Episode 7: Yi Wang of Liulishuo on Teaching English with AI

Yi Wang is founder and CEO of Liulishuo (a.k.a. LingoChamp), China’s leading mobile learning platform for spoken English with over 70 million users. It uses speech recognition technology to enhance the learning experience and provide learners with measurable and proven results. Within a few months of launch, Liulishuo rose to the top of the Apple app store in China, and was recently ranked by CB Insights as one of the 100 most promising artificial intelligence startups in the world in 2018.

Yi is a “sea turtle” (海归, overseas returnee) who returned to China after studying and working in the US. He received his Ph.D in computer science from Princeton University in 2009 and MSE and BE in electronic engineering from Tsinghua University. Before founding Liulishuo in 2012, Yi was a product manager at Google’s headquarters in Mountain View from 2009 to 2011. Yi has also worked as a product director at AdChina, responsible for its performance ads platform.

In this episode, we discuss questions like: Why did Yi choose to leave his comfortable job in Silicon Valley to start a new venture in China? What challenges must “sea turtles” overcome to successfully start a company in China? What makes Liulishuo so engaging to its millions of users? Will AI ever replace human teachers?

Transcript

HANS TUNG: Hi there. Welcome to the 996 podcast brought to you by GGV Capital, and co-produced by the Sinica podcast. On this show we interview movers and shakers of China’s tech industry, as well as tech leaders who have a U.S. China cross-border perspective. My name is Hans Tung. I’m the managing partner at GGV Capital and I’ve been working on startups and investing in them, both in the U.S. and China for the past 20 years.

ZARA ZHANG: My name is Zara Zhang. I’m an investment analyst at GGV Capital and a former journalist. Why is this show called 996? 996 is the work schedule that many Chinese founders have organically adopted. That is 9 a.m. to 9 p.m., six days a week.

HANS TUNG: To us, 996 captures the intensity, drive and speed of Chinese internet companies, many of which are moving faster than even their American counterparts.

LIULISHUO: Where does Tina come from? His name is Dan. Her name is Tina. She comes from China.

ZARA ZHANG: Hi everyone this is Zara from GGV Capital. What you just heard is an app called Liulishuo, also known as Lingo Champ that’s changing how millions of Chinese people learn English today, using speech recognition and artificial intelligence technology. I first heard of Liulishuo from my mom actually. Like many Chinese people she learned English in school when she was young but when she was preparing for her first trip to the U.S. to attend my graduation she still panicked about her ability to communicate with actual Americans. So she told me she now uses an app called Liulishuo every day where she speaks English sentences into her phone and is scored on her pronunciations.

The app works like a game where you must pass a certain score before proceeding to the next level, just like Angry Birds. She is literally addicted to the app now and sometimes even asks for my help where she’s stuck at a certain level. And her spoken English actually improved a lot and she was able to make some conversations with Americans when she came to visit. So GGV is actually one of the earliest investors in Liulishuo. Hans and I are super excited to be interviewing Wang Yi, the founder and CEO of Liulishuo here in Beijing, at GGV’s Evolving Lifestyle event, a huge conference that gathers many of China’s top entrepreneurs and investors where Yi was a speaker. You might hear cars honking as we’re close to street and this is Beijing. Yi, we’re super excited to have you join us today.

YI WANG: Well, thank you so much for having me Hans and Zara.

ZARA ZHANG: So let’s start with your product. If I’m someone who is trying to learn English as a Chinese person how would you describe Liulishuo in three sentences?

YI WANG: It’s an AI English teacher on a mobile phone that can help. You learn English more effectively than with conventional methods.

ZARA ZHANG: And how many users do you have?

YI WANG: So far we’ve got over 50 million registered users across the world, mostly in China but also in over 175 countries..

HANS TUNG: Wow, 50 million registered users.

ZARA ZHANG: And a lot of them are paying, right?

YI WANG: Yes. As of this July we’ve got over 600,000 paying customers, and more interestingly, over 80 percent of them were paying for the very first time.

HANS TUNG: How much on average do they pay?

YI WANG: On average they pay between $60 to $80.

HANS TUNG: A year?

YI WANG: Well, that’s how much they pay for the first time, but on average a year would cost about $150.

HANS TUNG: I see. And how many years did it take you to get to 50 million registered users?

YI WANG: Let’s see. The app was first launched on Valentine’s Day 2013, so it’s about 4 and a half years.

HANS TUNG: When Jenny first talked to me about this investment, you just started.

YI WANG: That’s right. We were actually, I think GGV’s earliest investment in, maybe May of 2013. Jenny was saying, “Hey, you should take $1 million at least from us. That’s probably the smallest amount we can put in.” I said, “No, I’m thinking about even less.” So that was a funny anecdote..

ZARA ZHANG: And I read that in the first two months since you launched you rose to number one in the Apple App Store for paid apps in China, right?

YI WANG: Yes, that’s also a very interesting kind of a historic event, because at that time in April of 2013 there was an earthquake in Sichuan, in China, and before that we were a free app. Then we said, “Okay, we’re going to make it paid and we’re going to donate all the revenue.” We didn’t know how much it’s going to be, but all the revenue went charity for the people who were in the earthquake zone. So we did that, and very quickly our app rose to the top ten of all paid apps in China market, and it stayed there for three weeks. The peak was number two across the board just right after, what was it, Temple Run.

That was a pretty interesting event. Then we were near the top of the leaderboard for a while, and I think in August of 2013 when we launched the NGO version we decide to make everything free to sort of spur the growth and we went from there. So I think that’s incredible because learning English, many people consider it to be a boring task, and learning takes a lot of effort, and for an education app to rise to the top of the App Store is quite incredible.

ZARA ZHANG: What do you think, what it is about Liulishuo that makes it so appealing to people and literally addictive to some of them?

YI WANG: That’s one of the questions that we thought a lot about when we were starting out, because traditionally learning is kind of boring, and especially English. Many people hate learning English because they run into a really bad English teacher at school, so they swear on never learning it again, and that they’re done with it for good. But when we were starting out we saw the need. That’s something you really didn’t think about when you were in the U.S. So that’s something that I’ve sort of got off the plane and started really observing the local market when we saw the need. But then we thought, “Okay, mobile is surging.” Pretend you’re in May of 2012 in China, you saw this karaoke app named “Changba” really surging. At first I thought, “It’s stupid, who would be singing to their cell phones?” But apparently a lot of people did across different age groups.

HANS TUNG: Yes, I remember.

