Launch HN: Eduflow (YC S17) – Async Social Learning
Learning needs to be active to be effective. For example, the best way to learn to code is to write a lot of code. Online learning still struggles with this. Mostly you get a playlist of videos to consume and a quiz at the end, or you’re required to participate in live video calls. Our product lets you build active and social learning experiences that work async.
We went through YC in 2017 with Peergrade (https://www.peergrade.io), a tool for online peer reviews. We got to a decent size, but ran into challenges. When we landed a deal with a company or a university, we would be a small part of the experience. We’d only get used a few times a year, and only a portion of instructors would be using peer reviews. Peer reviews were seen as a feature of a learning management system, not a product in itself.
We noticed that we could generalize Peergrade to make a more flexible tool that would be able to match their pedagogical designs and not force instructors to model their courses after our tool. At some point we were generalizing so much that we shifted focus to building a whole learning management system instead. That’s Eduflow!
Our design is modular. Your course starts out empty, and you add the learning activities that you want. Activities can depend on each other—for example, a group-based peer review flow where you first want learners to select their group members (group formation activity), then submit some kind of work together (submission activity), then give feedback to other groups (peer review activity) and finally to reflect on that feedback (feedback reflection activity). This modular approach has been non-trivial to implement but it allows instructors to build exactly the learning experience they want, so the experience for the learner is extremely guided. And it allows us to extend Eduflow by adding new activities which can then work with all the existing ones.
Some of our customers (such as Google, Adobe and Accenture) are large companies that use us to run training and onboarding for sales orgs, or to do peer reviews at scale. Others are educational institutions (such as ETH Zürich and McGill University) that use Eduflow to make their large courses more interactive.
We sell a subscription to our service, and we offer a freemium plan. Since raising some money after YC Demo Day in 2017, Eduflow has become a profitable 12-person company.
Looking forward to hearing your feedback and answering any questions you may have. Thank you :).
24 comments
[ 3.4 ms ] story [ 57.7 ms ] threadPersonally I don't believe in peer reviews for my class. I tried it but the problem is that the students don't know what they're doing yet - so when I reviewed their peer review comments, they often were extremely short and poorly written, or even plain wrong advice. Students themselves said they didn't benefit from the peer reviews, both giving and receiving.
But even if I did believe in peer learning (I'm not claiming it can never work) - how exactly would your LMS help with that?
1. You want to make sure students have time to do it well. Instead of asking students to give feedback to large amounts of work, reduce the workload. Rather good feedback to one submission than poor feedback to three.
2. Use really good guiding feedback rubrics. These should help students write feedback that is actually useful. Instead of asking "Give feedback on the style" which can lead to "Great style", you want to ask things like "Find a place in the code where the style is good, and explain why it is good".
3. Ask learners to give feedback on the feedback they receive, and use that to incentivize learners to give good feedback. In my own courses using peer review, around 20-30% of the final grade was based on feedback quality.
There are of course many other parts to this, but those are some of the most important. Eduflow has a bunch of functionality to support this, including a feedback reflection activity etc.
I browsed your website and it looks really nice. Beautifully designed. I have been in the education business for close to nine years and still fascinating to see so many innovations. I have a few questions and maybe some suggestions.
1. Async model (Moocs) has been known for low completion rate while instructor-led cohort courses aren't scalable for large firms, how have you tried to solve this problem besides the community angle? ( I have thought about this and want to see how you have approached the issue)
2. How is the completion rate in passive learning scenarios like a corporate onboarding/training as opposed to a university class in Eduflow?
3. have you tried running classes, I saw that you have templates have you have used them yourself how has been your feedback?
Not sure if you already do but it would be great to have a course dashboard with graphs along with the tables. A simple kanban board for the student and a teacher for easier course management. Lastly, give API access to the firm that could connect their internal tools or other no-code tools to EduFlow and give more contextual programmes.
Disclaimer-Take everything I say with a giant pile of salt. This is just after some cursory look and I obviously don't know anything about Eduflow other than what I saw for a few minutes on your website.
1. I guess this is what we hope we can find a solution for with Eduflow. Pure self-led learning is not very motivating to most. Learning together with others is a great source of motivation, but this often ends up resulting in endless Zoom calls. So we try to balance this by allowing async social things. Examples of this could be peer reviews or something simpler like discussion activities. We also support setting things up like matching people in groups, and then having groups meet regularly, basically studying together.
2. I think it depends too much on the specific course. Generally the completion rate can be extremely high because there is an external pressure to complete. If you are going through corporate onboarding or taking a university course with a grade, then you are quite likely to complete all activities. In end, we are not in charge of the actual course designs, just the tool that enables instructors to deliver it. So we see everything from 1% to 100% completion rates.
