Grinding Without Hope and Udemy’s Struggle With VCs
This is part 2 of a multi-part series in which I tell the story of Udemy. If you’re new, you might want to start with part 1.
---
It’s an odd feeling being an entrepreneur. At times, it feels like you have no hope. You stubbornly push forward despite hundreds of data points telling you to turn back. You pitch and pitch and pitch your heart out, while wondering if maybe the suits are right. Maybe your idea does suck. After all, 95% of startups at this stage fail.
For nearly a year, we desperately tried to raise our seed round. We had no savings and no family money. Eren and Oktay were on H1B visas and were technically not allowed to work on Udemy, lest they be deported.
As I got to know Eren and Oktay, I learned two things:
One, these guys are brilliant. Like, truly next-level, one-in-a-million type of brilliance.
Two, as Adeo liked to say, “Nobody can understand what the fuck they are saying.”
Like most countries, Turkey has a completely insane education system. The schools must hire “local” teachers, because the government is expected to keep jobs inside the country. So, subjects like English suffer, since you have very few native speakers who want to be teachers. So, your English teachers suck and you hold your country back, all in the name of giving local people jobs. The USA does the same thing, which is why my Spanish is no bueno.
You can see why we wanted to democratize access to education; all three of us had fallen prey in some sense to the modern education system.
So, I was going to be the glorified translator. The junior business guy with a $100 suit and a big mouth.
I took it like a man and worked for free, with no expectation of compensation. I joined a project at Accenture in Roanoke, Virginia, so I could use the travel stipend to make trips back to San Francisco each week.
Toiling with No Hope
My workload at Accenture jumped to 70 hours per week. I’d arrive at the soulless office park in Roanoke on Monday and stay there until 2am, perfecting Excel models for clients who would never use them.
On Thursday night, I’d fly from Roanoke to San Francisco, where an air mattress awaited me in Eren and Oktay’s spare bedroom. On Friday, I’d split time between Accenture and Udemy. We worked as a team on Saturday. On Sunday, I’d take half a day to see my girlfriend in Berkeley. Sunday night, I’d take the 11pm redeye to Chicago, get 4.5 hours of fitful sleep, transfer flights, change into formal clothes in the airport bathroom and do it all over again.
My job was to get users for the site and to pitch investors. I was doing a decent enough job, but it was slow going. Eren would stay up until 4am every night and then go to work again at 9am.
After four months of grinding, we finally had “the conversation.” Who would have what role? How would we split the equity?
Eren was obviously CEO, but what would I be? These two coveted software engineers were working at a venture-backed startup and I had no real experience. On top of that, they had worked on Udemy for 2 years prior to my arrival.
Eren was quite generous and offered me a co-founder role. We agreed to divide up the equity almost equally: 4x for me, 5x each for Eren and Oktay. I got the title “President,” which didn’t mean anything, but it felt about right.
When Paul Graham called Udemy a “tarpit”
I spent every spare minute emailing and setting up phone calls with investors and users.
I had tried everything, but nobody would bite. Eren and Oktay were still building the product. In retrospect, I think they had “launch anxiety” and were delaying the launch date.
The investors were cordial, but no one wanted to invest in education. Our exchange with Y Combinator is probably the best example of this:
Paul Graham:
You guys seem intrepid, but this idea is a tarpit. We get several vague proposals for tutoring marketplaces every cycle.
Try answering this hypothetical question: if someone wanted to build something that would make lots of money starting from the beginning, what would you suggest they do? Surely a tutoring marketplace would not be your first suggestion.
What do you need so much you'd pay for it?
Since 2/3 of you have experience in dating, why don't you work on something related to that?
Gagan:
Thanks for the opportunity to discuss this with you further. We're not a tutoring marketplace, and we must have done a poor job of messaging during our application. I apologize for that. We completely agree that tutoring marketplaces won't make much money - and have little potential to become big. That's because they focus on one-on-one learning and on replacing an interaction that requires face-to-face meetings. We're completely different. We are focusing on content that is one-to-many and are trying to become a place where users have thousands of courses they can watch and learn from. We plan to aggregate the educational content on the internet and provide users with a one-stop shop to learn.
Furthermore, people pay for online education regularly; its a $25B industry and over $1B is spent in the US alone on online courses that are not tied to a professional or university program. Lynda.com is a great example of a company that targets the same user base.
Ultimately, we offer tools to enable anyone to create a Lynda.com. Course creators don't have to be individuals and we don't expect them to interact one-on-one with users. We expect people to put up videos and then others to come watch and subscribe to those courses. We also expect some online instructors to hold live courses with tens of students in attendance, similar to EduFire's entrepreneurship series. There are thousands of in-person seminars held around the country on a wide range of topics, from OpenGL for the iPhone to Spirituality and Vedic medicine. Many of these can be replaced by online courses on our site.
