Case study: Eduflow’s pivots from Peergrade to Multiverse
How Eduflow began life as a peer grading solution, pivoted into the bigger LMS market and was then acquired by Multiverse.
👋 This week a great case study on EdTech pivots. We’re also hosting another online meetup on Tues 15 October, 4pm BST. Sign up here.
“The next time I build a startup, I'm going to make sure - absolutely sure - that if I can build it, there will be a customer,” David Kofoed Wind is reflecting on his journey from Peergrade to Eduflow to joining Multiverse.
“When I ask people about the problem, there needs to be a reaction like: ‘I would love to buy that thing! If you can solve this problem, I'm there!’ I need that kind of attitude before I start. Not like, ‘yeah, peer review… That'd be cool.’ Students maybe learning a bit more does not sell, unfortunately…” he says ruefully.
Eduflow, the evolution from David’s original product Peergrade, is one of the best EdTech products that most people have never heard of. A powerful Learning Management System that is simple and delightful to use for both educators and students.
Its strengths were spotted by Multiverse, who were looking to bolster their existing technology platform to support the company’s incredibly rapid growth. And in May 2023, Multiverse acquired Eduflow and its 10-strong team, who are all still there, making it an unusually successful integration.
This was the most recent step on an eight year journey to find product-market fit. David has taken the time to reflect on that journey with me today. Along the way we’ll explore finding your first customers, how to think about markets in EdTech and the messy reality of pivots.
“We did all the startup things. And it was fun. But I've learned so much. If I had to do this all over again, I would have done so many things differently.”
Beginnings
The story begins with David solving his own problem in 2015. He was doing a PhD in machine learning and teaching a course in computer science at the Technical University of Denmark. “The course got a lot larger than it was supposed to be, so I had a problem of grading all these students,” remembers David. “I couldn't find a peer grading solution out there, so I decided to build my own.”
David had previously worked on an algorithm to spot fraudulent voting as a research student, “we used the Eurovision Song Contest as the data set, where there's a lot of fraudulent voting!” He realised that he could apply it to the problem of students marking each other’s assignments.
“I started writing Peergrade on my own on the weekends to solve my own problem. Then it turned into a startup. But that was not really the plan. I didn't sit down to start a startup. But I guess if you if your startup muscle works, as soon as you start working on something that smells a little bit like a startup, you're like, maybe there's a business here… Maybe I should talk to people about it…”
He spoke to his supervisor. “He was a great guy - the kind of supervisor who ran 10 startups on his own. He was very entrepreneurial. So I tell him about it, and he says, ‘this is great! You should sell this to the department!’ I thought this sounded crazy, I already worked there!”
A meeting was arranged with the Head of Department, Rasmus, and David found himself pitching. “I explained this peer review thing. And Rasmus is, like, stone cold. He just looked at me,” David remembers. “So I said ‘Well, I guess, if I use this for my course, I don't need four teaching assistants. I guess I could do it with one, maybe?’ And now he says, ‘well, that sounds good. What does it cost?’ I say ‘$1,000 per course, for a semester, I guess.’ And he's like, ‘Okay, we'll take two licences, one for you and one for your supervisor.’”
Somewhat in shock, David returned to his office to figure out how to start a company so he could send an invoice. He also sent a text to his old high school friend, Malthe Jørgensen.
“He was one of my best friends, who was a better programmer than me. And I said, Malthe, I got myself into trouble here… I need you to help me write this piece of software. The only thing I can give you is equity in our company that we’ll create and you can be my one teaching assistant.”
Malthe agreed. A little later they were joined by their third cofounder, the designer Simon Lind who they found via LinkedIn. Peergrade was up and running.
Peergrade
He began looking for more customers. Fortunately, there were some easy targets close by. “I thought I'll just walk up and down the hall,” says David. “I know these people. I'll ask them, Do you have a lot of students? Do you want to do peer grading? I got a few customers right there.”
“Then, I went around the whole department. Then I wrote an email to the department heads of all the other departments, met with each of them and basically tried to sell a licence per department. Management and Economics was really easy. They really liked this entrepreneurial idea. And they had some very large courses.”
He carried on and did the same with other universities in Denmark. He realised that he had the perfect opportunity for a cold outreach: “Hey, I'm David. I'm a PhD student at DTU. I'm doing this research project on peer review. Can I kind of come and talk to you about it?”
He got an almost 100% response rate and started visiting people. “I started by telling them about peer review and why it's effective. Some of them knew about the concept already and I didn’t need to convince them. And then I say, ‘by the way, I’ve built some software that does exactly this. Pretty lucky, right?’ And that was my pitch for many years, because I was doing my PhD on the side for the first three years or so.”
