Class Notes on Lecture 1: From 0 to 1 – Technology and Globalization; Monopoly and Competition.
Background: Elon Musk, Pony Ma (马化腾), Satya Nadella and Mark Zuckerberg and others in the innovation and entrepreneurship world have recently given lectures at Tsinghua University (清华大学) and Beijing. Tsinghua graduates form the third largest group of alumni in Microsoft and the rate of unicorn generation in China has almost caught up with the US. More than a million copies of Zero to One were sold in China which was slightly more than the rest of the sales in the rest of the world combined at the time of the lectures. The title of Peter’s course in Mandarin is 创办新企业 中美新观察 which translates to New observations about starting new enterprises in China and the US. Even if you have read Zero to One and Blake Master’s class notes of Stanford’s CS183 before, in Tsinghua’s 60510261, Peter restates, reorganizes and develops his concepts and illustrates them in refreshing and revealing ways even as he adopts a more comparative and empirical approach. The following is the essay version from my class notes of Class 1, any errors and omissions are my cross to bear:
I: 0 to 1
The question that we should look at again and again is “How far do the lessons of Zero to One translate to China and how China is different and how it is the same”. I am extremely excited and honored to continue this conversation about the future of startups, technology and China.
One of the meanings of 0 to 1, is that somehow every moment in the history of innovation, science or in technology happens exactly once when someone discovers something new, hence it is hard to come up with a system for it. Experimental science starts with the number 2, it starts with experiments that you can repeat. The challenge is that some of the most important aspects of business has a one of a kind character and moments that are so important for entrepreneurship, innovation and the world. Hence there is no straightforward formula almost by definition that one can give in teaching about start-ups and entrepreneurship.
Does the question of 0 to 1 progress (e.g., going from the first typewriter to the first word processor) apply to and for China? As with many things you can always split them into three parts. For any proposition, you can ask the trichotomous question: Is it desirable, is it possible, and is it necessary? There are people who say that it is not possible and not necessary to do something new in China and I will address these critiques later in the course and ask if the answer is somewhat in between. There is a very big picture historical example in Japan from the 1950s to the 1980s and maybe all the way back to the Meiji Restoration which had this idea of copying things that worked extremely well and after 1995 per capita GDP in Japan stopped growing and hit a wall.
(Chart from countryeconomy.com)
This leads to a question that if China was only reliant on copying ideas, whether the same thing would happen and what China could do to avoid that. China might have to start thinking about doing new things beforehand, as it is not clear whether you can drastically and successfully transition from copying everything to not copying altogether.
II: Creating and Capturing Value
Moving from the more macro question of 0 to 1, the question people are more interested in is the more micro questions with regard to specific companies and businesses. The single big idea is that if you are starting a business and doing something new, you always want to be in a one of a kind monopoly business. One simple way of thinking of what a business does is that it creates X dollars of value and captures Y% of what it creates. The challenge however is that often people naively think that if X is large enough then X multiplied by Y is still very large. However X & Y are completely independent variables. In many areas of business, Y is close to 0% and in such a case, it does not matter how big your business is as it will be difficult to achieve significant profit margins. The domestic revenues of the US Airline industry are about 3 times bigger than Google’s global revenues. Additionally if society at large had to choose one we had to live without, we would probably choose Google’s search engine rather than the airline industry which indicates that airline technology is still more important. Yet Google’s market capitalization as a company is about 3 times more than the US airline industry combined, which means that there is almost a 10X difference between their market capitalization/revenues ratios because Google’s profit margins are so much higher. Within the United States and Europe, Google has had minimal competition in Europe and the United States. In business economics, the idea of monopoly is not very well understood or discussed very much in comparison to perfect competition for intellectual, psychological and social reasons which I will explore in today’s class.
In a world of perfect competition (such as the restaurant industry), it’s very easy to model in a static world without much changes and is somehow optimally efficient. However, it is not the best approximation in a dynamic world that changes a lot. When I studied economics at Stanford in my freshman year, someone asked the professor if all these businesses compete perfectly, how does anyone make any money? At the other end of the spectrum of these descriptive models, monopoly gives you an incentive to innovate, because if you are able to innovate you are able to capture a lot of value. It also orients companies to stable long-term planning. However in a static world, monopoly produces less output, charges high prices and maybe even stops innovation, thus becoming a bad thing. If you look at the context of economics, historically it emulated physics and the assumption was that it is easier to model and focus on perfect competition. Irving Fisher’s 1892 Yale PhD dissertation argued that there is one to one correspondence between physics and economics: A particle is an individual, space is a commodity, a force is a marginal utility or disutility, work is disutility and energy is a utility. There was this concept that economics was a science as precise as physics. Then there was this more complicated equation by Robert Solow describing the growth of an economy and the Solow residual which cannot be explained by capital or labor. This residual describes innovation, however often it is treated as a constant or a variable that is neglected.
