钱颖一对话张一鸣：对话实录（上、下）

Qian Yingyi in Dialogue with Zhang Yiming: Transcript, Parts 1 and 2

> Outlet: transcript released by Tsinghua University School of Economics and Management (its Sohu account) · dialogue held 2018-03-20, Part 2 published 2018-03-27 · [Part 1](https://www.sohu.com/a/226419447_641792) · [Part 2](https://www.sohu.com/a/226504510_641792)
> Only his own answers are kept. Prompts marked (Q) are the editor's one-line summaries, not the interviewer's wording.

**(Q)** You applied for biology, were reassigned to microelectronics, then switched to software engineering; what were the reasons?

Why did I apply for the biology department? At the time everyone said biology was the leader of the 21st century, so it was very hot. I was interested myself; in high school I took part in the biology olympiad and read 普通生物学 (General Biology), written by a Peking University teacher, which influenced me a lot. Biology, from cells to ecology, has rich and diverse species, but the laws behind it are very simple and elegant, and that offers many analogies whether you're designing a system or looking at a company's economic system.

As for microelectronics, the major I was reassigned to, I was also very interested in microelectronics myself. But after studying it for a while, I felt there were too few chances for experiments, or rather it was hard to turn what you learned from textbooks into practice. It happened that in our first and second years we also studied computers, so in the second semester of my second year I switched to software engineering. Studying computers has one advantage: as long as you have a computer and internet access, you can get learning resources yourself, and if you have an idea you can build it directly; even as a student you can release a product to a broad base of users. That's what finally drew me to computing. (The varied academic background) was also very rewarding for me; when you start a company you look at everything, and along the way I also took business administration as an elective.

**(Q)** Which experiences in your four years of university most influenced you later?

Our university is in Tianjin, unlike Beijing, so there weren't that many extracurricular activities to join. So I had a lot of time to read all sorts of "random" books. In my first and second years, besides classes, I read a great variety of books, with biographies a large part. The main thing reading does is shape your interests and taste. After reading biographies, in choosing a career later I was more patient about my career planning. You see that many very great people lived much the same kind of life when young, also made up of small things; everyone is an ordinary person. You need patience; keep going deep in one field and you'll achieve corresponding results. I've always felt that if you could only read a few of the world's books, two kinds are worth reading: first biographies, second textbooks.

**(Q)** You were already thinking about the efficiency of information flow at university; how does that connect to your later startups?

This is the main theme of my startups. I'm a heavy information consumer myself. Besides reading lots of biographies and miscellaneous books at university, in secondary school I also read lots of newspapers and magazines, even the gutters between pages. I felt then that transmitting information was something with a very big impact: differences in the efficiency of information transmission make very big differences to the efficiency of human society as a whole, to cooperation, and to people's understanding. When newspapers were first invented, they were a very luxurious product. I believe that in ten or twenty years, when the post-10s and post-20s grow up and learn that to transmit a piece of information back then you had to carve characters and mimeograph, and hand-deliver thousands or tens of thousands of characters to places hundreds of kilometers away, they'll find it very strange.

The flow of information itself has a very profound impact on society; I even think it may be the foundation of all other kinds of efficiency. The speed of information flow or the degree of order in how information is organized produces very big disputes (争议). When we studied computing we built MIS systems, storage systems and 3D display, also all centered on information; the IT industry is the information industry. Of these, the organization and distribution of information produces the biggest disputes. The information is already there; without effective organization and distribution, the value it produces is actually very limited. After graduating, whether I was doing a search engine or a social networking site (a search engine organizes and distributes information; a social networking site organizes information flow with people as nodes) or a recommendation engine, which organizes and distributes information with interests as the granularity, it was all basically centered on information distribution. Douyin too makes it easier for information to be created and distributed. Inside our company we have an efficiency engineering department devoted to optimizing the flow of information within the company.

> [Note: both instances of 「争议」 ("disputes") in this paragraph read 「增益」 ("gain") in The Paper's combined version; probably a transcription error; the original is kept.]

