udemy | Interview with its Co-founder & Chairman – Eren Bali
In San Francisco, we meet co-founder and Chairman of udemy, Eren Bali. He shares his story how he co-founded this startup and how the current business model works (supply and demand side), as well as what the current plans for near future, and some advice for young entrepreneurs.
The transcript of the interview is included below.
Martin: Hi, today we are in San Francisco with Udemy. Eren, who are you and what do you do?
Eren: Hi, my name is Eren Bali and I’m the co-founder, chairman and former CEO of Udemy. We are a market place for online courses. We have thousands of instructors around the world who are really passionate about their subjects and they are so skilled so they come to Udemy to create courses and share it with the rest of the world.
Martin: And, what did you do before you started Udemy?
Eren: I think while I was in college, I already knew I wanted to be an entrepreneur. So there wasn’t a lot of other things. I was in college, I went to school in Turkey, I grew up in Turkey. I went to college there to study computer science and mathematics. And in the last year of the college, we started our first company with my co-founder, who’s still with me on Udemy.
Initially, we were doing some other projects. We had created a 3D simulation for constructions. It was how we started. But I always wanted to work in education, so we built a company that were close, similar to Udemy in terms of supervision in Turkey, almost 8 years ago. So we launched it, but we realized that wasn’t the right time and the right place.
It wasn’t the right time because all this different section wasn’t ready, bandwidth was expensive and audio conversation wasn’t really working well. And Turkey wasn’t the right place. We needed tons of influences to make the idea work. And the product wasn’t right either, because we relied on niche market place. And in market place, liquidity is the biggest problem, meaning if a user comes to the site, there should be enough courses interesting for them so they will stick. And if you make it live, it’s even harder to make the liquidity first. So, there were a lot of mistakes and we had to shut it down in Turkey, but me and my co-founder still believe that was kind of a neat idea and it should have existed. So, we moved over to Silicon Valley, it took me 4 years in the process. And I worked as an engineer, like one of the first employees at a startup called SpeedDate. We did a pretty good job there. Actually the funny thing is, we built SpeedDate from the early code of Udemy. We had a video conferencing tool that we built for our live classrooms, we coordinated a video dating application for SpeedDate. It was a fun experience but we eventually knew that we wanted to try this idea again. And this time, it worked out.
Martin: OK. Great!
BUSINESS MODEL OF UDEMY
Martin: Can you tell us how the current business model of Udemy works?
Eren: Sure. So we have around 15,000 instructors, we have published courses on Udemy. They choose their course subject and build a course. Typical courses are like video base and some are interactivities, like quizzes and discussion boards and things like that.
But the big difference between the site like Youtube is, it’s not just small short-form content. The course I designed, to take you from knowing nothing about the subject to a decent level, right. So, and when they start publishing course, they can either publish it for free, I think it’s free for them. If they want to charge for their course, we take a cut from that. So in that sense, it’s pretty close to Apple’s app store model.
Martin: So basically it’s just a revenue share model?
Eren: Yes, revenue share model.
Martin: Okay, the instructors can free to choose what he wants or is there any fixed fee involved?
Eren: There’s no fixed fee. If it’s a free course, we think it’s good for the world, so we kind of do our part in it. If it’s paid, then we start making money.
Eren: Yes. We have over 4 million users right now. I think 50% of them are in the United States, 50% outside the US. We’ve seen a lot of growth in Germany, Spain, some of the European countries and even Asian countries, but probably Germany is like the second largest non-English market, at this point. So, usually our top verticals are development, design and business.
Martin: What do you mean by development?
Eren: Development means learning how to become a web developer, IPhone applications, mobile applications like game programming, those are usually the most popular subjects. And then there are other technological subjects like softwares, Excel, Office. Business categories are entrepreneurship, marketing, finance. And design categories are always pretty strong. But beside these, there are others like hobbies, like photography and yoga. Those are also big, but usually as I said technology, business and design are the leading categories.
In terms of demographics like, right now, there are more male users than female, which is the complete consequence of the development category. So unfortunately, there are still a lot more men, who are interested in development than women. We also want to contribute in changing it. But that’s the current state.
There are categories that are uniform, non-technical categories are uniformed between men and women. And in terms of Asia, we see age ranging from 18 to 50 – 55. There are more of younger users but that’s more because they are more younger as in internet. But proportionately it’s kind of interesting, some of the older user are more heavily represented in Udemy community than the rest of the Internet.
Martin: Interesting. When you started this type of market place, what has been the major obstacle for creating a lot of inventories, so people who come to your website see your value proposition?
