
Interviewed by Hans Tung and Rita Yang.
Walking on the streets of any big cities in China now, you will notice that there almost seem to be a spending war between EdTech companies, who are buying up ads from giant outdoor billboards to tiny screens in elevators.
EdTech is a sector with a notoriously high CAC, that is customer acquisition cost. Unless you have built a free tool that is so successful and widely adopted, that people are open to try whatever you put in front of them. This is exactly how China’s top learning app, Zuoyebang (作业帮), literal translation for that is “homework help”, managed to grow from a free photo search app to China’s largest K12 online education platform with 20 million transacting users.
For the first episode of our special EdTech series, you will be hearing from the founder of Zuoyebang Hou Jianbin, whose personal belief about education significantly shaped the way he built this company. Born and raised in a small town in China, He was a straight A student and has his life completely changed by the national college entrance exam. Got into China’s top university and a job offer from China’s search engine giant Baidu, where he built the product that eventually became one of the world’s most valued EdTech startups. Zuoyebang was spun off from Baidu in 2015.
For the second part of this episode, you will also be hearing from GGV managing partner Jenny Lee. Jenny led GGV’s investment into Zuoyebang in 2016. She shared what made Zuoyebang stand out among all the photo search apps that were out there, and what keeps Zuoyebang competitive as Covid-19 pushed a lot of new capital flowing into the market.
For the full transcript of the show, go to nextbn.ggvc.com
Join our listeners’ community, go to nextbn.ggvc.com/engage
Today on the show we have Hou Jianbin, founder and CEO of Zuoyebang. Zuoyebang is the largest K12 online education platform in China. Known for its photo search app that allows students and parents to get homework help and live streaming class, the company has more than 50 million daily active users and 170 million monthly active users. Its transacting users have grown 10 times over the last 2 years, reaching 20 million in total. This past summer, its transacting users have grown almost 4 times comparing to last year, reaching 7.8 million. Zuoyebang is a GGV portfolio.
Before founding Zuoyebang, Jianbin run Baidu’s suite of knowledge products. He is a straight-A student all the way from primary school to high school. He holds a bachelor and a master’s degree from Peking University. He is also an avid marathon runner.
[Rita]
Currently, the penetration rate of EdTech products in China is around 10%, which means 1 out of 10 students in China have used EdTech products before. That number, you once said in an interview, can grow to 30% in 5 years. How did you arrive at this prediction?
[Jianbin]
My estimate is based on only 2 things, 1 is the 90% mobile penetration rate in China, another is 50% of all students in China taking supplementary classes outside school, we expect this number to reach 60% to 70% in 5 years. Half of the students would be taking these classes online. That’s 30%.
[Hans]
7 out of 10 families in China have used Zuoyebang’s free photo search tool before. The massive user base you built upon these free tools is foundational to the company’s current success. Did you always know that you wanted to build an EdTech company and see free tools as the best way to acquire users? Or is it more of an afterthought?
[Jianbin]
Both. It is both intentional and natural progress of the company. We always knew there’s huge potential in building China’s biggest photo search for questions across subjects. When we tried live-streaming classes, it turns out it was very closely linked to the photo search tool. So we turned the tool into an important growth driver for live-streaming classes.
[Rita]
The majority of your transacting users are from the user base of the free tool model. This works out for you very well in an industry where CAC (customer acquisition cost) is high everywhere. What’s your secret in turning those free users to paid users?
[Jianbin]
Three things. One is to have the best photo search product in the market. We are 80% of this market and growing. It shows how much we continue to invest in this product. The second is to build a strong supply of curriculum across different subjects. Now we have more than 200 kinds of curriculum, covering all 12 grades. Last but not least, is the backend that connects our search tool and the live-streaming classes, how the user experience flow between our products and when to push what. These are all very important.
[Hans]
When and how did you know that live group classes will become the battlefield?
[Jianbin]
If you find yourself on a track with no real competitors, you are probably on the wrong track. This is particularly true in the highly competitive market in China. So, when we realize that group live streaming class makes a lot of sense in terms of business model, we knew it’s gonna get ugly.