III: Lies people tell (the business reason why people do not talk much about monopoly)
If you are a small business, you will tend to tell the story that your market is very small and that you are very unique. If you are a big business like Google, instead of saying that you are a monopoly in US search, you will say that you are just a small part of search advertising or even a smaller part of consumer technology:
The critical thing on some level isn’t the different types of stories people tell. As someone who is starting a company, you want to tell the true story, at least to yourself. Ask yourself, is the real market advertising, search or technology and how you get a large share of that market and not a market which is artificially or fictionally described.
IV: How to build a monopoly
If you are starting a new company you are always starting small and you want to start with markets that are relatively small, which is the opposite of how business schools focus on TAM (total addressable market) and going after really big markets to become a big business. Instead, you want to go after markets that are relatively small, takeover them and somehow expand and scale from there. Ebay, Amazon and Paypal all started with small sub-markets. For example, Paypal’s 30k “PowerSellers” on Ebay needed faster payments solutions and with them we were able to go from 0-30% of market share in the course of 3-4 months. I often get asked about or pitched on payment or other companies where start-ups always have a scheme that works slightly better for a very large market or the entire world. However the problem is that even though there are a lot of things that are broken and do not work really well, incremental improvements for large amounts of people are very difficult to roll out. It is hard to convince people to adopt it, unless there is a steep gradient or large delta in value for some subset of your initial customers and adopters. Small merchants that could not accept credit cards and had to accept cheques which took 7-10 days to process suddenly found that payments became instantaneous. And this was the kernel of 30k customers from which the Paypal business gradually grew. On the other hand, there are a lot of companies that create a small amount of value that is diffused among a large amount of people and it is difficult to get people to even start using it.
Facebook had a very dramatic beginning as well, starting in Harvard University in 2004 and going to 0-60% market share in 10 days before gradually building out from there. If you had presented Facebook as a business plan in February of 2004, people would have said that the market of 10, 000 students is too small, not interesting, will never get funded and this is a terrible business. The consistent critique that many of these internet businesses faced is that the business is too small and even though it expanded to other campuses, it cannot grow much more. However, the small initial markets were actually a sign of their dominance in a way. There is a very similar story with Alibaba here in China. Alibaba today seems to do everything from consumer retail, online payments, financing and all kinds of products on a global scale. However, it started out as a very specific B2B service connecting Chinese exporters to importers in other countries. It was initially focused on signing up small businesses in Jack Ma’s hometown before it built up a critical mass of buyers and suppliers. Only after you have established this initial position should you try to expand. On the other side, the kind of thing that does not work is to start big and get smaller. The most common mistake (that is exemplified by the first page of the business pitches of all clean technology firms in the U.S. over the last decade) is that we have an enormous market called energy which is worth trillions of dollars and we are going to have a small fraction of a small fraction of that market. The problem with an enormous market is that there are enormous amounts of competition. If you start with new thin film solar company you will have to compete with nine other thin film solar companies, ninety other solar companies trying something else and competition from China, fracking, windmills, etc. You never want to be a small fish in a very big ocean as this is a formula for disaster. It is almost a recognizable pattern or warning sign if founders really believe in markets that are too big.
One other way of describing this monopoly idea is in terms of two time frames:
t1 = time to take over a market
t2 = time (for you) to be copied by somebody else
If t2 < t1 it is very hard to take over the market. However if t1 < t2 , you can often take over the market before anybody else has noticed, i.e. 10 days for Facebook in Harvard and 3-4 months for Paypal’s Ebay’s PowerSellers. However, once you have taken over a market, it is often very hard for somebody else to copy you. In a large market, t1 might be ten, fifteen, twenty years and as it takes so long, so many people might have noticed you that its hard to take over the market. This is a slightly different way to get at this question of how to build a monopoly:
According to Pony Ma , “In America, when you bring an idea to market you usually have several months before competition pops up, allowing you to capture significant market share. In China, you can have hundreds of competitors within the first hours of going live. Ideas are not important in China – execution is.” Certainly there are aspects of this that are probably correct, for example China is very big, hence to some degree people are more competitive and pay attention more quickly. On the other hand, some of this is exaggerated, for example several months in the U.S. and a few hours in China is a very dramatic difference. Steve Jobs also said something to the effect that ideas do not matter, the execution of them does. So is execution all there is to it, or is it a self-serving thing on some level that Pony Ma should say such things if he wants you to work for Tencent? If you are told that ideas never matter, then it follows that you should never think for yourself and always work for a large existing company. The question of monopoly is always very tricky and somewhat obscured. The premise of this class is that ideas are not the only thing but they are important, they do matter, it is a place to start and you do need to execute them well. You should have both ideas and execution as you try to do new things.
