All Articles

Thoughts on Coursera's S-1

Coursera is an online learning platform that partners with universities to publish university-like courses which can be completed by anyone, from home. It recently added degrees which are real degrees with partner universities, such as a Masters in Software Engineering. These degrees cost between $1K-10K and the online courses vary from free to $99 or with a pay-for-a-certificate model.

Today Coursera’s S-1 was published, which is a financial document the company creates with the intention of listing its company publicly on a stock exchange, and this is called an IPO event. It lets the company’s shareholders sell their shares in the company and allows private and public investors to buy into the company.

Some of the interesting things from the S-1:

  1. Coursera grew from 46 million total registered users in 2019 to 77 million in 2020, boosted from a typical 20% YOY growth to 67% by covid.
  2. The S-1 doesn’t mention active users at all, suggesting the active users figure is disappointing, and with how short coursera’s courses are this wouldn’t surprise me. There’s also nothing on user retention. Both of these metrics are important indicators of Coursera’s performance.
  3. Last year the Degrees business of Coursera grew from $6 million to $12 million
  4. Coursera did $293 million in revenue in 2020, and the Cost of revenue was $138 million. This cost of revenue is mostly paying universities their cut for their content. Universities are getting slightly less than half the revenue generated for the non degree courses on Coursera.
  5. 86% of user acquisition is organic for Coursera, but they don’t list a channel breakdown unfortunately
  6. The CEO was compensated $14.9 million in 2020 in salary, stock and option grants. That’s more than 5% of Coursera’s revenue 🤯 This is likely because Coursera overperformed on their revenue targets (largely due to covid) and most of the compensation was performance based.
  7. 40% of the company is still owned by executives and founders
  8. They had a net loss of $67 million in 2020 and a similar loss each year before. However they spent $76 million on R&D in 2020 and their business fundamentals are strong. Their gross profit is at around 50%
  9. Most of Coursera’s revenue comes from Consumers at $193 million, next enterprises at $71 million and then degrees at $30 million
  10. Coursera recently became a certified B-Corp, and despite it’s .org domain is still a for profit company. Hackernews comments on the S-1 announcement frequently brought up a public sentiment of distaste for-profit education that Coursera has largely avoided.

revenue by segment

My thoughts

I love the idea of making education and learning more accessible to people and Coursera has been at the forefront of making this a reality. However I haven’t really enjoyed using Coursera’s product that much. It feels clunky, out of date and it lacks many of the engaging parts of other learning apps I use like Duolingo. Coursera has notoriously low completion rates like many other MOOC providers and back in 2014 during a hackathon I created a little Coursera extension that gamified your progress out of sheer frustration at not being engaged by Coursera. It literally is university copy and pasted online rather than university adapted for the 21st century.

Coursera’s main strength right now is cheap customer acquisition across their learning products. Customer brand awareness is high and this is useful for building credibility and authority over providing courses as well as convincing customers that a certificate from Coursera for completing a course has value in the job market.

Coursera takes a large cut (50%) of course revenue and the universities get the other half. This seems high for a platform that just hosts the content, but Coursera also acquires most of the users. Udemy, a platform for online learning with content created by anybody has an even higher percentage of revenue that they take.

Universities provide credibility to the content and reputation and certified masters/bachelor degrees, but as time goes on Coursera will strengthen its reputation and be in a position to create its own content without the need for the university. By aggregating content Coursera is collecting data in order to figure out the landscape of content and be able to create its own content in the future. Why have 30 computer science masters from different universities that are expensive when Coursera can provide a flexible masters with many learning paths for cheaper with the same or better credibility?

The role of universities, aggregators and Coursera in the future of online education

Universities with strong brands will also feel like they should get more than 50% of the revenue. They have stronger customer driving power and are going to other platforms or creating their own branded platform. Many of these elite universities (Harvard, MIT) are on edX instead of Coursera, and they also don’t offer equivalent bachelor or master degrees. They are protecting the brand of their in-person student degrees, which also pay significantly more tuition. This move to their own sites as an alternative to the marketplace is a mature trend online across industries - Amazon vs Shopify, etc. Organizations want to be in direct contact with their customers without having to go through the marketplace, more control over look and feel, and they want lower fees. Related: https://stratechery.com/2019/the-google-squeeze/

Part of the reason Coursera has such low customer acquisition costs is because people go to Coursera to search for content on how to learn something. However other platforms are increasingly competing for that right and it seems doubtful that Coursera is going to win the fight for where users start their search.

I expect Google to make moves in this area in the future, adding a domain specific search engine for learning, similar to the way they have done for flights and hotels. When they do, coursera may become more of a video hosting platform with quizzes for universities with a 50% revenue fee as they have less of an ability to drive customers to content. At that point Coursera will have had to made the jump to having certificates that carry weight in the career market or started innovating on content or I think it will struggle to be massively relevant.

Many people in tech think that the university system, particularly the fees in the US is fundamentally broken. It’s easy for software engineers to think that universities and degrees aren’t necessary when people in their profession often don’t require them. Universities are slow to change and Coursera is bringing “digital transformation” to some of the slowest to change organizations around. That business in the short to mid term will work great even if it feels like Coursera isn’t really making a big improvement in how people learn. Putting learning content online for free is great for the world and I hope Coursera finds a good way to balance is impact mission with it’s revenue-performance based compensation for its executives, as those can sometimes be in tension.

Coursera’s strengths and weaknesses

Strengths

Strong growth (67% YOY)

Strong gross margins (~50%) and fundamentals

Large network of content creating universities with large audiences and credibility

Strong brand awareness and trust. About a third of Hackernews responses were sharing a story of how a Coursera course impacted someone’s life and career.

Lots of consumers paying for courses. However how many of the 77 million registered users paid was not disclosed.

Weaknesses

Universities with the strongest brands are mostly not choosing Coursera. EdX, a smaller competitor to Coursera has MIT and Harvard in their course catalog. Universities can put their content in multiple places to increase distribution, and Stanford is both on Coursera and edX.

Coursera’s certificates don’t mean much in the workplace or for hiring managers right now, and Coursera monetizes around these. About a third of hackernews comments on the S-1 were commenting that the certificates had no little to no perceived value. Most of Coursera’s courses have tests that are graded automatically by software, so most of the tests are multiple choice which makes them not the best tests of aptitude as well as easy to cheat on or have software complete the course for you. Coursera needs to make courses and tests more rigorous in order to build credibility, but that will likely increase the costs to them per student if it involves humans marking results, and that means courses with credible certificates will have to remain paid.

I’m speculating here but I don’t think Coursera is winning the best enterprise customers. Coursera feels like they are catering to large slow moving companies that want “digital transformation” but is not winning top startups or companies known to be very innovative like Google. This means Coursera’s product is being developed for enterprise features and the focus is on satisfying these customers rather than driving end user outcomes which are very difficult to measure. While this is a great way to build revenue, in the long term the company may not be focused on solving the problem of learning as much as it is focused on convincing their customers that they are doing a good job with learning.

As a user of Coursera I find the product lacking. It’s not very engaging, it has no streaks, no gamification, few social features and feels like videos and multiple choice quizzes in a white word document. Sure the universities creating the content are probably not the best content creators, but would it hurt to add a bit more design to the product, add some more pictures the quizzes and help people set learning goals, like Duolingo does?

You can find the S-1 here: https://www.sec.gov/Archives/edgar/data/1651562/000119312521071525/d65490ds1.htm