I have three criteria for what makes a great EdTech product. One is product experience; how much learning can actually happen. Second is if the team can constantly deliver a new curriculum both in terms of content and tech. Lastly is if the business model makes sense. Some people might think 1 on 1 class provides the best user experience, but you cannot find enough good teachers to deliver these classes. That’s why the group class makes sense. With this model, the CAC and gross margin will improve over time. We started group classes fairly early in 2015. There were already people doing that, but the scale was limited.
[Rita]
In the 1st stage of Zuoyebang, you said that building a user base with free tools is like a lightning war that needs to finish fast, where the competition in group live classes will drag on. What is the specific organizational adjustment you have to make as a leader for this transition?
[Jianbin]
The biggest organizational change is that we merged the photo search team and the group class team. It is now one business unit instead of two separate teams. In other words, the search tool is in service of the group class business. It is our core business. We used to have each team collaborating; now it is driven by one business. It is more efficient this way. For a very long time, we were all in for the photo search product. By focusing on that product and growing users, we raised the bar pretty high and it’s harder for competitors to catch up. The same tactic won’t work for group classes. It is a product that requires continuous investment in content and operation to deliver the best user experience. We need to be patient and do the right thing at the right time.
[Hans]
Any difficulties in merging these 2 teams?
[Jianbin]
Nothing more difficult than the daily struggles of running a fast-growing company. The market is so competitive and exciting that we are always, always in some kind of adjustment.
[Rita]
K12 education is an extremely fragmented industry in China, even though it is more accurate offline than online. One of your competitors, who started offline and have done incredibly well on its online class, Xueersi, seem to do well. It is part of the public company TAL with a lot of capital and resources on the ground. How do you see this particular competitor, especially in the lower-tier cities?
[Jianbin]
Xueersi was definitely the pioneers for group live-streaming classes. They started doing it as early as 2009. The parent company TAL is already very successful in its offline tutoring centres. Despite its leading position in the market, its market share is still quite small, given how big the market is. The growth of offline centres is a very different game than online. For tech companies like us, we are pretty lightweight and adjust more quickly to the market.
Another thing worth noting is that our user base for group live streaming class is quite different than Xueersi. Most of its live-streaming class users are in first to the second-tier city, while ours are in lower-tier cities. Our users are pretty representative of the overall composition of china’s students. It’s not because we are more popular among lower-tier cities but that’s how the entire user base look like in china. This is also because 70% of our search tool users are from lower tiers cities.
[Hans]
During COVID-19 in China, because of the shelter-in-place policy, the users for Zuoyebang’s live classes grow 15 folds. With millions and millions of students online at the same time, how did Zuoyebang manage it?
[Jianbin]
Well, it is all about the daily grind. All the hard work we put in every day since the first day of our existence. We were very intentional in building a strong team from day one. 2 out of 4 co-founders of this company are developers. We only hire the best out of the best for our team. Also, Zuoyebang is a tech-driven company as you can tell from our photo search product. It was a pretty tech-heavy product. In short, a long-term commitment to building a world-class tech team is how we were able to handle the 15 folds growth during the height of COVID in China.
[Rita]
From last year to now, your team has grown 6 folds, reaching almost 20,000 people. How did this shape you as a leader?
[Jianbin]
The most important change for me personally is how I spent my time. I used to focus more on the business side of the company. Now I have to think about the organization. When you have a 20,000 people company, you can no longer lead the company with just business. You need to run a highly effective organization.
We were growing too fast in the last couple of years, so there’s a lot of organizational catchups left to do. We have put more structures and systems in place now, including company culture, value. We have also bumped up the internal IT system on our priority list.
[Hans]
Zuoyebang as your first startup, what experience would you like to share with us?
[Jianbin]
I’m not that successful yet. The limited success I have had over the past 5 years is probably more luck. I tend to be more practical when it comes to building companies. First, you need to identify a problem that is big enough for a valid business. Next is to decide if you can put together the right team for it.