Eren: That’s why we’re straight forward. In any market place, like Udemy, there are 2 types of users. There is an asymmetric market place. The most important problem and probably the only problem that you should really worry about, is the chicken and egg problem. If there are enough courses, why would the students come, and if there’s not enough students why would instructors create courses. So this is a huge problem in the beginning, and I said that kind of converge into competitive advantage as you become larger. Because once there’s Udemy, it’s really hard for the competitor to do this, because all types of instructors would want to create their courses on Udemy. So, as I said there’s an obstacle in the beginning that becomes a big advantage as you grow.
Martin: How did you take tackle in the first place?
Eren: We did a lot of tricks. The number one thing is you have to worry about is liquidity, meaning when you are an instructor can you get him enough users and if there are users can you give them enough courses. So, we tested a lot of things. We tried to do some BD, Business Development, it didn’t work out. We try to launch it kind of vertical, that didn’t worked out as well either.
I think what worked for us was going and doing a lot of unscalable things to make the first video instructor successful.
Martin: What does it mean?
Eren: Actually, I kind of give you the birth story. So, I was in a breakfast meeting in one of our investors house. And AirBnB was just getting traction at that point. They were big but they were just getting off the ground.
And I was asking these questions to everybody. I would say how did you get the first 100,000 users? How did you solve the initial traction problem? And he kind of.., Brian gave me such a suggestion: “do things that don’t scale”. So meaning, there are things like SEO and advertising are scalable things in getting users. Those tend not to work in the early days of a startup because most of the scalable methodologies were better on scale.
As you get more inbound pages and you get inbound links, your SEO goes up and advertising is better. But they don’t really work in the beginning. And there are things which are unscalable, which you can not do as a large company but for small service can do those. And those unscalable ways of doing are actually what makes you successful.
So what AirBnB was doing was they were going to houses of each host and taking professional photos of room. They were making the place look really good. And this is completely unscalable, but it was creating a magical experience. So we kind of mimic the same idea. Like I said earlier, when he said those 3 words that immediately clicked for me. We went back and said, you know what let’s help the creation of the first few courses.
We want to create 3 entrepreneurial courses because we knew people who could teach those. Even that didn’t work out, because it was too much work. Initially without their proof point. So instead, me and my co-founder, Gagan, organized entrepreneurial events. The first one was called “Raising Money for Statups”.
And instead of asking people to create a full course, we brought the subject in 6 pieces, like the first one was getting leads, then doing a pitch tag and then like the last one was closing the round. So you kind of add what the Laissez faire raising money course would be.
And when we invited people, we said “Okay, you are going to talk about this particular subject. So when the event finished, we were able to video-taped everything, edited it and converted it into a course. Actually we start charging for it. It was 30 dollars. Even with the first course, we got a lot of sales from it because it was a pretty good course. It was the first time where people uses or get practical advice about raising money. Not like high level “follow your passion” bullshit. It was like, this is exactly the email you sent, this is the type interview you get, this is when you email them back. This was very tactical. So the tactical aspect really worked out.
It was 2 more of those events, we get the first 3 paid courses. Before, I was kind of passive about that, we were also crawling all the open courses from universities and we were publishing it on Udemy with their permission, that helped us get some traction but they were all free courses. It was still not proofing financial model here. So the first 3 courses we made were proofing the fact that there was this one person in the world who could buy them. And they were actually like almost 1,000 people who bought that courses.
Martin: Did you just wanted to check whether there was some kind of willingness to pay for such a course?
Martin: Okay, and this was the major force why you decided not to put it for free but want to charge?
Eren: Yes, I remember there was an email discussion with our investors, and got their opinions. There was a big debate whether course should be free or paid. My visual was that, although a lot of entrepreneurial talks like this are free, they’re usually motivational, they’re not tactical.
When I go to a conference and give free talk, I don’t talk about, like this is like how we do. It’s just motivational, the most strategic maybe. But the tactical piece was missing and if we charge 30 dollars for it, and if it makes 0.1% easier to raise money, it’s definitely worth it. And we tell them that we would like to see whether people would pay for it. I said people would pay and they were super happy with it. Nobody thought that 30 dollars was too much for it. So, we launched 3 of those courses in entrepreneurial subjects and then we found some newsletters which has relevant traffic to Udemy. So we started promoting those courses from these newsletters. And people who signed up to this newsletters really like the courses. And then, what we did next was, we went to some other instructors who weren’t affiliated with us and we said, would you create a course. We have all this distribution of newsletters signed up to promote your course. They said yes, why not try it. They promoted their courses in the newsletters and we went to more newsletters and said, you know what, we do this course, would you mind to promote and get their share? They said yes and we kind of, similarly we jumped back and forth between the supply and demand. As I said, the key things like liquidity and making your early customers happy, right. We made them really happy by actually do a lot of distribution work for the instructors and a lot of content work for potential distribution channels. So we were like being their technological platform between the distribution channels and the supply side, the content creators. At one point, we had enough users accumulate on Udemy that we didn’t rely on external distributions anymore. So many people create their courses, we had enough users on Udemy so that their course would be successful.