V: The Contrarian Question
One of the ways I like to frame this is with a contrarian question, where there is a business version and an intellectual version.
The business version is always: “What valuable company is nobody starting?”
The intellectual version is always, “What important truth do very few people agree with you on?”
The assignment for the class is to have all of you answer this surprisingly difficult question in the small group seminars. I am often asked what my own answers to these questions are, and I have a lot of practice and a lot of answers. One slightly facile answer that I always give is that this sounds like an easy and straightforward question but it’s actually a very difficult question to answer because originality is hard and we are always under tremendous pressure to give the answers that everyone already knows to be right. In a job interview, it’s a very uncomfortable question for people to answer because the correct answer is one that the interviewer will disagree with. If the interviewer knows your answer, that is a bad answer. A good answer will elicit a reaction of “this is ridiculous, I’ve never heard of such a thing”, which gets to a sense of why we are very uncomfortable with this kind of question or the innovation if often points to.
Answers to this question are often answers about the future: “How will the future look different from the present in a way that people have not yet understood?”
The business and intellectual questions are actually often very similar questions and a good answer to the contrarian question(s) offers a refreshing perspective on the future – on something that may happen or may be possible. This monopoly idea that is so central to business -- how will the future look different from the present in a way that is not understood -- ends up not being accessed intellectually and socially. You should think of almost every good startup as providing a good answer to this question.
Pre-facebook, there were all sorts of theories about social networking and people said social networking was important in one way or another. Reid Hoffman started SocialNet in 1997, seven years before Facebook, and there was another company called Second Life which started a year before Facebook. People pretended to be virtual cats and dogs interacting in cyberspace. The usual question everyone was asking in 2004 was “How do you network on the internet?” The harder question that Facebook focused on and answered is “How do you translate their actual identity onto the internet?” Facebook starting at Harvard was perhaps the right place to start because Harvard students wanted to know who the people around them were and the real identities of their fellow students. One of the early competitors of Facebook in social networking was Myspace which started in Los Angeles where you have a few movie stars and a lot of people who want to be movie stars. Basically, everyone (on Myspace) is pretending to be something different from who they are. Myspace was about fake identity and Facebook was about real identity. Most of the social networking sites were somehow about pretending to be fake and the real trick was to get people to represent themselves as who they actually were and it was a different problem that Facebook somehow solved:
There are a lot of different perspectives on the future and there are many contrarian answers to these questions. They are very easy to see with 20/20 hindsight but very hard to see in advance. One of the mainstream perspectives that is too pessimistic is that we have already figured everything out, and there are no new answers left because everybody knows everything. Therefore if you have an idea different from everybody else, it just has to be wrong. Most of the time, people are right about things such as 1+1. For instance, 1 + 1 =3 can also be very contrarian and very wrong and does not count as a good answer. There are many new secrets like Facebook’s that are underexplored that we should uncover. However, we are tempted too easily to say that everything has already been thought of rather than to say that there are some very open questions that still remain. We’ve also talked about how monopoly gets ignored because of intellectual frameworks in economics and fictional stories that business people tell, but the deeply trained and ingrained social psychology of competition also helps to obscure the importance of monopoly.
VI: The Psychology of Competition (which will be discussed more in depth separately)
VII: Q & A
Q1: According to your lecture, when you want to startup a company you should ask yourself these three questions: Whether it is desirable, possible and necessary? How do you figure out if your idea is necessary, for instance before the IPhone came out, maybe Steve Jobs knew that people might like the touch screen function, but he may not have been sure that its necessary.