YI WANG: And I thought this built-in microphone thing is really unique. It could change people’s behavior. So we thought, “Okay, if they were so into singing into their cell phones, maybe they could also practice English. But then how can we make them stick? So we thought, Angry Bird was very popular at the time can we gamify it? What would be the key elements of gamification that we could use? Instant feedback. So what kind of instant feedback? Maybe some feedback on their pronunciation to keep their score, to give them some indication of how good their pronunciation is. At that time I called my college buddy Ling Hui who was a research scientist at Google in Mountain View back then, specializing in speech recognition and data mining. I’m like, “Hey, can you do this?” He said, “Yeah, of course.” And I’m like, “Okay, so why don’t we do something together?”

So in a nutshell we thought, “Okay, let’s build a very simple utility app, push some new content to people every day, give them level including games, give them feedback, and see if they stick around. At the first stage we were thinking first can we make it simple, but easy to use, so that we can get more users, accumulate the users, and keep growing. It’s fairly simple. So just like the founder of UC, Hexiaopeng, was talking about on the stage, at the first stage you just want to survive. So that’s what we did. And within the first week, one day we were looking at our traffic and it just shut off. Then we realized, within a few hours, people were sending us screenshots of us being featured on Apple’s App Store in mainland China, Hong Kong and even Japan. So that was our first big wave of users because of Apple’s recommendations.

ZARA ZHANG: So I think your company became profitable this February. Do you have any tips on entrepreneurs with regards to monetization and how you started thinking about it early on?

YI WANG: Yeah, I mean looking back I think I can share more about our growth history and some key decisions we made along the way. Before we decided to enter this market we asked a few questions. We thought, okay we’re going to do something consumer-facing, and we saw a big need of Chinese people to learn English. Even with very ineffective conventional methods people are putting like $5,000 or more a year into the traditional brick-and-mortar schools. So we were like, “Okay, clearly they are not getting much in return.”.

HANS TUNG: They’re not getting full benefit in return.

YI WANG: Yeah, statistics show that half of the people who paid that much money never go back again to the school after two months. Remember they pay for 12 or 24 months altogether.

HANS TUNG: They just stop going, even though they still have money there.

YI WANG: Yeah. I was talking to one of the owners of a very brand school and I said, “Dude, this is the gym model. You pay and you never go.” And he said, “Yes”, very bluntly. I clearly saw a big gap in the value exchange, how much people are paying and how much they’re getting in return. So I see a big market. I see huge inefficiency in the system. Then we got excited, because whenever you see this big gap in value exchange and a deficiency, it means technology can come in and help improve that efficiency. That’s very creation to us. All three co-founders of this company we were engineers by training, so we thought what we can do is we can use technology and product to improve the efficiency of that system, and we can get value out of it.

HANS TUNG: So what role does AI play specifically and what can of kind additional services can you provide that makes it sticky?

YI WANG: So I’m going to talk about AI in two aspects. One is sort of the more user value aspect, and the other is the business model aspect. From the user value aspect we realize there are two major trends, just like the theme of this conference. In China there’s a very popular terminology being talked about a lot in recent years called “xiaofeishengji”, consumption upgrade. We see two keywords in consumption upgrade. One is the quality aspect of that which oftentimes translates into a more personalized service or product. The other is sort of the time or efficiency aspect that people are trying to pay money to save their time, so that they can waste their time on the things they love. So these are the two trends. And these two trends are nowhere to be seen in education. It’s very expensive. What makes it more expensive is people not only pay a huge amount but they need to spend time to get a result, you can’t get it instantly.

So we thought what if we can reduce the price they pay and make them reach their learning goals faster. That will be a very compelling argument. So we thought what the future of learning is going to look like? We literally closed our eyes and dreamed about it for quite a while. We thought, two words, personalization and high efficiency, these are the two keywords. So we thought what could make it possible? We realized you can only go so far by throwing human teachers in front of the students. We’ve been using this model for thousands of years, and we see a lot of one-on-one models both online and offline, but they also only go so far, and by the way they have really bad economics. So we thought what’s next. We thought it should be something like the Disney movie Baymax where you have some intelligent virtual teacher. It understands you really well. It understands your learning better than yourself, than any human teacher possible, because of the interactions, you interact with this product.

HANS TUNG: And they have more knowledge, more data, so they know what suits you.

YI WANG: Exactly. And it never forgets. It serves you 24/7, and it can cross-reference from hundreds of thousands or even millions of users. Then, by the way, I’m going to throw this out to get your input, but I think this is a higher level of shared economy. When we talk about share economy we talk about sharing physical goods, space, whatever. Now this is a share of people’s knowledge. We are able to hire the best educators, cognitive neuroscientists, AI scientists, designers, product managers across the board, and they would never go to a classroom to teach because they thought it’s going to be a waste of their time, but now they are dumping their knowledge and expertise into this engine.

HANS TUNG: Everyone can have access to it.

YI WANG: Exactly. The variable cost is so low that it’s negligible. So this is a whole new model of really pushing the boundary of human learning forward..

HANS TUNG: And the engine can learn from working with so many instruments.

YI WANG: Just look at AlphaGo. It’s an extreme example but we really think that as time goes by the brain of this AI teacher is going to be more and more powerful. It can really surpass most human teachers, and we’ve already demonstrated that. Even before we launched this AI English teacher in July of 2016 we did a two-month experiment. We used each ETS’s TOEIC bridge test as a benchmark because it maps to the Common European Framework which is a canonical level system of language proficiency. The results that came in were very interesting. The users learning from our AI teacher on their cell phone were able to get up one level in their system with only about 36.5 hours of study within a two-month period, and they recommended at least 100 hours of study to get up to that level, so we pretty much tripled the learning efficiency. We think this is just an exciting starting point. We know our product has a lot of areas to improve still, but we think this is a very exciting direction.

ZARA ZHANG: So as someone from China who studies overseas a lot of my friends went to English schools like New Oriental Education and others, do you think that you guys can replace those schools with apps like Liulishuo, and replace actual teachers, or do you think they still have a role to play in this new world of AI?

YI WANG: I get that question a lot. I think what we’ve shown in the past a year or so is that for adults definitely we think the AI teacher can do most of the work, if not all. Today our users on average spend over five hours a week on the app, and they’ve been getting very noticeable improvement. Most of our users will say this really exceeds their expectations. They didn’t think that with a $15 dollar monthly subscription model they can get this kind of improvement. But if you go down the age group, if you’re talking about teenagers or even kids, I think human interaction is still needed, but the AI aspect can really help a lot. Let me give you another extreme example. We did a social responsibility project in the Qinghai highlands, in rural China, in a Tibetan orphanage school last year. They don’t have English teachers there. So we donated a hundred smartphones and we made our paid teacher product available for free to them. After nine months of using this product two to three times a week, each time for 45 minutes..