Martin: At what point did you recognize that you’ve reached that critical mass?
Eren: The critical mass is not just one single number. So, similarly as a market place, you go up in phases. The first phase is you do whatever possible to get the first few customers and first people on the supply side happier, right.
And then, over time, like at that point, you kind of switch into kind of more a scaled approach to getting instructors and the scale that meaning that, it was not like me and my co-founder meeting with each instructor in their house to convince them. Instead, we had people, like business people who are reaching out to instructors to kind of convince them to teach courses. So it was kind of like sales model of reaching out to actively instructors.
And in the user side, we start things slightly more scalable and slightly more scalable thing. Like maybe some SEO worked out and we had more partners, we had like an affiliate network and we had a growing organic traffic that was the largest traffic driver.
And then over time, that changed also, because right now there’s so many people come to Udemy to teach a course. I think every month, 10,000 people start creating courses. Not all of them do it, but a lot of people are willing. So we don’t have to do outreach anymore.
We have a large content team whose job is just helping the instructors that already coming. We don’t reach out to people anymore, except for strategic courses. As I said, now is a different phase and overtime, even that phase has changed. Now, honestly the only thing that the content team can do is barely: they do work and create new verticals, and they do more highly work in managing verticals and working at the course that’s scalable and article variable like data science team analysis, like which course are doing better job and which courses are doing worse job.
So, that has come to a new phase, we’re data driven and big data and even more scientific. As I said, it’s phases, it’s just as you grow, you start changing how you do stuff.
Martin: Let’s talk briefly about the corporate strategy of Udemy. I totally understand that once we have this scalable mass that you create a competitive advantage. If you put it into a single company perspective, how do you perceive having other competitors like Coursera, etc.? What’s your competitive advantage over them and what makes you be able to build up on that?
Eren: That’s a good question. I mean there are a lot of competitors. There are primary competitors which mean there are other market places for teaching. Right now we are the largest, we are leading that one, so we’re not too worried there. But then, there are publishing companies, companies which produce content and sell it.
So in users side, they might be considered as competitors. And then there are companies like Coursera, which we are not really competing with in the market. I think it’s just our psychological competitors. We are more like friends with them, so we kind meet each other and help each other if possible. Those are really different. Coursera for example, they are focused on higher education. There are companies which are really focused completely on K-12 kind of academics. K12 international academy is focused on K-12. And Udemy’s focus has always been skill based adult training. So it’s not K-12 or higher education, it’s like after that. We believe that learning has to be lifelong. And our courses are usually more tactical, you learn how to build an IPhone application, you don’t learn the computer science fundamentals as much. Although we have courses around it, but that’s not our strong suit. Our strong suit is, I’m going to learn something very practical and I’m going do something with it.
So that has been large enough differentiator, just some notes that we start a lot early than those other companies. I think we could say, we kind of, we start this online course market. So there were companies before us doing some sort of learning but constructive online course that anybody can take, like we were the company that initiated that move, and hopefully we have some help in the popularizing concept like MOOCs. Coursera just came around 2 years after us. And when they launched, we already have our model which was working. I think they look at our model and kind of, look higher education could do the same thing.
Martin: You beautifully described the different phases of a company. What would be the next phase of Udemy?
Eren: I think the next phase is right now, we solved the problem of creating an ecosystem about casual learning. We have done that. As instructors are making enough money to live their lives, we signed about 6,000 checks every months, like 6,000 people are making money from us every month. And some of them are making millions of dollars. And for some of them it’s like passive income.
But I think the most important that we’ve created is the job. Teaching online is formidable job right now. So the next phase is taking on demand learning from where it is right now to its next level. So Udemy is and should always be the frontier of learning online, especially on demands.
Because you can also do 1-on-1 live classes, that’s a different world. Skills on demand learning is our focus and we are awaiting a new phase to do this and we are building a lot of science about what makes a course better. So we obsessively analyze action that users are doing during the courses and now we can actually we have thousand of courses, so we have that unique position where we can crawl what aspect of the course make it stick better.