**(Q)** You worked at four companies from 2005 to 2012; how did you plan and adjust yourself over those seven years?

Before that my startups were all fairly technical. The startup in social networking gave me the chance to observe users, which helped a lot in understanding users and products intuitively. Social products deal with people, and many users give you feedback directly. Starting a company with Wang Xing, that startup paid great attention to product quality; it was truly very sensitive to a single pixel, the slightest color difference. When I started a company on my own, 99fang was a rehearsal for founding Toutiao. At 99fang I was CEO. Before, I'd finish my duties and report problems upward, but a CEO can't report upward; he has to make independent decisions. Hiring and firing can't be pushed onto others. It was a rehearsal for founding Toutiao. I was very lucky that in the end, when the mobile internet wave rose, I had all these experiences.

**(Q)** Was founding Toutiao a long-held intention or chance?

This startup was also atypical. The search engine company I ran was doing all right, and I actively chose to found a new company and choose a new direction. Later our business was quite badly affected, and I handed it over to another partner; it was actually a pretty active choice.

**(Q)** Why make this choice actively?

Mainly because I saw that the mobile internet wave was very, very big, and I'd felt this wave even earlier, in 2008. In 2007 Apple released the first iPhone, and when I bought one I was very shocked: you could build a website or write a program on it. Put it in your pocket: it was hard to imagine a website lying in your pocket; you could write a website and open the browser to visit the site you'd written. Putting an operating system on a mobile device was highly disruptive.

Later, while we were doing the social networking site, we found that mobile internet was rising by a few ten-thousandths a week; still very small, but rising all the time. In 2011, when we were doing search, it was rising very fast, and we made several internet apps for property, which were also very successful. By the end of 2011 I basically judged that the whole medium of information transmission would change, from print as the medium to phones as the medium: real-time, two-way, multimodal, able to transmit information on all kinds of subjects, with many sensors too; the whole transmission of information was changing enormously. In that situation, compared with doing a vertical business versus a platform business, the chance to build a platform product was very rare, so I actively chose it.

This also has to do with reading biographies. People easily feel nothing about big things, and generally feel nothing about big turning points either; usually they only feel it afterward. In 2011 I really did feel that technological progress would make a field appear, and sensed that major, far-reaching innovation and far-reaching change would come, and this change still hasn't stopped. Six years have passed now, and from 2011 to 2016 the whole ranking of listed companies changed because of mobile internet, with the per-capita output of the top five changing tenfold. We're now internationalizing overseas and have been to many countries, and we find mobile internet has changed the economic, cultural and social face of many countries. This trend hasn't stopped, and video is also a sub-wave on the mobile internet, brought by faster mobile network speeds. [Translator's note: the source reads 发生10倍的变化, literally "underwent a 10-times change".]

In 2011 I felt it fairly strongly: this really was a huge wave. Generally you only live life once and can't examine it; you're in the middle of it. Reading biographies gives you the chance to examine; although it's someone else's life, you identify with it as you read. As you read, you can see how people change within huge waves.

**(Q)** You've said not to compete in others' core areas but to pursue a different area; how do you find such an area?

With big waves, it's about timing. There will always be waves, but they have a certain periodicity. Wait for the moment; different eras have waves in different fields. You really do need to find your own interest; some call it interest, some call it aspiration, some say passion or ambition. The opportunity in a startup is that you look at the world from a different angle and see something different, something you might see ahead of others, and in the end you create it.

If you treat society as one overall system, you'll find you're not fighting for their share but creating new increments in society. If you can do that, it's win-win. For you, it's not doing things over again; you're not copying a business. For society, there's a new increment. If it's not your interest, you might give up because of fluctuations in the industry. If it is your interest and you believe in what you see, you can get past the influence of the industry or of public opinion.

**(Q)** How do you see these two points: seeing from a different angle, and sticking to your interest and aspiration?

Actually, from what I've observed, we've hired many people, and many of my classmates and friends have worked a long time in all kinds of industries and roles; there aren't that many people in society who truly have a strong interest and seriously do one thing for a long time. Those who manage it often do pretty well.