HANS TUNG: How many students did this?

YI WANG: About 100. And they are the graduating class of junior high school, 9th grade. Their average score in the English exam improved by almost 20 points on a 100 point scale.

HANS TUNG: Wow, that’s a huge improvement, 20 percent.

YI WANG: Yes, it’s pretty remarkable. And as I said their starting level isn’t that high, but still there are teachers and students who were very impressed. So that’s in an extreme environment. I don’t envision that our teenagers would only be studying from our AI teacher, but we actually have a lot of English teachers on our app as our users. They will say, “Okay, this actually saves a lot of my time. I make assignments on your app for my students to practice listening and speaking.” So I think it’s really not about replacing the teachers role but also actually alleviating some of that burden and help them spend their precious time on the other, more the motivation and personal character traits, and the other aspects of education, of a really well-rounded human being.

I think that’s a very valuable aspect of education as well. I think that’s going to take a longer time for AI to play a significant role. So maybe some of the U.S. audience didn’t know this but in some of China’s less developed areas like Yunnan province, 70 percent of their primary schools and secondary schools lack English teachers. In Heilongjiang province the number is 50 percent. There is really a huge deficit of English teachers.

HANS TUNG: So out of 1.3 billion people in China 500 million people live in the city, so there’s another 800 million people, and most of them don’t have access to English teachers.

YI WANG: Or just in general quality teachers. So it’s a huge problem..

HANS TUNG: So today you English and you can also teach other subjects or Chinese later as well.

YI WANG: I think so, yes, I think the general argument holds very well across the board. We think artificial intelligence embodied in user friendly services can really go a long way. It can help equalize the field so people with different socio-economic status can have equal access to quality education, and they can really bring down the cost and improve efficiency. Think about that. A lot of 6th grade students go to bed at 11:00 p.m., even at midnight. They have too much homework to do. A lot of that work is a waste of time. People know that but they still do it. So a lot of the problems can be solved by improving the efficiency, bringing down the cost and also equal access. I think AI can really go a long way. We didn’t think about AI this term particularly back in 2012, but we thought there must be some kind of technology with a highly interactive product getting people’s data. We sort of massed expert knowledge and baked into the product and then it can evolve.

HANS TUNG: So at one level you have English, Chinese, and other academic subjects that you can teach in school. Another way of looking at another level is there could be more professional usage with training for this type of service. Can you share more about along that axis and what new things could be done with your approach?

YI WANG: Yes. Let me give you a couple of examples. We work with a very well-known education expert named Lance Knowles. He lives in San Francisco, but in the past two years he spent almost half of his time in Shanghai and flying back and forth. His learning theory which we use extensively in our product called our RHR, recursive hierarchical recognition, it’s just one of the many learning theories you can find, but it’s a very unique one. He told me once a football club, a soccer club in Belgium was using his learning theory to train the players. And they’ve been training in more of a human centered way by breaking down the movements of the players and making them get better faster.

You can definitely imagine some kind of product, maybe even with VI or AI as an element to really help people’s professional sports training, with replacing the human coaching element with a gamified product and with technology. And another aspect is one of my friends who was doing the training of car mechanics, she actually got into a very in-depth chat with me the other day asking me could she use AI to really revamp her course. Now it’s very labor intensive, it takes a long time and depending on the trainers quality the outcome of the class there is quite a bit. We started talking about the huge market of professional training China and also how really AI and other technologies bring it to the next level.

HANS TUNG: For something as car mechanics that is quite complicated, how do you use AI, and virtual teacher to guide them through answering all the questions that the users may have, and that they may run into?

YI WANG: I think you can start from something relatively simple. We were talking about that for some checkup procedures or something that you would normally require memorizing a pretty thick manual, or something, you can actually make them some kind of games.

HANS TUNG: Help them remember critical information.

YI WANG: Yes, exactly. One key element of the Liulishuo experience on the mobile phone is that it’s in small bites. You can really use it in a fragmented time, and if you have five minutes you can fire away a few challenges, you’ll remember something, and next time you just pick up from there. So it’s really breaking down monolithic course into small pieces and use algorithms to stitch the elements together in an individualized way. I think that’s the gist of it. But also obviously across different domains and subjects we need a domain expert to come in and produce the content. That’s why we are an increasingly more and more a content heavy company as well.

HANS TUNG: A lot of PGC professional generated content.

YI WANG: Exactly. We had a community that was more UGC than we had PGC, so that’s a topic Zara and I talked about. And really starting from day one we were like a 100 percent tech company, and then we started to build up the curriculum design team, the content team, and so on. It’s really about looking at user needs, and building your product and service around that.

HANS TUNG: When users give you feedback that’s UGC, but when your professional content creators help you to create better content that becomes PGC, and that’s very valuable to get even more UGC out of it.

YI WANG: Yes, so content heavy. And also another interesting element of that is in our model not only the professionals generate content, the users actually feed their personal content to us. Collectively they create a lot of value, and you don’t get that value individually, so that’s a very unique network effect.

HANS TUNG: I remember you had told me earlier that you can learn English for a while, but if you need to use English to get a job in a job interview situation, it is a different kind of application of English skills, and it requires more practice and a finite amount of time to ace an interview. It would be nice if there were modules that someone can go through.

YI WANG: Yes, we have that kind of modules and we literally have examples of like a security guard at an office building training on our app for half a year and he got an office job, literally got a huge pay raise. We have a lot of different examples like that. We have grandpas retired, deeply worried about not being able to talk to his grandson leaving San Francisco, started using our app, and on his next visit they were finally able to communicate using English. I think he can be self-sufficient and just go shopping and everything.

ZARA ZHANG: When users first open the app they fill out a brief survey to indicate their interest and why they want to learn English. Whether you want to study abroad, or shop abroad, be able to talk to foreigners or get a job, pass IELTS, TOEFL, or whatever.

YI WANG: That’s right. We really tried to personalize the content for them and they explore, but yeah, we tried to make it relevant in a way you really see the exciting things you can do after you hit your language learning goal can be one of the most powerful motivations.

HANS TUNG: I want to switch topic to more personal questions.

ZARA ZHANG: Yeah, let’s talk about your background. I think you were the “gaokaozhuangyuan” of Hangzhou city.

YI WANG: Yeah.

HANS TUNG: You aced your college.

ZARA ZHANG: Yeah, that’s a huge thing, especially Hangzhou.

YI WANG: Yeah, with some luck.

ZARA ZHANG: And then you went to a Tsinghua, and then you went to Princeton for graduate school.

YI WANG: Yes.

