others. The amount of information that we can obtain and process is also limited; if we study some industries, we have to ignore others.

We can’t make all the money in the world, and we need to be aware of our boundaries.

I usually hold four to five companies in my stock portfolio and change one or two every year. I have never touched some great companies such as Alibaba, Weibo, Qihoo and NetEase. If the timing is right, buying these companies can make a lot of money. At some points, I did have the impulse to buy them, but held back nevertheless.

In retrospect, I resisted the temptation to buy them for one or several of the reasons below:

1)    I wasn’t sure how it could “monopolize core and scarce resources.”

2)    My knowledge of the company wasn’t deep enough, and I didn’t have the time to research.

3)    To buy the stock, I would have had to sell some others, like Tencent.

We also constantly review and question ourselves – if Toutiao, Pinduoduo and Qudian showed up again as early stage companies, would we recognize them?  

Honestly, we don’t have the answer. Limited by our knowledge and comprehension, at early stage, we may still be unable to tell whether they can monopolize scarce resources or if the entrepreneurs are great. We can and will learn to extend our boundaries, but before then, some opportunities just don’t belong to us.

You might want to challenge me – why don’t you hire talents who can complement your knowledge and comprehension? Many other funds have done so – besides the TMT team, they have a consumer team and a healthcare team, and under each team there are sub-groups like consumer internet, SaaS, smart hardware, frontier tech, online education, etc.

Yes, hiring more talent is a real temptation. We only have four full-time employees in the investment team. Whenever I felt exhausted, or I had to pass on a deal because I didn’t have time to research, the thought came to me – should we hire more people?

I held back eventually. I had two concerns. The first is, the more people we have, the more time I’ll need to spend on management. Then, I’ll be farther from the frontier, and the quality of my information would drop. The second is, the more people we have, the more time we’ll need to spend on communication, and our ability to process information would drop.

When the size gets bigger, a team will start losing its ability to see the essence and making conclusions too quickly. I can give you a few examples.

Many VC funds had believed that JD’s business model is too “heavy.” But its cash-flow turned positive a long time ago. Not just JD, but Walmart is also a company with “negative working capital.” This means that the growth of their businesses will not consume cash, but instead increase cash.

Another common mistake is viewing “unit economics” in a static way. When VIPShop went public in 2012, its gross profit per order could not cover fulfillment costs. At that time, the whole market held a skeptical view of its business model, but the company actually had a high barrier to entry and growing pricing power. After the IPO, its cash flow soon turned positive. In the following quarters, its margin continued to improve, and fulfillment costs continued to drop, and the company became highly profitable.

The funniest misperception I have seen is that some people simply think the “internet of things” has network effects. The definition of network effects is “when a new node is added to the network, each node’s experience will improve.” Some applications of the internet of things may exhibit economies of scale, but most of the applications simply add a sensor to collect data through the internet. Although it is called a “network,” the experience of every node is the same regardless of how many nodes there are in the network – there’s certainly no network effects.

There is another reason why we don’t want to increase the size of our team – we have a unique understanding of teamwork.

We believe “democracy and equity” will lead to low-quality investment decisions. In group-discussions, people tend to pretend to know something even if they don’t because of the pressure of “making contributions.” We think all communications are inefficient, because language itself is imperfect.

We believe that a group with “group thinking” is the most inefficient. One independent thinking person is better than that. A collaborated group of independent thinkers is the most efficient.

So, we have adopted some unique approaches to teamwork. Let me give you a few examples.

We have a culture of “making independent observations.” We hope we are not affected by other people’s opinions. In our company, if you always try to influence with your opinions, you would feel unwelcome. It’s OK to have opinions, but if you have no observations or thinking behind your opinions, they won’t be respected. Whenever we discuss an issue, we will ask for your opinion, but immediately we will ask you, “What is your observation?” and, “What is your logic?”

We do not encourage group discussions. We use our weekly meeting to share knowledge and one-on-one meetings to discuss deals. We found that discussions in group settings usually become debates. To debate, you need to have a clear standpoint, then you automatically wear a pair of tinted glasses, filtering facts in favor of your ideas and eventually getting yourself convinced. Only in one-on-one discussions can you “sit down together and figure things out.”

We usually have multiple people working on the same deal, but we do it independently, and we don’t “divide and conquer.” Everyone does his own customer interviews and his own data analysis and writes his own investment memos with his independent conclusions.

Our philosophy for hiring is also unique.

We do not put a lot of weight on background, experience or track records. What we value most is the candidate’s thinking habits – curiosity, independence and objectivity – and a passion for entrepreneurship and early-stage investing.

The third element of our execution is “sincerity.”

Early-stage investing is hard. You stay up late often and travel frequently. You have to constantly fight your own intuition and common sense and suffer from an extremely long feedback cycle.

The worse thing is that you need to say no to people every day, sometimes making you forget who you are. So, it is important to be a little idealistic, and hold a sense of mission in your heart.

Our mission is to create long-term value through innovation and entrepreneurship. We have defined this from the beginning.

So, when we are confused and don’t know how to make decisions, we often ask ourselves if we are doing something that creates long-term value. If the answer is no, then we don’t do it.

Finally, I would like to give you some words of advice – wealth comes from one or two big climaxes, but happiness relies on many small climaxes.

Big climaxes will raise the threshold of happiness, so I always remind myself that decoupling happiness with wealth is key to continued happiness. I’m good at enjoying the “process.” I found that one my happiest moments is having a strategy discussion with a CEO in a coffee shop. We met at 10pm and finished at 2 a.m. when the coffee shop closes. Walking out and seeing the empty streets, I felt extremely fulfilled and gratified.

I am very lucky to have the support from you, our investors. With your support, we have the chance to start a VC firm with a unique approach, and to help some exceptional CEOs build great companies and create long-term value.

I think this is my luckiest accomplishment. Thank you all!