**(Q)** What role does mathematics play in your way of thinking and working?

When I was studying I couldn't see it either (laughs). Mathematics really is important, but at the same time I think Chinese (语文) is also important. Language and mathematics are both tools for describing things; sometimes mathematics suits the description, sometimes language. Language, like mathematics, is also a kind of computation. Legal questions, for example, also have causal analysis and nested concepts, very similar to writing programs or mathematical proofs. When language describes things, the range and degree of different words can also be drawn on coordinates.

It's just that compared with mathematics, quantifying descriptions in language is harder, and many problems can't be solved well if they can't be described very precisely, so mathematics is needed. Take recommendation: we build a user profile for each user, and many people think it means putting lots of tags on each user and describing them with lots of words. Actually no; it's seen as a vector, a relationship in a space. (Mathematics) is very important for describing the ad system, including describing the diversity, degree of convergence and degree of generalization of content; all of these can only be improved through quantitative description.

**(Q)** How do you get the post-90s, post-00s and even post-10s to like your products?

This is actually quite challenging, and I take it very seriously myself. In an earlier question, you said needs are changing, products are evolving, and it's very important to stay young, and products are all used by young people.

> [Note: "You said needs are changing, products are evolving…" is him restating what Qian Yingyi said earlier, not his own claim.]

**(Q)** Age keeps changing; how do you stay young?

Use Douyin more (laughs). We have an all-hands meeting every two months, and one employee asked: how can the company become very mature? My view is the opposite: we precisely need to stay young and not casually take a negative, critical attitude toward new things. When new things appear there's always a reason; you must experience them, try them, and watch how they develop. We ask a lot of our team. Very early on, I found the management team didn't use Douyin, and I was very anxious. I required them to shoot two videos a month and get a certain number of likes, using compulsory means to keep everyone young. Our company culture is "always day one" (始终创业), similar to Amazon's Always Day 1: always think as if it's the company's first day, always think about what users are thinking.

**(Q)** How can university students meet the innovation needs of future technology companies?

As an employer, we prefer people who can learn many kinds of knowledge and skills. For example, many of our company's product managers switched over from engineering, some are designers who became product managers, our head of human resources studied electronics and switched, and our head of administration studied computer science and switched. A purely matching major isn't that crucial; what's needed more is being able to learn many kinds of knowledge. Keeping the ability to learn may be more important than accumulating knowledge. Today's internet lets you get knowledge anytime; I think the ability to organize your own knowledge structure and update it may be more important.

**(Q)** Your interests were broad at university; how do you see that looking back?

If I had the chance to study again, I'd study more.

**(Q)** If you studied again, name two things you'd want to learn.

I'd want to learn more mathematics.

**(Q)** What else?

I'd want to learn more physics.

**(Q)** People now read in fragments; what advice do you have for students?

I think reading Toutiao, Douyin, Wukong Q&A or Xigua Video can indeed enrich and broaden your horizons and supplement a lot of information. But I still think that if you want to study a foundational knowledge structure, you should read books, especially textbooks.

**(Q)** Why read textbooks?

Textbooks are the most concentrated, distilled books of human knowledge. If you're interested in some field and read general popular science books, a lot of that is general interest reading. In terms of concentration, the ones that best grasp the essentials are certainly textbooks. At school, having teachers push you to finish studying textbooks is still very important.

**(Q)** What is the most important thing for Toutiao in 2018?

For our company, which has now turned six, 2018 has three most important things. For the company itself, we hope to keep improving the company, whether management, culture or systems. For the business, we hope to go global. If (globalization) succeeds, it will be a milestone for us, and also a landmark for Chinese internet companies. Third, as a member of society and a platform company, we hope to do more in corporate social responsibility. In the past we cared more about business growth; now we want to set higher requirements for the management team and business heads: not only caring about the business, not only caring about globalization, but also having a broader perspective to recognize and take on more corporate social responsibility.