ZARA ZHANG: So you speak perfect English now, but I remember you said before that when you first came to the U.S. to Princeton, your first meal was at McDonald’s, and you had difficulty communicating with the counter help at McDonald’s, which used very simple English, and that made you think about why that is the case. So can you share more about that experience and why do you think even though you got top scores in English tests in China it was still difficult for you to communicate with actual Americans?

YI WANG: I think that’s going to be a lifetime memory now, that was September 1st 2005, and I was at a McDonald’s near the New York airport. I was trying to just pick a few items, but that waiter didn’t understand. I finally said, “Okay, number two”, and I picked a set meal. Now I remember that I got almost perfect score in my TOEFL exam, only missing 10 points in the 677 scale. But I think at that time I was mostly doing drills in my grammar, or maybe writing, or something else, but my listening and speaking didn’t get that much exposure back then. So when I met my later advisor at Princeton I was most worried about his speech rate, can I really follow and have a meaningful conversation and discussion about my research topic with her.

But then I really spent an entire year watching the news, talking to my American fellow students, just that I get all the practice that I could, and that really helped. So I think it’s really about, not just getting your language education from the books and from doing drills, but really use it in practice. And practice makes perfect. It applies very well in a skill type subject like language. Unfortunately even today the whole education system and the setup of language learning courses in China is more like lecturing some kind of knowledge. The teachers take the center stage and just preach.

ZARA ZHANG: You just memorize.

YI WANG: So that’s really counterproductive. So when we started out building this app I had many of my fellow classmates in mind. I’m like, “If they would travel back in time and use my app what they need and what I need the most?” So we really thought a lot of listening and speaking practice inputs and outputs are crucial, and also helping to get them access to authentic dialogues. So we hired very talented writers to write authentic dialogues, and we have the top tier of voice actors and actresses help us record the materials. So we’re trying to give people as an authentic learning method as possible..

ZARA ZHANG: Sort of transporting native English speakers to people who may not otherwise have access to those living in China.

YI WANG: Exactly. That we think both helps in the learning setting and also from a motivation perspective as it gets them really interested.

HANS TUNG: So you were at Google before.

YI WANG: That’s right.

HANS TUNG: I remember for most of the last decade and at the beginning of this decade being a returnee, a sea turtle coming back to China to do an internet startup is not easy. Most people would consider that background, that grassroots enough to go “jiediqi” to know what’s going on. But it seems like with what you’re doing now, your returning background is actually a huge plus for you.

YI WANG: Yeah, that’s a really good point Hans. I think I did a couple of things right, looking back. The first thing is I really took my time before coming back to China, and even after coming back to China, to really observe and learn. I saw a lot of examples of Chinese techies with some kind of technology, coming back..

HANS TUNG: From California.

YI WANG: Yeah, using that as a golden hammer and looking for the nails to hit on. And the results weren’t good. I did things the opposite way. I had a U.S. patent in virtual router migration, in that working space, and I could have raised a huge amount back then, and it’s still a very hot software defined network. I could just do a startup in the Valley. But I didn’t do that. I thought, “Okay, China has huge market. Let me come back and really recondition myself find out the user needs.”

HANS TUNG: Right. To see the market and see what the users want.

YI WANG: Yes, and then wear my PM hat on and build a solution, build a product.

HANS TUNG: So having that PM background is actually key.

YI WANG: Yes, I think so yes. Now I think there are a few startups in this related field that started before my startup and that didn’t go that well, because they have only single founder type background, engineer background.

HANS TUNG: No product management experience.

YI WANG: No product manager experience, no co-founders, and they come back with a golden hammer style to build products to see if they can stick. So I think, yes, having great co-founders is a great plus, and also having a product perspective. If I weren’t doing this startup with this approach maybe I was not going to be Hui’s co-founder, and he would probably be a successful research scientist at Google.

HANS TUNG: Or do something else.

YI WANG: Yeah, so it’s really identifying the user needs and problems first. And then as you build the solution you sort of assemble the team.

HANS TUNG: To help you do this.

YI WANG: Yes. And also if you look at the spectrum, our approach is pretty Silicon Valley.

HANS TUNG: Yes.

YI WANG: We use a very AI, tech driven approach, product driven approach. So I think find the right problem to solve, and find the right path.

HANS TUNG: Yes, and people and tools.

YI WANG: Yes, it is really important. If I were to do more, I don’t think I would be the best founder for social networking or shared bikes.

HANS TUNG: Right.

YI WANG: Maybe not. So I think really just find the right problem, the right space, and you pick the right approach. I think your previous experience including your overseas experience can be a huge plus.

ZARA ZHANG: Would you say Liulishuo is a more product driven or technology driven company?

YI WANG: Honestly I don’t make that distinction. I think people don’t consume technology, they only consumer product. So good technology will be embodied in the product. Take Tesla as an example.

HANS TUNG: Right.

ZARA ZHANG: Can you talk about some of the iterations that you’ve been through with the product and the major decisions you’ve made along the way and things you’ve learned?

YI WANG: Yes, the first stage was very simple we’ve tried to build a very simple utility app that people want to use. And then that spoken English practice evolved into China’s largest language learning community online. Then we realized that people have the need and they are paying. They use our app daily but they would tell us, “Hey, I just paid $5,000 to that offline school.” So we thought, “Why don’t we offer a systematic solution that people want to pay for?” People don’t pay for this piecemeal “toy”, but they would pay for a systematic course.

So we went out talking to the top publishers, Pearsons, Oxfords and Cambridges of the world, and tried to license their content, but we realized all the famous textbooks were designed for the offline custom teaching environment. They weren’t fit for someone laying on their couch, studying in solo mode on their cell phones. We thought, “Okay, why don’t we build it from scratch if the perfect solution we envision doesn’t exist today?” So then we’ve assembled our curriculum design team which was definitely not easy, and we found our U.S. and China experts and started doing that. So the third phase is really going from just a utility app and a community into a full-fledged learning technology company, and with a systematic, leveled, AI English teacher course. That’s sort of the three steps we’ve taken so far.

HANS TUNG: You mentioned having that project management experience was important, so you don’t just take your technology look for a problem to solve. How did you gain that product management experience at Google, especially in the Silicon Valley?

YI WANG: I think the shortest answer is really follow your heart. Back when I was doing my Ph.D. in computer science at Princeton I remember I didn’t even know there was this position called Product Management back then. I was talking to a recruiter from Google and she said, “Hey, what would you like to do during intersession?” I was like, “I have this tech background, but I’m also interested in business. I like to solve real problems.” So she said, “Oh maybe you could try the product manager role, it’s kind of a bridge between both.” I’m like, “Oh, interesting.”.

HANS TUNG: But you hadn’t had business background before, not an MBA?