I think we did identify the right problem to solve. The key thing is that we didn’t know that everything can happen so quickly. Unlike businesses that can grow exponentially, people usually grow more linearly. So, we have to make sure that we can catch up with the organization set up and help our team to evolve as quickly. Not easy. The good thing about us though, is that all our co-founders are still around.
[Rita]
Hans, actually, this is a question for you. You met Jianbin being around four or five years ago, where you had a very exciting conversation around the edtech company that he’s building. Can you share with us? What was your thoughts back then? And how the company has grown over the past few years has, you know, surprised you or you kind of always knew that it’s going to be that great.
[Hans Tung]
I was introduced to Jianbin by Erica and Jenny, Jenny helped to pioneer our EdTech investment practice out of China first. So, I was intrigued about what this team can do. My key reservation was that this is a team that grew up inside Baidu. So as a standalone company, can they fight grassroots style and be able to go toe-to-toe with other people, without a lot of support as a self-sufficient, independent company, that was my number one concern. After meeting with Jianbin, I was impressed by how hungry and how thoughtful he is. He knew how to take the risk but he carefully planned things. So, he is very methodical to the empire to have what he wants to build. And the free app is very attractive, it’s very easy to use. Therefore, you can acquire a lot of users easily. You have a question that you’re not sure, take a photo, set it up. Someone’s going to respond to you. So, it is such a natural way for users to use it, then that he can sell additional premium services on top of that. So, it’s a model that we’re familiar that can work if the team is hungry and can do the grassroots thing of guerilla marketing and win. One thing I did not worry about was that given the background, they know how to manage scale when they can reach scale, and they have the good back end to support their growth, what I wasn’t sure of is whether they can fight on their own knuckle to knuckle, hand to hand combat, and can still win. And obviously, Jenny and Erica, a lot of confidence they can, were very willing to be supportive. And I have to say that what they have achieved today is beyond my imagination. So, I’m very impressed with the progress and growth momentum they have done in the last four or five years. Obviously, COVID-19 played a role but how they respond to COVID-19 was even more impressive. And now it is one of the most respected and valued companies in China. So, we’re very happy for them.
[Rita]
We heard that you are not a big fan of luck. You were the only person who didn’t gamble when the whole team went to a casino. You believe in hard work and the 10,000-hour rule. You were a straight-A student from a very young age. How does this personal belief shape the vision and culture at Zuoyebang?
[Jianbin]
Great organizations are usually also unique in their own ways. The founding team might be what makes the company unique. For me, I have three personal beliefs that probably shaped the way this company work. First is that I believe that greatness is nurtured, and everyone can be shaped into greatness. We often ask ourselves what the value is we actually bring to people. As an education company, we know that to improve the way people learn could change their lives. Some people believe that you need to be born into greatness, we happen to believe that anyone who is willing to put in the hard work should have access to better ways of learning. The second thing I believe is that excellence breeds excellence. By hiring the best talent, we make sure everyone we serve is inspired to do better. The third thing I believe is that great organizations are driven by its own vision and mission. Its employees are also self-starters. We hope the way leadership works in this company is to inspire people into excellence instead of pushing them into it.
[Hans]
What motivates you?
[Jianbin]
To push my limits.
Recently I am reading The Aims of Education by Alfred Whitehead. He thinks that the aim of education is to develop oneself. This is something I believe deeply. The way I define self-development is first you know who you are, and what excites you. The goal can be grand or small, one must have one to live a meaningful life. Then you are willing to work for it. This is what education means.
An example of me doing this is that I started running. I was very bad at running for a long time. I run 400 miles last year in total and wanted to sign up for a marathon this year. It might not happen because of COVID. But this is how I keep myself energized and I hope to bring this to more people by building this company.
[Rita]
You have a very high level of self-discipline, and you always expect the same from the people who work with you. That is a very common trait among successful entrepreneurs. What would be your advice for founders like yourself, who have high standards for themselves and the team?
[Jianbin]
I don’t think I am very disciplined. I have a colleague who always sends everything way before the deadline. I do think the more responsibility you have, the less time you have for yourself. Everything comes at a price. If you are a celebrity, you almost need to give up privacy. If you’re in finance, you cannot make a mistake on numbers. If you build a company, you will have very little time for yourself. Young entrepreneurs need to think it through before they jump into entrepreneurship.