Part 2: Globalization, and from "BAT" to "TMD"

**(Q)** Why did Toutiao think about globalization so early?

First, because in this wave of mobile internet, Chinese companies and foreign companies started at relatively the same time. In the 1998, 1999 generation, foreign companies came out two or three years earlier, Chinese companies did "copy to China", had no chance overseas, and could only rely on localization to build their business in China. We feel the internet is interconnected and in future there will certainly be global competition. Chinese companies now, like American companies, are "Born to be global". If you can't allocate globally, you can't use global economies of scale, including markets, organization, human resources and so on. In this era China itself has opportunities, but only by using the advantage of global allocation can you achieve better results. The marginal cost of the internet falls very fast; with the same investment, if another market's population is 5 times yours, you can't sustain competition in the long run. We've seen good precedents from companies ahead of us: China's electronics manufacturing and machinery manufacturing have already been very successful, and among ICT (information and communications technology) companies, Huawei has also been very successful. When we launched globalization in 2015, the team wasn't that confident, because no internet platform company had ever gone overseas before.

**(Q)** What was your reason for sticking with globalization in 2015?

The most common reason (for not internationalizing) is that every country and region has a different culture. But I always encouraged everyone with one example: Huawei. Huawei's products need pre-sales, implementation, deployment and after-sales, and can be sold both to developed countries and to Africa. I told my colleagues, if a company that needs this much localization can go overseas, we certainly can too.

**(Q)** Huawei's globalization was much harder than yours; how do you see it?

For different cultures and languages, does the product need a lot of localization? Actually no. The best global products, even software products, Windows, Office, Facebook, YouTube, don't need it; the content only needs to be localized, and the product is universal. Our strategy is a global product plus localized content. Actually many companies do internationalization, and in the end what they do is set up an international department and develop new products for local markets. We don't; we set our vision as "a global platform for creation and exchange", hoping to be one unified platform.

**(Q)** Why build one unified product instead of developing new products for each locality?

Because the technology and the recommendation system can be shared, and with a certain amount of localized operations it can suit the locality. The cups are all the same; the drinks are different. Platform products can do this.

**(Q)** In the 2015 internal discussion, did you insist on it, or did everyone agree?

Some colleagues did have this confidence, but everyone knew it wouldn't be easy. And in fact it really wasn't easy. Now we can already see the direction, and it's fairly bright. Now in Japan, 1 in every 10 Japanese citizens is our user, with over ten million monthly actives, already a very high proportion. On the app charts in many countries we're in the top few, including the overseas version of our Douyin and the overseas version of Toutiao. Although there are still many shortcomings, we already see a big opportunity.

**(Q)** What Huawei did was more complex and harder; what did you see in it?

It really did take more effort; I looked into the course of Huawei's overseas expansion. From 1999 until before 2004 there was almost no progress at all. Without resilience, you'd think Chinese companies weren't suited to globalization, and you'd fail. But after 2005 Huawei's overseas progress was very fast.

**(Q)** Where is your competitiveness compared with American companies, and what is your model for overseas growth next?

Chinese internet companies' exploration in applications, including R&D, operations and even commercialization, is actually ahead. The main bottleneck to growth is overseas organizational capability; few overseas companies (海外公司) have successfully carried out globalization. We should play to our strengths, keep doing application-level innovation and efficient execution better, and at the same time learn from overseas companies, manage well, so the company can adapt to operating globally. Globalization means hiring global talent and building international teams well; there are still some challenges here at the moment, since many foreigners actually don't adapt to coming to work in China. [Translator's note: 海外公司, "overseas companies", is as written; the context may mean Chinese companies operating overseas.]

**(Q)** Operating globally requires hiring talent globally; how do you do it?

We're also setting up offices in many countries, on the principle of "Talent First", giving priority to talent. IT is so advanced now that we open offices wherever the talent is. This is also something we're learning from Huawei.

**(Q)** Do you also recruit scientists around the world, like Huawei?

We're doing that too. The advantage is that we can hire suitable talent according to each country's characteristics. Huawei set up a mathematics institute in Russia because Eastern Europe and Russia are very strong in mathematics. We now hire in each country according to its characteristics and according to the distribution of languages. We have many foreign employees overseas.