They signed up the University of Copenhagen as a client and began going university-by-university, department-by-department. They discovered that, much like with Rasmus, their product was an attractive way to cut costs and save money on grading in a way that felt like it also could improve student learning. They realised that they could solve the same problem for high schools.
“I wrote to my old high school teacher. Malthe and I went to high school together,” he says. “I said we're doing this startup thing. Do you think peer grading makes sense in high schools?” They got a very positive response, which David found initially surprising. “I thought the whole problem was the number of students and high schools had small classes. And she replied, ‘But David, I teach many classes…’” And so they began to get traction with schools too.
They attracted some Angel investment and got accepted onto the Y Combinator programme in San Francisco. They were doing a real startup!
“We're gonna take it to the moon. Let's make it huge!” he reminisces. “I was touring around, evangelising peer review in all the high schools and universities, and the other two were just coding, coding, coding and designing. We kept getting more and more clever about peer review. We found more and more interesting insights and features, and we built fast. We quickly got to the best peer review product in the market because there were no real competitors.”
This is the bit he describes as the “boring” middle bit where they just worked hard. They got an office, hired more people, and watched Silicon Valley together. They eventually signed up all of the universities in Denmark and half the high schools. High market penetration, but in a small country.
The first messy pivot
“There was no real market for formative assessment products,” says David, just exam software. “And we didn't want to be an exam product.” Almost by osmosis, they started to get pulled further into the rest of teaching and the lines began to get blurry.
He describes the kinds of requests they got from teachers: “After a peer review, it would be nice if I could do a teacher review. Or what if they self review before they give a peer review, and then they can compare the two. And what if they do peer reviewing groups and submissions in different formats… and suddenly we're thinking it's starting to look a little bit like a learning management system. So for a long time we were trying to figure out what we were.”
While at YC, they built a live version that could be done during a lesson. But immediately got feedback that it would be good if they could offer a combination of sync and async.
“There's always a twist, there's always a teacher with a plan that's slightly different than the one you imagined,” says David. “This happened over and over again.”
They spotted another client, a business school, setting up an assignment with no reviewers. When they investigated it turned out they were just using it as an assignment submission. They asked why they weren’t using Blackboard their VLE to do this. “The students are in Peergrade. We like Peergrade and we just want it all in one place,” came the reply.
“I thought, okay, this is a very clear message,” says David. He notes that this is one of the big learnings: you need to keep talking to customers and really understand how they want to use your product and what pain points you are solving for them. They may not be the ones you think they are.
Meanwhile, their partners that were just using the product for peer grading weren’t it much. “Even our most loyal fans only taught a course or two per semester and used it twice in the course.”
The idea of building an Learning Management System began to grow in their minds.
By this point, it was summer 2019. They had been running Peergrade for four years. Whilst not yet profitable, they had been breaking even for most of that time.
“Peergrade was working and growing linearly, but not exploding,” says David. “And there was no clear path to explode. We'd used all the growth hacks at this point we could come up with. And so we thought we could try this alternative product strategy and keep running Peergrade in the meantime. So that's what we decided to do.”
Building an LMS
The team was aware that there were significant basic expectations you needed to meet to enter the Learning Management System market. Unlike many pivots, where a team doubles down on one feature, this was the reverse: they would need to offer a lot more.
“We couldn’t have started there,” says David. “But we had a million dollars in the bank. We had a 10 person team that was very effective. We already had authentication, some dashboards. We had a bunch of security stuff that was already working. We could take all of our integrations with us, and so we had a lot of the table stakes. We didn't have to invent all of those things again. So it was a great time for us to see if we could do it at no risk.”
There was a lot of competition. But they felt they could easily do better than Moodle, Canvas and the incumbents. Initially, they got a positive reaction to the idea. “But then they would come with an Excel sheet with 300+ requirements. We would get these insane procurement sheets. We just looked at each other and said, ‘holy sh*t.’”
With Peergrade, they had been able to sell to individual professors and departments and saw usage from day one. “But buying a LMS is not a professor's decision. A peer review tool was like a plug-in, so we could quickly get to something useful in peer review. With an LMS it took a very, very long time.”
They began by trying to migrate their existing customers. They managed to do this for around half, but eventually gave up. They saw some excitement from those that weren’t yet using a LMS but relying on Google Docs. They got some traction with ‘makers’ who were already using tools like Teachable but looking for more feedback and assessment tools.
And then they started to get into corporate L&D, including Google. He says that L&D is so much easier than working with universities who take years to make decisions. He reckons the sweet spot is scale ups who need an LMS but don’t demand lots of integration.
“It was still pretty tough,” he says. “It took four years to really break into enterprise. But that was when it got potentially quite exciting, because the deals were starting to be more of a more interesting size.”