That said, we're not opposed to changing the business plan. The technology we've got is amazing and it wows everyone who sees it (including investors and CEO's of startups at TechCrunch50 - we got tons of follow-up). We've got a live learning platform that is extremely easy-to-use, flexible and scalable. We'd love the opportunity to show you it before you write us off (because it could be re-purposed for something else if necessary).
Finally, re: 2 of us are experienced in dating. They've also got experience in online video, viral marketing and user-generated content. A lot of the same things that made them successful with SpeedDate.com can transfer to Udemy. If you still aren't convinced, let us give you a demo. The site is literally weeks away from launch, so you'll at least see a functioning product during our meeting and can make a determination there if you think we could morph into something that isn't a tarpit.
Trevor Blackwell:
Can you guess what fields will be popular for your kind of teaching? Using office software like Lynda? K-8? Quantum mechanics? Astrology?
Gagan:
Thanks for the followup. We've got a good idea based on what has been successful on the internet and via video/DVD-based learning to-date.
We expect this to not appeal much to K-12 (that market is fairly saturated). Instead, we see this site as being a place to learn the stuff people don't want to go to school for but still want to learn. Here are some categories that we think will be successful:
Casual learning (see The Teaching Company: teach12.com)
These are general knowledge courses on anything from "String Quartets of Beethoven" to "Art of Critical Decision Making" to "American Civil War"
For people who want to learn the interesting material of college classes without attending a university or paying the tuition fees
Students pay $100 for 24-lecture DVD sets
Company was acquired by a private equity firm recently
IT Education (see Lynda and the hundreds of indie sites such as javavideos.net)
Self-explanatory. For people who are interested in learning a trade and either earning a living off of it or casually coding / using photoshop in their free time.
This is an enormous industry dominated by various technical schools and sites like lynda.com.
Test Preparation
Sites such as brighstorm.com and tutorvista.com are seeing solid traction in this space. We think they cost too much and don't provide enough teacher options, and that's why we can compete. Also, tutor sites don't allow for recorded do-it-yourself content and brightstorm doesn't allow for one-on-one question and answer.
Ultimately, there are dozens more - yoga, cooking, how-to videos, photography. But we assure you we won't try to conquer them all in our early days. We will focus on a few (2-3) specific verticals and build density in those to gain traction, evolve our product, and prove the concept. Then, we will slowly expand into other verticals.
---
What’s fascinating about this is:
Paul Graham actually dug this up for me. I admire him deeply, and it shows how much of a class act he is that he would give us permission to share this.
YC quickly changed their tune, and later even joined forces with Geoff Ralston’s Imagine K12, an education-focused incubator. Things change, and timing matters.
---
We applied to YC three times but never got an interview. Simultaneously, I was pitching angel investors, and by February 2010, I was finally getting some interest. Eren even had a Turkish billionaire thinking of investing!
Juggling Accenture and Udemy, while simultaneously writing for TechCrunch, was getting to be too much.
I had saved about $5,000 by that point, and my co-founders still had full-time jobs. I remember going to them and saying,
“Guys, I think we can raise money. But if this doesn’t work out, I’ve got almost no savings, so will you cover me if that happens?”
We finally incorporated the business. I quit Accenture, flew back to San Francisco, and immediately started hitting the pavement.
Within 3 weeks, we had met everyone on our lead list. They all said no.
We had one ace up our sleeve: the billionaire from Turkey. He was going to invest $500K.
Unfortunately, he dropped off. Eren emailed him several times, but we never heard from him again. We were stuck.
It was a decisive failure. It stung. I was frustrated.
I still thought Udemy was a good idea. The arguments made by investors didn’t make sense to me; they seemed to not “get it.” I took the feedback and constantly improved the pitch, but it wasn’t enough. Pretty soon, we’d exhausted all of our contacts. We knew we had to do something different before we could move forward.
Why the pitch failed
In retrospect, there were a few major problems with our pitch:
First, we hadn’t launched yet. It was a major red flag, but nobody told us directly. Investors don’t want to gain a reputation for offending entrepreneurs, so they tend to be indirect.
Second, I was pitching, but Eren was CEO. We don’t know how much this affected things, but it was certainly a bit atypical.
Finally, we’d picked an industry nobody liked. At the time, online education seemed like a “tarpit” to most investors.
So we decided we had to get traction. Traction. Traction. That was the goal.
But how? We were a two-sided marketplace—traction would mean getting both instructors and students. How do you get instructors to teach on Udemy without any students? How do you get students to learn without any instructors? We couldn’t have teachers without students any more than we could have students without teachers.
Next post: Solving the chicken-and-egg problem, and raising a seed round.
---
Thanks again for reading this. I write sporadically, but when I do publish something, I try to make it interesting. Subscribe here so you’ll never miss a post.