YI WANG: No, not MBA. I interned at Bain. I was a summer analyst there.

ZARA ZHANG: And you got an offer from McKinsey.

YI WANG: That’s right.

HANS TUNG: So you did a lot of preparation work.

YI WANG: That’s right. In preparation for my McKinsey interview I actually read the Wall Street Journal for 18 months.

HANS TUNG: You forced yourself through it?

YI WANG: Yes, to really be familiar with the business jargon and whatnot, so that’s a really good preparation.

HANS TUNG: Did you do any case study?

YI WANG: Yes, I did. In the summer of 2008 I was interning as Cisco writing my algorithm into the big fat routers, but after a busy day I would not go back to my apartment. I actually drove down the busy central to Stanford campus, and I would do a case interview practice with my fellow student and friends at Stanford.

HANS TUNG: Interesting.

YI WANG: Yes, we did that for the whole summer.

ZARA ZHANG: So you always knew that you didn’t just want to be an engineer, you wanted to be in the business world.

YI WANG: That’s right. I see a lot of really talented researchers, scientists and professors crunching out really beautiful technologies.

HANS TUNG: But they’re stuck in a lab.

YI WANG: Yes, they have no interest in starting a company. They just want to crunch out papers which I have not a huge deal of interest in. But then on the other side you see great businesses that could be even better.

HANS TUNG: If they had used technology correctly.

YI WANG: Exactly. They have the mindset to use it, and they know how to use it. So I see there’s a lack of bridge in between, so I’m like, “I can be that bridge.” I realized that I actually was the only Chinese that got an offer from McKinsey back in 2008 as a consultant, right after Lehman went bust, I literally had an interview with McKinsey the Friday after it went bankrupt. But then I got the offer from Google and I was like, “Yeah, if I could learn how to build a product from Google, that’s probably the best preparation you can get if you want to become an entrepreneur.” So that’s why I went to Google.

HANS TUNG: We see a lot of engineers, obviously including Chinese, that don’t have that business background, so they are “stuck” doing just technology and refining algorithms, and it’s harder over a period time of five, ten years to then become an entrepreneur, because they don’t have a comprehensive enough background on business to complement the technology.

YI WANG: I can totally relate to what you said. I’ve seen examples around that. I think if you have that inspiration you’ve got to cross that chasm at some point.

HANS TUNG: This is what we’re trying to do at GGV, we’re trying to figure out how to help engineers who want gain that biz background and practice, so your experiences are very valuable.

YI WANG: Honestly in a lot of engineers, whether they are Chinese or other backgrounds, I see huge potential. Just give them the right exposure, the right environment, and they can grow into a more well-rounded founders, I have no doubts.

ZARA ZHANG: Did you always know that you were going to return to China or did you also consider staying in the U.S.?

YI WANG: Honestly I always knew. But I still remember very clearly when I bought a little condo in Sunnyvale with my wife, we were talking about Google being such a great company we will at least work there for four or five years. But I left within two years.

HANS TUNG: Really.

YI WANG: Yes, even a little bit less than two years.

HANS TUNG: I’m surprised you had the desire to be there for four or five years and leave after only two.

YI WANG: I didn’t. So what prompted that change? I think the sense of my personal growth has started to plateau a little bit. And also when I visited China I saw..

HANS TUNG: Tremendous excitement.

YI WANG: Tremendous momentum and energy. You can’t see it but you feel it. So both factors combined, I thought, “If not now, when? Let’s just do it.”.

HANS TUNG: Interesting.

ZARA ZHANG: How did your lifestyle change when you first returned to China from Google, because our podcast is called 996.

YI WANG: I no longer own a car. I take a very crowded subway. I tried to save money and actually I lived with my college buddy who is also a returnee from Harvard doing startup. We shared an apartment.

HANS TUNG: This was in Shanghai.

YI WANG: Yes, but actually not in a prime location, so we had a one hour subway commute with two transfers at the beginning. Long hours, obviously. But it was great. I was watching people using their cell phones in subways. And I would buy the same kind of breakfast they buy at the convenience store, or even just on the street, and just really live the way they live and watch. I didn’t have the time to watch so much TV, but I would talk with them over lunch, and my younger Chinese colleagues would tell me about what they watch, and I would just listen to their conversations.

Actually the first spark came from my lunch conversation with my colleague. I didn’t start my company right away. I joined Ad China which was one of the leading ad networks back then. My younger colleagues would ask me, “Should we be watching American TV shows, more movies? We pay so much to different language schools, but we get so little in return.” I’m like, “Dude, you pay so much, and now you ask me should you be watching more TV shows and movies?” And I’m like, “Clearly you are clueless.” I started to look into this market and I saw huge demand, I saw how inefficient the current was.

HANS TUNG: So it wasn’t your first product when you first came back?

YI WANG: No, I actually thought about a lot of different directions. I even thought if we can do something with a better compression algorithm.

HANS TUNG: So something more technical?

YI WANG: To reduce the bandwidth of TV streaming, video streaming or something, or enterprise. Actually I explored a lot of different directions. Then I finally I zoomed in on language learning on mobile, because I saw a huge user need, and I saw a clear growth potential in mobile in general.

ZARA ZHANG: So would you say you decided to start a company and then went out to find a problem in the market, or was it you saw the problem first and then you wanted to start a company?

YI WANG: I clearly told myself it’s the right time to leave Google..

HANS TUNG: And do something.

YI WANG: And I had four choices I could stay in the Bay Area and I was getting a lot of LinkedIn private messages from recruiters. So I could’ve join a smaller startup, or I could’ve done it on my own in the Bay Area. Or I could come back to China to do both. I’m like, “Too many choices, let’s eliminate half of them, so I would just move back to China.” And then I said, “Okay, now you are back in China. There’s actually not that big of a hurry. You should take your time, really learn the market, and really find the right user need first than do other things”, so that’s what I did.

ZARA ZHANG: Google recently launched a new headset that can automatically translate everything you hear in a few languages.

YI WANG: Yeah, I saw that.

ZARA ZHANG: What do you think of that product, and do you think in this age with these products will people still need to learn foreign languages in the future?

YI WANG: I really want to try that. So later this month when I visit the U.S. I will try to get a pair. Yeah, I think that if that works well it’s going to help solve some of the problems. Like my parents when they travel abroad maybe they can use that. But I don’t think that’s going to completely eliminate the language learning market. In fact it’s unimaginable with technologies like that that the school system would eliminate English as a subject, for example. I think the natural way for people to communicate is through their traditional conversation kind of method. People value that. And also the process of learning a language is also a process of understanding that culture, and also it’s a process of building up self-confidence. We see that comment a lot from our users.