3. We are currently running a course called Instructional Design Principles for Course Creation (https://www.eduflow.com/academy/instructional-design-princip...) on Eduflow. This is going well, and I think the places where we could improve is probably more about the actual content of the course (it was a bit too intensive for some learners). Before building Eduflow I taught data science with peer review (using Peergrade), and that was a pretty good success!
4. We are actually working on adding a bunch of dashboards to Eduflow right now. It is not just something instructors are asking for - it is also one of the most popular requests from larger customers. We offer an API already (https://docs.eduflow.com/) which gives access to data in Eduflow and manipulation. We also support Zapier which is pretty popular, especially for smaller customers.
Hope this is somewhat useful, otherwise you can just let me know and I will try to elaborate :).
You are using social accountability, be it online or in-person (office setting), to improve the completion rate. I am sure they have worked well for the clients you work with. Here are the few ways I have tried for corporate clients and have seen the completion rate increase in most cases
1. The candidate has to pay a high entry fee (for example USD 1000) and whatever percentage of the course they complete beyond the minimum 60%, they get that percentage of their fee refunded. So if someone finishes the entire course, the course is free for him and if someone does only half or less of the course they pay the full fee. The client then donates the money on the candidate's behalf to a partner NGO.
2. Another is a reward at every level. I don't have enough evidence to say these words, as I tried only once.
3. Course completion is tied to performance metrics but it has to be done with care because the goal is learning rather than a certificate.
4. Displaying % completed for the entire cohort (which can have side effects but is a good motivation device and can contribute towards social accountability.
I did these experiments by combining different tools but my case isn't representative of anything in the industry. So I must repeat take everything I say with a kilo of salt.
On the contrary, I am a bit skeptical personally about the long-term effects of adding gamification to everything we do. In the end, intrinsic motivation is the strongest force, and I worry (mostly on a personal level) that so much pressure on external motivating factors will make us immune eventually.
In Peergrade, we had a small amount of built-in gamification like this. We would show you how much feedback you had provided compared to the class average. This led to fairly significant improvements in both feedback amounts and quality.
Your biggest challenge will be figuring out exactly what market segment you are targeting. I see that you are doing Corporate and Educational institutions. Very different ballgame in terms of their needs, sales process etc. Trust me because I am doing this daily :)
All the best. Always good to see competition in this space because even though there are so many players, I think we are still in the early days on eLearning in general.
I take a lot of inspiration from Notion, Airtable, Trello, Typeform. They solve a generic problem and are almost mental models in product form. The modularity of Eduflow means that our customers are using us for very different things internally, which means that:
1. There are many gateways to using Eduflow. If you are looking for a solution to a specific problem, we can probably do it pretty well.
2. If we are used across departments and functions, then we become a lot more sticky, useful and valuable long term.
95%+ of our customers come inbound today, but you are totally right about the different way to approach these different types of clients.
As a seed investor who is a bit cautious in investing in EdTech, I am wondering if you get any requests from universities to plug into their established LMS (hooking into their login api for example)? And if they don't request that, do you get lots of issues related to user privacy and handling student data (going through compliance,...)?
Only asking because we have previously identified these as one of the two highest risks in selling LMS software to universities.
Also am curious to know what is your rough revenue share breakdown between academia and corporate? I a am guessing it is very skewed towards corporate, as we have previously identified another risk with EdTech is that it is very difficult to charge universities much (schools are even exponentially more difficult).
1. This is actually something we do a lot. Due to our history of being a peer review plugin, we have always been able to plug into an LMS through the LTI standard. When we built Eduflow we kept this option, so Eduflow can be plugged into another LMS as an LTI tool. Often this is how we start our work with a large corporate client - we will function as a plug-in to their existing LMS to offer them additional functionality. Over time our usage grows, and for some organizations we end up becoming the main LMS.
2. Actually we have more than half of our revenue coming from higher education so far. This is mostly due to our long history selling to universities - but universities are actually able to pay decent amounts, and when you land them as customers they are extremely loyal and very rarely churn. Schools on the other hand are very challenging - we have decided not to focus on the K-12 market at this point due to difficulties monetizing it.
1- What is your route to market when you sell to a new university, do you sell directly to the instructors, or sell to some kind of central function with the university ("learning technology department")? The issue here is that the centralised functions are much more risk averse to paying for new LMS once they invest in supporting one.
2- Also, do you notice a difference between geography/type of universities and how much they are willing to pay? For example from our previous investments, usually it is easier to monotise European business schools (where departments have more autonomy over budget), but selling to large top-tier public universities is much more difficult.
2. Somewhat. We have been able to work with all sorts of institutions, mostly in Europe and North America. We have all Danish universities as customers today, but that is also our home turf.
They built a new platform because internet bandwidth had increased to the point where it was competing with real time educational platforms.
Happy to pay, if the service is not expensive.