Martin: I would call it educational analytics.
Eren: Yes, more like educational analytics, but people call it big data education. But then you see, that a lot of assumption people make about education is incorrect. Even like assumption made by professors who have been doing this for a long time or educational scientists, they have a lot of small studies, observations, now we have real data. And we see a lot of them are incorrect, and you see it in a lot of large scale across verticals. It’s not like one vertical, we can see like this learning theory can I apply it in this vertical but not that vertical. We are in the position, I want Udemy to become a company which is pushing the boundaries in on demand learning. This is the main thing and other stuff is opening up internationally, so we opened our first Europe office.
Eren: In Dublin, I really want Berlin. I went to Berlin a year ago, it’s a great city. I really love that. Hopefully, we have an office there too at some point, but the first is in Dublin for more practical reasons. Europe market is very important and so we start opening up to South America, Europe, East Asia.
And the other thing is actually making the model work in some other niche verticals. We are pretty strong in the top 3 – 4 verticals but I would expect musicalcourse would be very successful in Udemy. So we already have somebody, it’s not like as strong as we wanted. Like maybe hobby courses and some other sophistication base courses. There’s always new vertical that we wanted to have, build liquidity at.
Martin: In terms of market development, how do you perceived the offline education versus the online education, and even maybe a specific sub segments in between?
Eren: If you look at the education market or the world virtually first divided by age group. There’s K-12 was a separate base, right. Second it’s about learning release, like education means something different for them. And there’s higher education, college, maybe grad degrees, and there’s lifelong learning.
And lifelong learning can happen at home, you’re learning for yourself, it could be at work, or it could be in some kind of institution. If you look at them, for each of the verticals there’s an offline element to it, and there’s an online element to it. So in K-12 for example, still dominantly offline. There’s some online movement but it’s very small in proportion.
Lifelong learning actually is the part where online gets a much larger chunk. That’s why we actually focus on that. And higher education is kind of small slowly transitioning, it’s a slow transition, it’s not going to happen overnight. For us, with lifelong learning, like offline, there’s some offline component, you could go to a certification program in a place, or you could take a high end boot camp. Those tend to be really expensive and limited to a very small local.
We are happy to admit that might be a better experience, so I don’t claim that learning online is better than in person, because if you can find a top Stanford professor, or like an amazing person to sit next to you and give you a course directly to you, it might be better if you can access that. But it’s going to be 100 times more expensive than an online course. And you’ll be kind of limited by the time and location, so to me that education is for elite group of people in the world.
And for everybody else, that’s not a solution, that’s why we started Udemy. So, I think on demand and offline are sometimes used in conjunction together to a hybrid learning. But for majority of people in the world, I think online education is the only viable way to get them up to speed with the skill that they need for jobs or even hobbies. I think online is the only viable way to do that.
Martin: In terms of certification of a specific skill, do you think that having this kind of certification later on in the future or do you really think it’s just about showing people who you have the skills?
Eren: This is the point where I and the rest of the market is in disagreement. A lot of companies are trying to get certification up, maybe try to raise the money, all the inventors were saying, why don’t you use certification, I would only invest if there’s a degree or diploma behind this.
I kept saying no, does not make sense. Because people look at universities and they’re just trying to move that model while you go to the university, there’s some learning there, there’s a certification that you have diploma. It does work in a small constraint environment, but when you go to internet, things changed. You can’t move things from physical world and just digitized it.
So, in world of internet, you can access a lot more people. Accessing a lot more people also make it harder to make a reputable certification. Because Stanford or Harvard has a pretty strong brand in their certification, if Harvard has 5 million students, they wouldn’t be nearly as good brand. So the value of the brand of your certification and the number of people who are able to get it, they are inversely correlated. So that’s why, I think if you want to help hundreds of millions of people, you can’t at the same time try to create that high, strong brand in your certification.
So I thought, in the world of open digital education, these 2 things should be separated. There should be companies like Udemy, whose job is only to help you learn your subject because you want to learn it for something. I don’t care for what, but you want to learn a subject. I think our company or other organization which feel whose job is going to be certifying that you are good at a particular skill, assessing your skill. I think trying to merge them together, will really break the experience, it makes both parts function worst.
So from that, I believe we should separate, so we never get into that. We do like course completion certificate, like light weight badge, but we’re not trying to make it completely a diploma. So we never thought this is critical.