HANS TUNG: They may know the grammar, they may know the vocabulary, but they don’t have the confidence to say it and practice it.

YI WANG: Exactly, and they feel good when they actually are being understood and they are able to communicate to the people they are interested in. So I think it’s a learning experience. And then when you reach that point you’re able to do the things that you weren’t able to do before, so I think it’s the whole package. So yeah I don’t know if that answers your question.

HANS TUNG: When Jenny first told me about your team I was surprised that no one on your team had any teaching background.

YI WANG: That’s no longer true but at the beginning yes.

HANS TUNG: Back in 2013.

YI WANG: Exactly.

HANS TUNG: And so because teaching English has been, as you said done, for over a decade there are a lot of academic study on it, and what is the best approach and so forth. Now those traditional schools may not have the ability to innovate but they understood some theory behind teaching English, or teaching in general. So how did you initially develop that prototype product and handle those iterations initially not having that background to get to some kind of product market fit?

YI WANG: The honest answer is we just did what we feel is right.

HANS TUNG: As users.

YI WANG: Exactly. As someone who learned English the hard way. We thought we weren’t trying to build the silver bullet at the beginning. I still don’t think we have a silver bullet today. It’s an evolving process. The product is never finished. So at the beginning we thought, let’s just push maybe one dialogue of his/her level and interest per day, and it’s interesting to them, they can practice, get feedback, that’s good enough. That doesn’t require that much of curriculum design, or deep sort of research and education. But now we do need that. So we built our curriculum design team at the end of 2014..

HANS TUNG: How big is that team now?

YI WANG: About 40 people.

HANS TUNG: Wow. Very sizable.

YI WANG: Yes. I actually don’t think Elon Musk had all the right technicians and mechanics, engineers when he started out Tesla. You build the team along the way. You bring the right people at the right time. I think if we were to get the best curriculum design experts back in 2013 we wouldn’t be able to get them. So I think you should build the team along the way.

ZARA ZHANG: Can you talk about the geographical spread of Liulishuo’s users and what kind of users they are? Are they students, or white collar workers, what’s the breakdown?

YI WANG: We have one third of our users being K12, so they’re under 18, and we have a little bit less than a third users being college students, and bit more than one third are white collar workers or adults.

ZARA ZHANG: And do you have students outside of China?

YI WANG: Yes, now we have less than 10 percent of our users being abroad, so they’re mostly in China, because you would need to be able to read Chinese.

ZARA ZHANG: So the app is only in Chinese now.

YI WANG: The interface is in Chinese only for now. In China we are all over. We have our users in first tier cities, but also second, third, fourth tier, and even the countryside.

ZARA ZHANG: And you mentioned that your company can serve as a force for equalizing education within China. People in Tibet may not have access to good English teachers.

YI WANG: That’s correct. We see huge potential there. This year we have set up a dedicated team working on social responsibility and working with NGOs, improving access to quality education in rural areas. We are hoping to reach 100 schools by the end of this year.

ZARA ZHANG: Would you consider expanding to languages other than English?

YI WANG: Yes, it’s always on our roadmap, so that’s a plan and we’ll do that.

ZARA ZHANG: What are three new things that users can expect in the next year?

YI WANG: Let’s see. They will definitely see more intelligent, efficient AI English teachers on their cell phones. Number two, they will see some very cool, state of the art, kind of technology that’s related to language learning. I think they will also see some of our new exciting products, aside from this app.

ZARA ZHANG: I think Liulishuo also has the world’s largest database of Chinese peoples pronunciation in English. What do you do with all that data?

YI WANG: We crunch the data and we keep improving our language assessment engine. Today we have, what we believe, the most accurate speech recognition engine of Chinese people speaking English thanks to that huge dataset. And also building on top of that dataset now, our engine was not only able to accurately give feedback on people’s pronunciation, but also on grammar, on vocabulary, but also fluency and coherence, basically all major aspects of people’s speech. We also developed assessment engine on writing. We’ve been building a full stack AI engine to help people learn a new language.

ZARA ZHANG: Yeah, so it’s easy to tell people when they’re wrong, but I think it might be harder to teach them what is right when it comes to pronunciation.

YI WANG: Yes, that’s right. That’s an area we’ve been making a lot of investment. We have recently set up a small office in San Mateo with a former AI scientist from Facebook being our head, and we’ve been hiring very aggressively, both in China and the U.S., to really get the best people who are excited about what we do, and who can also contribute with their talent to build the next generation of technology products and content. So I think these are the three pillars to solve the problem you mentioned.

ZARA ZHANG: How big is your U.S. team now?

YI WANG: Right now it’s a few people. We hope to get to 10 by the end of this year. But we are very highly selective, we want to hire the best.

ZARA ZHANG: That’s really exciting.

YI WANG: Yeah.

ZARA ZHANG: I wanted to talk a bit more about monetization, because a lot of westerners may have the impression that Chinese users are less willing to pay for knowledge and content, but I think that’s been changing drastically in the past few years.

YI WANG: Totally.

ZARA ZHANG: You see all these “zhishifufei” projects like Dedao or Ximalaya where people pay subscription fees to access knowledge. I think Liulishuo is another reflection of that. Do you find it surprising that Chinese people are paying so much for knowledge?

YI WANG: Yeah, totally. I think it’s an unstoppable trend thanks to the surge of easy access to quality content as well as mobile payment. On top of what you just mentioned, these are really good examples, but I want to point out one thing which is education is very unique in China.

ZARA ZHANG: Because parents care so much about it.

YI WANG: Exactly. Chinese value education more than probably anything else in prioritizing their spending. 75 percent of the total spending of raising a child in China was spent on education. That’s a mind-boggling figure if you talk to someone in the U.S.

ZARA ZHANG: Yeah. Another education company that’s been making a lot of waves is VIPKID. Do you see them as a competitor? How do you think your product compares to theirs?

YI WANG: I don’t view them as a competitor because they focus mostly on kids, and so far are paying customers are mostly adults, or young adults. Number two, we take very different approaches. They are trying to bridge human teachers and students. They are more like a platform. We think we are on a different “innovation curve”. We think we are trying to do what human teachers can do, or do it much better than the human teachers. We build AI teachers. So I think the approach is very different and the gene, the DNA of the companies is also very different. I think both have their sweet spot.

ZARA ZHANG: What would be your top advice for “liuxueshen”, or overseas returnees, Chinese students who study abroad and want to start a new company in China?

YI WANG: Call GGV. Yes, number one, call GGV, talk to GGV. Honestly, I think we were very fortunate to have GGV as our very first investor, and being our continuing supporter. GGV’s very unique perspective, in both the U.S. market and China, and also its understanding in both technology and commerce, I think it really makes a very unique experience. By the way it has a really good reputation across both markets and it helps. Secondly I think it’s really about finding your calling. If you are just a hardcore technologist or engineer then you may have a different decision-making process than someone who just graduated from MBA program, for example.