A1. You can use the trichotomy in many areas. The standard thing people do in a startup context is wonder often about whether there are features that are either 1) nice to have or 2) need to have - and its often a fine line to go from one thing to another. The more common failure is to think that it’s not necessary for them to do it, as others will do it. Then there is a more general version where people make anti-startup arguments. For instance, an anti-startup argument is that it’s undesirable because you have to work really hard, it’s risky and a lot goes wrong. It’s not possible to start a company because ideas get stolen too quickly, the market is too big, people won’t believe it, you won't get funding, its bad and it’s not necessary because it will all get done by a bigger coy with more resources. This is very abstract but you can come up with this structure where you tell yourself you should never start a company all the time because its 0 out of 3. Obviously, this is an incorrect argument, but it’s not easy to tell what is wrong with it. When I started PayPal, one of the big push backs I got was that it was unnecessary to start a new payments company because everything could be done by existing banks. “Even if this business with email and money could be nice to have, it’s not necessary – certainly not necessary for you to do it as it can all be done by big banks which have more resources, people and marketing.” The reason such arguments tend to be wrong, is because of institutional barriers within these large organizations where you have to somehow set up a new team within this organization and pull resources from other people. Even though in theory you are much better off doing it at a big bank, in practice it’s not. Argument for startups in general is that even though it’s very hard, problematic and undesirable, yet somehow within existing structures it’s hard to do new things. For some reasons a lot of big companies get to be very bureaucratic and slow, they don't think they are slow but they really are. Hence, there are more openings for new companies than what one thought was possible.
Q2. When it comes to innovation, you seem to mainly focus on technology. In China, the conversation is more about process and business model innovation such as online to offline and peer to peer lending and not so much about inventing new technologies. Instead it’s about finding new ways to apply cutting edge technologies to streamline and increase the efficiency of an operation. Do you think there is a balance in looking for innovation in the business model and in the technology?
A2. Certainly there is innovation in applying technology and in business models. The thing is you have to be honest with yourself and about the barriers to entry. Business process innovation can be a real form of zero to one innovation, but you have to ask yourself if it’s easy to copy in one way or another. A simple rhetorical device is that if people talk to you about the technology, such as talking to a biotech scientist who invented a new medicine, talk about the business aspect. If they talk about business, then talk about the technology. Talk about what people do not want to talk about and get to very uncomfortable and difficult territory. If it's a financial technology company, a hard question is usually what their proprietary technology is and what really differentiates them. If they do not have much technology, maybe they still have a good answer in their marketing or distribution channel or roll-out speed that differentiates them but then you have to drill down and figure out what their cost, market and efficiency is. If you line up ten different companies in the same area, you often get more insight than talking to one. You meet someone new and their business process and distribution innovation may sound like a great and exciting idea. However, after you talk to nine more people, it all begins to sound very similar and you conclude that none of them quite work. Monopolies can take many forms in reaching new markets but whatever people tell other people, at least be honest with yourself.
Q3. Companies can find it hard to expand beyond their small markets as they and their expertise is too narrowly focused or are unable connect to new markets?
A3. Certainly there are these cases where you takeover small markets and are unable to scale beyond that and it is always a little hard to predict how far these things can scale. You are still quite well off even if what you have is a small market and are the best in the world at what you do, even if you cannot expand. This is still a better place to be than in a big market with a lot of competition. I was born in Germany where many of my relatives are from and we have this very interesting phenomenon called the Mittelstand, which are hundred year old hundred fifty year old companies that are specialized in extremely narrow markets where they are the best in the world. Even though there is no natural way for them to expand beyond that, it is still actually a very valuable place to be.
Q4. Which sectors have the most promising future here in China?
A4. I always dislike answering the sector as it's like a formula. If I could tell you what sector it is, there would be a lot of people already doing it. The similar question about good trends, such as SAS, cloud computing, etc is incredibly misleading. These trends and sectors are just buzzwords which are facts and evidence that they are not differentiated and good answers to what is true and important that few people would agree with you about. If I have to give you a detailed explanation about a new sector you have never heard about then I may as well start the company myself and not tell you about it. If you heard an elevator pitch like ‘mobile platform for SAS enterprises to bring big data to the cloud”, my instinct is that the buzzwords are an indication that the business is undifferentiated and the company has not solved the most important problem. Many of the companies that are really differentiated do not even have the language or words to describe what they are doing. If Facebook described itself as a social networking site, it would be a very misleading description which did not capture essential differences between it and other social networking sites. Google was also in fact different in 1999 from other search engines because it was machine powered search but somehow there was no language to describe this new category. The challenge of discovering a new category of one and figuring out how to execute it is the main challenge of building a new business. It is something like Godel’s incompleteness theorem, if I proved it was complete or gave u an answer u know it would be wrong.
(Omitted a small amount of material that has already appeared in http://blakemasters.com/peter-thiels-cs183-startup and Zero to One)
Comments