What we do more often is a real life certification, which is accepted in the market like Cisco network administration certificate, or Adobe design certificate, or Microsoft Office certificate. So we try to get, we have a lot of courses that we are preparing for those existing certifications. So people can go and take a course in Udemy to prepare themselves and they go to Cisco to get the network certificate. And Cisco network certificate will always be more valuable than a certificate that we made up.
ADVICE TO ENTREPRENEURS FROM EREN BALI
Martin: Let’s talk about what type of advice, you can share with entrepreneurs or what type of lesson you learned. I mean in one of your startups you failed, then you work in another startup and now you’re quite successful with Udemy. What are the major learning that you can share with our readers?
Eren: Sure. There’re a lot. So I start what comes to my mind first.
- First of all, when you build a startup, you can look at how things currently work or how other people do things. And you can just assume the world is in a certain way because it is. That’s the wrong way. Because as an entrepreneur, you have to forget how currently the world works and say how should it work or could this work. Because there are thousands of reasons why Udemy wouldn’t work. That what some investors brought up to us, and they said that they passed on us, they say this is never going to work. Because most of those reasons they were coming up with, were just how things are currently were happening. But if you kind of forget about them and you say that there are millions of people and there are also millions of experts in the world and billions of people who love to learn from them. Why wouldn’t this exist? Just starting the first principle and kind of ignoring how things are currently happening, is really important.
- The second part is that you need a laser focus in execution. Pretty often people try to do a lot of different things. I mean, testing a lot of things is Okay, but you have to be laser focus on what you execute. For example, for us in the beginning, we said that we need to figure out the early traction problem. We have to solve the chicken-and-egg problem. Everything we were doing was to solve that problem. And once we grow older, there was a scale problem and quality problem. So we always were very focus on what the problem we were solving. You said “be focused” is like a dummy advice, because people think they have to do less things, sometimes you do more things but your purpose has to be extremely clear. That’s the focus you have to be very clear.
Martin: And did you already think, like when you wanted to create this kind of product market fit, in terms of what are the scaling problem or quality problems, or did you just take one problem at a time?
Eren: Especially when young, we try to take, just tackle a problem at a time. Because I can admit in the first year, there’re for the courses we had the quality wasn’t very high. Because we were just barely trying to make the course successful, so we had to have a low barrier, but as we grew we start increasing the barrier and started really focusing in the quality. So you can’t just try to achieve everything at the same time. That kind of change per company.
If you’re building a utility tool for example, building a network and utility tool are very different. If you’re building a utility tool, you have to be laser focus on the experience of users from day one. You should think, are my first 10 users extremely happy about this. But if you’re building a network business, then the way to make your users happy is usually rely on getting the other part of the market so then you have to build a separate model. So the model changes but you kind of have a real purpose of what you’re trying to achieve and try to achieve one and two things at a time.
Martin: What else that you learn?
Eren: I mean like so much. I think we should start with things which more like counterintuitive side I’ve learned. Because I think a lot of standard advice you hear everywhere.
For example, in terms of product success, especially if you are coming from an engineer background, you’re usually convinced as important are hard. Like things which are hard just seem to be important to you, because they’re hard and you’re an engineer and you think because you solved it you’ll be having success. That might be true for school but it’s really untrue for business. What is really valuable and important could be something simple to you, you just need to consistently do it. For example, for us, the live education component was a lot technically harder and cooler for them because it’s a sexy product. We even have raised money for what it’s worth, but it wasn’t really bringing much value for users. We’re on demand, kind of crude video courses actually being valuable, and I think people were ignoring the importance of this, because it wasn’t a new technological achievement.
It was easier to do. And it wasn’t a success for other people because they weren’t consistent in doing this on a lot of subjects. I still see this, people who want to get into education, they wanted to do something big and they always have cruel ideas about how to teach something. They just ignore certain things because they’re straight forward. I think the value you can provide is not like crazy, weird ideas. It’s consistently doing something good and iterating step by step.
Martin: Okay. Great. Do you want to share any further?
Eren: Not much, but if you’re more for Germany, I would say, Germany is one of my favorite markets. I think we have a unique position to observe the learning skill of doing learning courses in different countries. We see that a lot of people are real passionate about learning and their engagement are really high for a lot of German users. So if this kind of clear that people in Germany, they really care about bettering themselves and learning new things. And there’s a lot less age fragmentation which is also very cool to see. So we’re pretty happy with the demographic that we are getting from there, so we’ll definitely double count on Germany as a country.
Martin: Great! Eren, thank you very much and if you want to start your own company and learn how to code, maybe look at Udemy.
If you have been using the internet for a number of years, you must be able to recall a point when, …