Identify your role and what you aspire to become. But regardless if you think China is a growing market which I totally agree to, then really take the time to understand the market, even if your role will still be CTO, you still want to understand the market. In the past 18 months or so you see a huge wave of new AI startups, a bunch of technologists with their golden hammer looking for problems to solve.

HANS TUNG: Right. They don’t have a good end user usage case.

YI WANG: Yes, I think a lot of them will have a really hard time. You really want to get the order straight. You want to identify the problem first then build the product to solve it. Turning things the other way can be challenging. As you go through this thinking and analysis process it’s also a very good way to find the right people and build a team. And I would strongly advise against being a single founder these days. I think having a good team is half of the success.

HANS TUNG: What do you think is the optimal mix initially? Two or three people?

YI WANG: I think two or three, these are two good numbers.

ZARA ZHANG: You didn’t jump straight into entrepreneurship when you came back, you worked at an online advertising firm first, and that helped you to get to know end users.

YI WANG: Yes, I did a lot of preparation. I prepared heavily for my McKinsey interview and the consulting aspect of that, and also in my last year at Princeton I took only one class which was a high-tech entrepreneurship class taught by a very experienced veteran who used to teach at HBS, and we used HBS case studies. So that really helped prepare and develop my way of thinking. But coming back to your question I joined Ad China not because of their technology, but because of its reputation in marketing, operations and sales. These are the things I know I weren’t definitely exposed to at Google and in the Valley, and I know I wouldn’t need if I were going to be the CEO of a new company. So that’s sort of all the preparation I needed.

HANS TUNG: That’s smart. You were well-prepared before you did this?

YI WANG: I tried.

ZARA ZHANG: Shall we jump into the quick-fire questions? The first one is an entrepreneur that you admire the most and why?

YI WANG: I admire quite a few people, but if I had to pick one I would say Jeff Bezos. I don’t know him personally yet, but I know he’s a Princeton alumni. The key thing is I think he’s very visionary. He started from building an online bookstore, but I think he always had a very big dream, and ambitious goals. You don’t read a lot about him. He doesn’t give speeches that often. He just keeps..

HANS TUNG: Plugging away.

YI WANG: Yeah, exactly. Plugging away and just making things happen. And once you build a reputation, a track record, even Wall Street understands his way of doing things. So the valuation, or whatever, it’s just a sign and it reflects people’s confidence in his approach to doing things, so I really admire that. And also I always make this analogy that 99 percent education companies, whether online or offline today on the market, they are more like the Walmart model. Their core competency is in operation efficiency.

They need to try really hard to squeeze out maybe a little bit of profitability, and they employ a large workforce. They throw more people at the problem. But we are definitely more of an Amazon model. Our core competency is technology, product and data, that’s also our barrier. And once we hit profitability, which we did this past February, then you have a much nicer gross margin, and the unit economics is much better. So we definitely admire that model better. And we know given the scale, the revenue, and the size of the company, Amazon is definitely a very nimble company comparing to the traditional approaches.

HANS TUNG: And people still work very hard even though they’re very big already.

YI WANG: Exactly. Work hard, be focused on the most impactful problem to solve everything. We try to model us after that.

HANS TUNG: Share with us something you read recently that inspired you a lot. It could be recent or a while back, things that inspire you.

YI WANG: Yes, there’s quite a few things. Some of them are in Chinese and people may not resonate with it in the U.S., but I would say you know I recently reread the book from Ben Horowitz, The Hard Thing About Hard Things. I really like that book. I always enjoy reading it, especially the first four chapters. For me that’s very relaxing. It’s his problem, not my problem. It’s a tough experience, but at the same time I’m kind of traveling, I’m putting myself in his shoes, I’m traveling through time, and I’m like, “What would I do? What decision would I make?” This is what I make had I been in that situation?” I always draw strength from his experience.

HANS TUNG: You’re not the only one who’s suffering.

YI WANG: No, definitely not. And comparing to his problems back then, I’m like..

HANS TUNG: You’re in a much better position.

YI WANG: Yes, exactly. So I always tell this to my colleagues, things are never easy if you want to make something new, so just let’s keep plugging away.

ZARA ZHANG: What keeps you up at night and what motivates you to get up in the morning?

YI WANG: I sleep pretty well. I go to bed pretty late but I sleep generally pretty well. I think that’s a good thing.

ZARA ZHANG: What’s a thing that you worry about?

YI WANG: I think at this stage one of the things I think about the most is really how to build the best team to really make the future of learning a reality. Because my partners and I have a deep conviction that education, or a better term would be learning, is definitely long overdue for a revamp. In the first three industrial revolutions they revamped a whole lot of industries but not education, so I think the fourth revolution is definitely going to change. It’s both exciting and also challenging. It’s definitely an interdisciplinary effort, so that’s why now we have Professor Nick Turk-Browne who is an expert in cognitive neuroscience from Yale as our advisor. We have Professor Kai Li in computer science from Princeton as an advisor, but also we have a professor Dan Schwartz, the dean of the School of Education at Stanford as our advisor.

We’re getting experts from different disciplines to work with us, to really think out of the box how the future of learning should look like, and then make it happen. I think it’s a very exciting thing to really get the right people together, and still being efficient and productive, and making progress is a challenge. I’m still learning it from different people including my advisors at GGV.

HANS TUNG: So when you wake up in the morning what gets you excited?

YI WANG: It’s really just thinking about how lucky myself and my team is in the history of human beings, really. I was reading A Brief History of Tomorrow.

HANS TUNG: Yes. That’s a good book.

YI WANG: So we were very lucky being born in China at this age and also have this background, going overseas and really opening our eyes and seeing what’s possible. And also we had a technology background and it’s all blessings. And now we’re in this market in China right. Everything is booming and we’re starting very low which also is a blessing for anyone who wants to make a difference. I think almost all the stars are aligned to just make it happen. So these are some very exciting times.

HANS TUNG: If anybody asked me the same question that’s an answer I’d give as well. We’re very, very lucky.

ZARA ZHANG: One last question. What is one thing you believe in that other people may not?

YI WANG: I believe in maybe five to ten years, the majority of subjects people learn will be possible to augment by artificial intelligence in a major way. The scope is definitely not limited to skill based subjects, but for all subjects.

HANS TUNG: If that happens it could be the single most important factor to change the inequality in access to good teachers.

YI WANG: Exactly. I think technology, or more specifically artificial intelligence powered learning products, will be the biggest game changer for education equality.