**(Q)** Is the talent you attract in Japan local?

Local.

**(Q)** Not just ethnic Chinese?

Not just ethnic Chinese; to truly do localized operations, local talent is still very important.

**(Q)** At what point does Toutiao count as a global company?

I've set myself a small goal in my heart: more than half of users from overseas.

**(Q)** Roughly what percentage of users are overseas now?

10%.

**(Q)** Is the gap from 10% to 50% big?

The gap isn't that big.

**(Q)** How many years from 10% to 50%, conservatively?

Keeping in mind the likelihood of delivering, three years.

**(Q)** Did the term "TMD" come out of a get-together between you, Wang Xing and Cheng Wei?

There was no such phrase at the get-together; the timing of when the media started using the term roughly matches. [Translator's note: TMD = Toutiao, Meituan (Wang Xing) and Didi (Cheng Wei).]

**(Q)** BAT is spread from north to south, in Beijing, Hangzhou and Shenzhen?

Yes, from north to south.

**(Q)** Why are all three TMD companies in Beijing; chance or pattern?

Beijing's overall basic conditions are fairly good; most importantly, there's lots of talent. One characteristic of this wave is that it's faster than the last. Tencent reached 10 billion US dollars in 2008, which actually wasn't as fast as people imagine: Tencent was founded in 1998, so it took ten years. But "TMD" took much, much less time than those three companies, because in Beijing, a place with abundant factors of production, including capital and talent, you don't need much accumulation time. So all of us being active in Beijing has some element of chance, but Beijing's basic conditions are very, very important.

**(Q)** You and Wang Xing are both from Longyan in Fujian; what is Longyan's environment like?

Longyan's environment, or rather Fujian as a whole, got digitization and internet access fairly early. When I was in high school, the internet was already widespread, which may have something to do with both Wang Xing's and my startups: you come into contact with the internet earlier and develop an interest earlier. Fujian got the internet fairly early, and there are quite a lot of internet entrepreneurs from Fujian as a whole.

**(Q)** Will the platform landscape formed by BAT become an obstacle for new startups?

Although the internet has economies of scale, in fact we see that cases of successful startups or investments are still increasing, and the overall trend in entrepreneurship and innovation keeps getting better.

First, because economies of scale are certainly important, but innovation is more important. A good environment is also very much needed: whether through law or industry self-regulation, leaders must not abuse economies of scale, and a good competitive environment must be maintained.

Second, whether Toutiao, Meituan or Didi, the birth of this batch of companies has to do with the mobile wave. Meituan was in 2010, we and Didi in 2012, and Meituan's rapid growth was also after 2012. When there's a big shift, like crustal movement, the whole ecosystem has to change again, and at that point new species have a chance to expand quickly. "When the big wave comes" is a very important foundation; big waves are often brought by technological innovation, for example computers, the internet, then the mobile internet; a big wave is an important moment. Waves come one after another, only is the future wave necessarily in the internet? Not necessarily. So-called technology is using new, better methods to create things; since it's using new, better methods to create things, new waves will certainly be created. For new entrepreneurs, opportunities on the whole are still better.

**(Q)** However strong BAT is, there are places it hasn't seen; how did you rise?

First, find the market opportunity accurately: see the blue ocean half a step ahead. At the same time, in an innovative environment, a company that doesn't advance falls back; it's actually very important for a company to stay sensitive to the outside. Yahoo rode the internet wave and seized portal navigation, but didn't seize search, didn't seize social, and after missing two small waves, Yahoo ceased to exist. This also reminds companies to stay vital, or they'll be swallowed by more vital companies.