I don’t believe in giving advice in general, especially for adults. People are all individuals with his or her set of values. Sometimes when I talk to my new employees, sharing all my 11 years of learnings. I am actually not sure if they can really get it. Just like when I was thinking about doing startups, Robin Li (from Baidu) asked me, have you really thought through? Doing a startup is very very hard. And I know how much he went through, so he was not being dramatic. I thought I had made a fully informed decision. Now I think about it, I had no idea the kind of “hard” he was referring to.
[Rita]
Quickfire questions:
Who or what was your biggest teacher?
[Jianbin]
Experience. When it comes to wisdom, some people are born with it. Some people can learn from the way other people live. I happen to be the type that has to learn by experiencing it myself.
If you can invite 3 people to a dinner party, alive or dead, who would they be?
[Jianbin]
My family.
[Rita]
What success mean to you?
[Jianbin]
This might sound pretentious. First of all, I am not successful yet. And I don’t believe there’s such thing as success. What is success? When I was in high school, success was getting into a big-name school. I did that, but what’s after that? If you think about it, everyone will be dead and that’s when we can draw some kind of conclusion about a person’s life. With all that happened in a person’s life, I just don’t think there’s such thing as success.
What would you write on your tombstone?
[Jianbin]
This is a person who has lived to the fullest. I just cannot live the fact that I didn’t give everything I could in everything I do. This goes back to my student’s life. I knew that I could be No.1, so I will give my best and be No.1. I cannot live with “could have” s. It is a lot easier to become the No.1 than to keep the No.1. So, I am always worried that I did not give my best. It’s very hard for me to not care about the outcome. I want to have a good outcome for my best work.
What is the book you would recommend?
[Jianbin]
The Road Less Travelled.
[Rita]
Hello, Jenny. Welcome to the show. Again,
[Jenny]
It’s good to be here.
[Rita]
So tell us when and how did you first meet Jianbin from Zuoyebang.
[Jenny]
I cannot remember the exact date. But it has to be sometime in 2014. The first meeting I remembered is in the most popular Hotel in Beijing, where a lot of the deals are being done. And we actually had our first meeting in the lobby, just talking about case consideration when he was about to spin off Zuoyebang from Baidu.
[Rita]
So back then, you’ve already started investing in EdTech. Was it a very hot scene in China back then?
[Jenny]
No, the hottest scene back then was in the shared economy. But that was everything from ride sharing, food sharing, even a pet’s cleaning. For EdTech, it was still relatively early. Traditional investors were more investors in offline education. I think online was really new at that time.
[Rita]
The founding team of Zuoyebang was first-time founders, even though the team used to be within Baidu. The business model of a company that does photo search tools back then requires a lot of money to run. So what did you see in him or the company that made you decide to invest in him?
[Jenny]
The photo search app at that time was quite competitive. Typically, investors saw this tool could be an entrance into the path for the students. At that time, there were a lot of other startups. Zuoyebang was not the first to do this. In fact, I would say they were in the second cohort of companies to try and attempt this. So even prior to the meeting with Zuoyebang, we’ve been tracking the space, we’ve been looking at the competitors, we’ve been trying the products. But I would say that, when Zuoyebang launched a product, we have not made any investments, because we felt that, to do this, it’s really not an EdTech service deal. It is a technology. So it’s tech before education, and it has to be perfect. Perfect in the sense that one, the accuracy of photo recognition has to be there, because you’re scanning the questions, and the ability to pick up the questions, how you recognize the question, whether it’s a math question, English question, physics question, actually is very important. And then the second part is that students generally are very impatient, if you’re doing your homework, you just want it to be done. So you could go off and play games. And the latency, the time it takes for your questions to be scanned and answered, has to be done within the microseconds. Students are not going to wait seconds, or even 5,10 seconds for answers to be written. So, the processing time has to be short. And then finally, it has to be accurate. if the students go through this process, and not have an accurate response, then over time, they will move on to other apps. So it was that time where we will try and we had like the standard question, but where we were testing out all the apps with the same set of questions to look at the what I mentioned, recognition, the reaction time. And finally, the accuracy piece, and through tracking these various apps, Zuoyebang came up tops, in terms of the ability to do all three. I think the team from Baidu is very strong testimony to this, the team prior to founding Zuoyebang, Baidu was already doing a lot of these fundamental building blocks that could today, you know, being what we see as the back end for Zuoyebang. Of course, there’s a lot more refinements and variants over the years. But definitely those were the early blocks. And it was really critical to make sure that, as a tool that’s out there in the market, consumers can download 10 tools and try them. But this is like the last one that remains standing. So that was very good validation.