HANS TUNG: Especially on Android, on very cheap smartphones that can easily be sold in third world countries.

YI WANG: Exactly. That’s super exciting.

ZARA ZHANG: All right, we’re very excited to see where Liulishuo will grow in the next five to ten years and thank you so much for joining us.

HANS TUNG: Thank you. It was a pleasure.

YI WANG: Thank you for having me.

ZARA ZHANG: Now you will hear a short interview with Jenny Lee, managing partner GGV Capital, who drives our investment effort in AI and frontier tech, and what made her want to invest in Liulishuo.

ZARA ZHANG: Hi Jenny, thanks for joining us. You lead GGV’s investment in Liulishuo, could you share a bit more about why we made that investment and what you saw in the company?

JENNY: Thanks Zara. When I first met Wang Yi almost four years ago this was when he first started the company, formerly called Lingo Champ. At that time the company was based in Hangzhou, there were two or three of them, and they were still trying to encourage on of the co-founders from Google to come back and join them. He’s really the NLP expert behind the voice engine that they have. I think what was really attractive about the business model at that time was Wang Yi was thinking about revamping or disrupting the whole education sector by combining gamification elements to his product, as well as leveraging the voice engine to automate the process, and thereby removing the reliance on having extra teachers teach English. The mobile phone is an excellent platform for the interaction.

And when I first took a look at the product I was pretty impressed. I used to look at gaming investments as well and I think that that combination of making learning language fun through gamification using the mobile phone was really the way to go. So that was a long time ago, four years ago, back when no one was talking about AI. Today I think after the last four years of hard work the team has really built a great product. They’ve also built a great technical team, globally from U.S. and from China in the NLP space, so we’re excited about the investment.

HANS TUNG: Was there anything about the team itself that attracted you? I know you liked the product, but what about the team themselves?

JENNY: Yeah, this team is a very interesting team. So Wang Yi and his other co-founders Ben and Hui, both have elements of global education, they have work experience working overseas, but being Chinese they understand the pain points that Chinese have when they’re learning a new language, in this case English. So the right complementary background in, both technology product sense, the right appreciation for this need and pain point I think was really great as well. Wang Yi is very aggressive yet a learning CEO. He never ceases to amaze me. At first we really see the investor in the investment, and you can tell that he really wants to learn.

He’s the one that has his own views, but at the same time where he’s not sure or he doesn’t have experience in, he will ask 100 guys if he can, and if he has a 1000 guys to ask, he will ask a 1000 guys. And that’s how he learns, from getting input from people more experienced than himself. He’s not going to just take what you give him and accept that. He will think about it, he will challenge you, and then he would talk to another 100 guys, but when he thinks you’re right he will change very fast. So I think the bet is not just on the product and the sector but also on him. He’s a very special CEO in my opinion.

HANS TUNG: That’s a sign of good young CEO who can do that.

ZARA ZHANG: What’s the biggest change you’ve seen in him since the investment happened.

JENNY: I think that over the last four years he definitely has grown. Back then he was learning how to hire, when to fire, how to incentivize. I think over the last four years definitely he’s getting a lot more experience having seen and met so many people. So as a CEO I think this is critical, assessing, hiring the right management team. I think he’s done that really over the four years and really matured. And so his assessment now and picking the management team members to join him has got better and better. Our most recent hire is in our view a super CFO, was part of the GGV ecosystem that he managed to convince her to come out of retirement to join him in this growth going forward.

So I think that’s one, really learning, managing, and finding the right talent. I think it’s important. And then the second part he really matured in is really focusing and saying no. In this new category where it’s education plus AI you could do everything. You could target adults, you could target the kids, you could target K-12, and you can also change a product from learning English to learning Chinese. You can learn Japanese, you can learn Korean, because the voice engine itself is actually very flexible. But knowing when to do what I think is very important. And so he has also over the years learned to say no, to cut up projects, to leave them for later to develop and really focus on what’s core, which right now is the China market for English learning for adults.

We may go up and down each group but I think that China by itself is an extremely big market, so we should take on that market first. We do have all the building blocks and the back end of technology to switch the engine to multiple languages, even when they decide to go international. I think on that front it’s never easy for CEOs to say no to things but he’s really matured along the way as well.

HANS TUNG: Do you see the product being able to teach other subject matters besides languages down the road?

JENNY: Well, I think the way the engine is geared for is actually for language right now. I think to go from language to teaching, say physics and chemistry and math is a little hard. So a little bit far-off. But I think if you look at the demographic itself the business isn’t just about education or learning a language, it’s also evolving into a lifetime of knowledge and learning. Between education and learning, learning is a much bigger business, a much longer time in terms of LTV, so I think that’s a big opportunity there.

YI WANG: What kind of advice have you given to the company along the way?

JENNY: Given that we were a so early seed investor, as I mentioned we helped him interview candidates, we helped him think through packages for how you should incentivize, how you should hire people, when to fire. I think the whole Aesop setup was actually one of the value add that our portfolio services team helped him to get off the road and started. The other aspect is that the company over the four years also had very good successful fundraising rounds. Always fundraise when you don’t need the capital, and so in that sense we have also helped him to think through financing, when to have financing, which are the partners to choose from. I believe that we have also built a very good, complementary shareholding investor base for the company.

ZARA ZHANG: Thank you Jenny.

JENNY: Thank you.

HANS TUNG: Thanks for listening to this episode of 996. By the way, we also produce a weekly e-mail newsletter in English, also called 996, which has a roundup of the week’s most important happenings in tech in China. Subscribers have told us it is informative and fun to read. The newsletter also features original content and analysis from Zara and me. Subscribe at 996.GGVC.com

ZARA ZHANG: GGV Capital is a multi-stage venture capital firm based in Silicon Valley, Shanghai and Beijing. We have been partnering with leading technology entrepreneurs for the past 18 years, from seed to pre-IPO. With $3.8 billion in capital under management across eight funds, GGV invests in globally minded entrepreneurs in consumer, internet, e-commerce, frontier tech and enterprise. GGV has invested in over 280 companies with 29 IPOs and 22 unicorns. Portfolio companies include Airbnb, Alibaba, Ctrip, Didi Chuxing, Domo, HashiCorp, Hello-Bike, Houzz, Keep, Slack, Square, Wish, Toutiao, Xiaohongshu, YY, and others. Find out more at GGVC.com.

ZARA ZHANG: If you have any feedback on this podcast or would like to recommend a guest please e-mail us at 996@ggvc.com. This podcast is co-produced by our friend and business partner Kaiser Kuo, the host of the wonderful Sinica podcast. It covers China’s economic, political and cultural issues.