**(Q)** What do the BAT and TMD generations of entrepreneurs have in common, and how do they differ?

I've had contact with all of them, and we've learned a lot from the "BAT" entrepreneurs; for example, I read 腾讯传 (The Biography of Tencent). What we have in common, I think, is grasping user experience, being user-centered, creating value for users, and valuing talent. As for differences: first, for startups around 2012, the domestic startup environment as a whole was fairly good; internet penetration was already high, and users adapted (to our products and services) right away. Second, it's much easier for us to raise money and hire. I believe that in 1998, 1999, before the internet bubble, it was a bit better, but in 2001, 2002 hiring wouldn't have been this easy. Now internet companies are still very competitive in hiring, and computer science admission scores are said to keep rising, because jobs are easy to find.

> [Note: here the original has an interjection by Qian Yingyi, "Mainly competing with our school." (主要跟我们学院竞争。), following Zhang Yiming's remark that computer science scores keep rising; it is not a question. Zhang Yiming's next words, "Yes,", respond to the interjection, and he then continues listing his third and fourth points. The interjection note is kept at its original position, and no prompt is set.]

Yes. Third, the supply of talent keeps growing. Fourth, the new generation of entrepreneurs tried globalization earlier; not only are we trying globalization, Meituan and Didi are too, because we didn't start behind foreign companies; it was almost simultaneous. Fifth, more daring in opening new businesses; these two (Meituan and Didi) are opening new businesses against each other. Less afraid of failure, higher tolerance, better conditions; on the whole more inclined to take risks, and businesses expand faster than before. You can also see that this batch of companies from 2012 grew faster.

**(Q)** Is the TMD generation more willing to try new things?

The 2012 batch of entrepreneurs hardly ever ran into "the company's out of money"; as long as the business was good, there was no such thing as being unable to raise money. The environment got better.

**(Q)** What is the essential difference between the internet's "first half" and "second half", and where is the future direction?

As I recall, the phrase "first half, second half" comes from the disappearance of the user dividend: everyone who should have a smartphone has one, all the software that should be installed is installed, and there's no new-user dividend left. From this angle, that's how typical internet applications are.

Developing further, after the user dividend from simple market and marketing means disappears, there are several directions for breaking through. Depth: scenarios get deeper, providing the same users with services along a deeper, stronger value chain; scenarios get richer, the value chain gets deeper. Breadth, with two methods: either go global, or open up new dimensions, like Mobike or ofo putting chips in bicycles and bringing in the Internet of Things, which you can think of as an extension; now there's also the industrial internet, IOT. None of these are in the traditional consumer internet or enterprise internet fields; they've all extended.

There's also one very important dimension, one I have fairly high hopes for in Chinese internet technology companies. In the past we did to C businesses; actually the harder thing is B-side business. The databases, cloud computing, chips and payment systems used by to C products are actually the lower layers of the ICT industry. If, after finishing the C side, you can move upstream into B-side infrastructure, and if that succeeds, it would raise the level of Chinese technology companies.

**(Q)** Do companies in other countries have more of an advantage in to B?

Amazon in cloud computing, Orcale and WareForce on the B side are all very good. Whether capturing the user dividend, marketing or social spread, you mostly have to go global before you can enter the harder work upstream.

> [Note: "Orcale" and "WareForce" are as in the original, meaning Oracle and Salesforce; other versions don't include this passage, so it can't be compared; the original is kept.]

**(Q)** TMD are valued even higher than American unicorns of the same period; where do Chinese companies' advantages lie?

We can't say we have much of an advantage. In ride-hailing, for example, regulation is local, and most countries' markets have been localized. If you look at single markets, China after all has a large population and fast growth in total economic output, so single-market valuations are very high, and China doesn't have that many car owners. Airbnb has globalized very well, but it's a relatively slow business. In the end, to achieve bigger results, you have to break through either in the value chain or in globalization. In the same few years, we can't count as having surpassed them by much, but at least we're on a par. In their first few years "BAT" were basically 1/5 of the market value of American companies; Baidu at its IPO could be benchmarked at 1/5 of Google's market value, which was already very high at the time. Now it has to do with the growth of China's whole economy, and also with the basic environment.

Copyright: © LookMinds. Our own edited text and translation, not a substitute for the original; rights in the original belong to the author and the platform. Please see the original link.