[Rita]
For the photo search tool, Zuoyebang told us 7 out of 10 students in China have used that tool before. So I guess it speaks to how good the product is. EdTech has become an extremely hot market recently given the COVID-19. What are some of the competitive moats do you think Zuoyebang has to keep its leader position in a highly competitive, also heavily capitalized market?
[Jenny]
We started looking at the whole EdTech scene four or five years ago, Back then there was no COVID, we did not predict there’s going to be a COVID or acceleration, what we saw was that technology really has come quite far. And it can be now utilized to a great extent to bring content online, reducing the reliance on schools, physical schools. And then two, the technology, it could also in some way automate the learning and teaching process to free up the teachers’ time. And so, we can scale the teachers a lot more so. So that was done four or five years ago, right.
And I would say that, you need time to build a competitive moat. And therefore the fact that four or five years ago, this was not a super-hot sector, it was a sector that did allow startups like Zuoyebang actually to have three, four years to really build this competitive moat. And in this case, the competitive mode is a few for sure. I think one, the backend database has to be big enough to cover all the curriculums across the different school systems in China, the technology has to react fast enough. And then of course, to go from one user to seven users out of 10 users definitely takes time. And that’s why it’s the network effect that’s been built over the last four or five years. I think today for a new startup who wants to come into this exact space that they are in, it’s going to require not just capital and technology, but also a lot more innovation. And it’s not going to be easy. I think right now they have a very, very strong leadership position.
[Rita]
Besides Zuoyebang, GGV also has two more EdTech companies that also helped democratize the access to great teachers. One is Vedantu in India, another is Ruangguru in Indonesia. So what were some of the biggest takeaways from the Zuoyebang deal that helped inform decisions in investing in Vedantu and Ruangguru?
[Jenny]
I think GGV’s view on EdTech is that it is a global trend, it’s a long-term trend. And there’s a lot of global insights that we can learn from some of the market leaders in China like a Zuoyebang. But ultimately, the EdTech flywheel has to localize the investment, the companies that can be successful have to be localized, it has to be local companies. Why? And I think, number one, education tends to be very tied to specific and local countries in specific curriculum. Number two, it’s a very important pillar of any country, and therefore, there will be regulations that will support the growth of local companies. And then thirdly, I think that we are talking about the EdTech part. And so, when we are trying to move schools online, we’re trying to move the whole teaching experience online. It also requires the existing back end infrastructure and network to be able to support this.
So, I think, while we look at China’s lens and say this is a huge market, we also have to understand that there are fundamentals in place to push this forward. The back end is great, we’re now talking about 5G. So that was never the issue. In China, anyone with a phone, a smartphone can actually get online access to service. And number two, payment is now so widespread and has so many different forms, anyone who has a WeChat account, Alipay account, any other forms of mobile payment, actually can pay for the services as well. So that’s kind of the China context. And I think it’s important to understand that and not just blindly said, this is the Zuoyebang for China, or this is for India, or this is a Zuoyebang for Indonesia. So, I think that’s the lens we have, right, the global lens, the global insights. And then when we look at local markets like India, Indonesia, we then have to take on what I just mentioned, I think number one, we were looking for local companies, local champions, local leaders. I think in India, we see potential as two reasons why we also felt that it’s time for live streaming classes to take front and center station. India is because of the network. The second issue I mentioned, the network is great, it opens up. And then thirdly, payment, payment is with the unified ID in India as well, it’s lowering the barrier to paying for such services. So in that sense, I think the live streaming model would also go very well in India, which is why we chose the local leader here. In Indonesia, though Indonesia’s situation is a little different from China, India, the wild Internet has arrived, it’s only arrived at certain parts of Indonesia and certain demographics. And so, for education to reach more widespread, the model has to take on a little bit more of, you know, not an online model, but more of a download model. And that’s why I think Ruangguru has the local champion, our model to go to market, it’s basically based on more of a lighter recorded video where users can download it in one shot, and then view it at the offline scenario. So again, these are little nuances. But it’s important to understand that once you can understand that you can also see that over time, these models are all going to evolve and morph as the underlying fundamentals change. If broadband arrives in a very big way to Indonesia, then I think that the live streaming model will also take off and become an interesting user experience for the learning part.
[Rita]
It’s really helpful. What are some common traps you see for EdTech companies, especially those who want to start from offering free tools and gradually build their business model?
[Jenny]
You know, I’m one of the few investors who actually likes tool businesses, because I feel that, you know, tools, the best use, done and design developed really goes a long way to alleviating or resolving a specific need. In the case of Zuoyebang, it has homework assistance. So I think it’s a great way to really understand users, and then remove the barriers of payment to allow, you know, very good utility like that to be delivered right to all the users who need it. And if the application of features is really good, then you’re going to see this huge pickup. Because then people will download it, and they don’t have to worry about payment. So, I think that has value for the consumer, it’s a very good way to go about launching the service, especially if you can identify what is that niche? Or what is the specific pain point that you’re trying to resolve? I think the traps that the companies go through is that they get all ecstatic about the free tool and become free for a very long time and maybe forever. So, I think that’s the biggest issue. And so, we will encourage companies that it has to be a free tool. But you know, over certain times, I think startups are not doing this, you know, completely social, ready to have to survive, they have the teams to support. And so, I think any viable company needs to think about monetization. And so, I think that to think about monetization later is probably a mistake. I think that for startups, when they’re thinking about going to the market, they should have the product. I think free is one way to test the user response. But over time, if the tool is of value, then it needs to be paid. And it could happen in situations where the tools remain free, but the community itself pays for the services, like in this case, with Zuoyebang, they’re paying for good content to teachers online. Now, in some cases, the tools can come and beat it with so much content, like the Ruangguru application you can have users paying for via subscription, access to content as well. So I think there are different types of monetization, the right thing to do is actually to think about all things on the first day, but then have a three year, five year even 10 year horizon, to build up the entire go-to market roadmap. But you need to have that roadmap, because having that roadmap will help you to guide what the second generation version of the tool is going to look like.
[Rita]
Okay, last question. So K12 is a big chunk of the EdTech market. What are some other emerging trends in this sector? Either within k12 Or besides k12? that excites you at the moment?
[Jenny]
Over the last four years, we have invested in now over 10 companies, closer to 15 companies in the tech scene. So I would say, if we look at just K12 categories, first we were looking at kind of in the classroom, and that relates to homework, which is a high frequency activity, we have also invested in the activity of K12 after school. So essentially, whether it’s after school tutoring, and we go from curriculum base to non curriculum base, which then centers around holistic education, I think learning shouldn’t just be about test taking and getting the right grades. I think that all the countries around the world are trying very hard to ensure that our next generation also develops in an all around fashion. So I think around holistic education areas like logical thinking, arts and crafts, music, programming, those are all areas that are also being disrupted in different forms from technology. And so that’s definitely an area that we have invested, and we will continue to spend time on. And then I think that we are also looking at technology, which is technology that moves away from reliance on schools, and then completely from teachers as well. And we have a new investment that’s in that category that actually leverage technology without teachers in the loop that can be leveraged in schools all around the world. That’s a different story that we’re going to talk about.
[Rita]
Interesting. All right. It’s always great to have you on the show. Jenny, thank you so much for your time.
I would like to receive news and updates from GGV.