But after four years they realised that they were back in a similar situation to four years earlier. Solid linear growth but nothing that looked explosive. And once again, out of ideas for how to change the pace of growth.
Lack of focus
David puts this down to the flip side of their enormous strength, offering a simple but highly flexible product. They appealed to lots of people, but lacked focus.
“It was clear to us, our investors and everyone we asked for advice that we had no focus. We had fourth grade students in schools, we had Google, we had ex-Teachable customers... Everything under the sun,” says David.
“We flip-flopped for years about whether this was our biggest problem or our biggest advantage. But, even on the days where we were clear that this was our biggest challenge, we didn’t know what to do about it. Because if we picked one of those markets, we’d kill 90% of the business.”
He says that they were leading quite a stable and good life, but that it got a bit repetitive. “We were not pursuing novel ideas, mostly looking at that big 300 item Excel sheet, and building the next one on the list. You start to think, do I want to be doing this when I’m 50 or do I want to do something else one day?”
It was then that they started to get their first acquisition offers.
“We had gotten no acquisition offers for the first six years or so. That was not really a thing that happened to us. But we said, if an acquisition offer comes up, we're going to take the conversation,” he remembers. “And then we had the first two conversations in 2022.”
The first was with a German HR company looking for an onboarding product but there was no chemistry with the founders. Then a Canadian company who had a portfolio of businesses but no course platform. But they didn’t feel it was a good enough offer when they were still growing, albeit slowly.
Then Multiverse got in touch.
Multiverse
Multiverse was founded in 2016 by Euan Blair to provide an alternative to university education for early career starters. The business quickly grew fuelled by the UK apprenticeship levy and a growing demand for skills in data and tech.
“Multiverse was very successful at selling apprenticeships and growing their business,” says David, “But a lot of the growth came so rapidly that its tech stack wasn’t scaling at the same pace. They were using some form of Moodle as their learning platform. But it was becoming clear that to really scale Multiverse into a big tech operation, they needed to own more of their own software so they could change it faster.”
They had gone looking for an LMS they could acquire and begun speaking to a range of companies. “And then there was a good fit with us because we were a LMS that could do everything. So no matter what they came up with, we were like, ‘yeah, you can do that.’”
He elaborates, “they couldn't go out and buy Canvas, right? Canvas is an average product with a mega business. But we were a very good product with a fairly mediocre business. So it was a really good fit and a good acquisition for everyone.”
The team of ten joined Multiverse. One who lived in Indonesia, couldn’t due to employment law but the rest of them are all still there, nearly 18-months later.
“We became our own squad, and eventually our own ‘pillar’, which is a structure we have at Multiverse,” says David, now Senior Product Director. “Malthe, Simon and I still work together side by side, and all the original Eduflow engineers are still working with us, along with some new people. Today, there’s about 20-25 people in our group. Eight of us are the original team. A couple of the Eduflow team went to other departments, like learning.”
And what about the product? “The core product actually hasn't really changed. It's all the stuff where it needs to integrate into Multiverse. All the challenges are on the surface.”
He says that whilst they would love to spend time improving the learning experience, there are much bigger problems to solve to support the scale of Multiverse. Coaching capacity, how to make content faster and a lot of other things have a bigger impact on learner and client satisfaction than continuing to build more learning features.
Learning
So what has he learned from this incredible journey? What are the things he would do differently?
“People should just start,” he says. “I talked to a lot of potential startup founders who are scared of doing the wrong thing and so they don't get started. It's much better to do the wrong thing than nothing at all. There are things that I will do differently next time. Now that I am David 2.0. But could I actually have done that the first time around? I'm not sure. Maybe I had to learn from the experience…”
So given that, how would he approach his next startup?
“I would start with the market demand,” he says without hesitation. “One of the best ways to start a company is to scratch your own itch, like I did with Peergrade, because you know that at least one person really needs the product. But you don’t know if there really is a market of people who want to buy the thing. That wasn’t something that we really thought enough about.”
He reflects further. “With Eduflow, we went into a market that was much more competitive. That was both good and bad. There was lots of competition. But it also meant that the market was very validated and there was plenty of opportunity to build a better product.”
He sums up: “I would much rather have a large amount of technical risk or product risk than market risk. I'm not willing to take a market risk again.”
He says that the important thing is that it is a real problem that you can potentially tackle in a new way. “I often think that if Simon, Malthe and I sat down and looked at a bunch of products, we would probably find a category where we could do something better if we put our minds to it. Let’s take a market that works, and let's just build a better product than the others.”
I ask him about where the opportunities are in EdTech. “It's such a hard market. But Education is going to be disrupted much more than most other things by AI,” he says. “I just know that if I had GPT in my school days, I would have done all my homework that way.”
He also adds that AI will change the way that people will approach startups. He wouldn’t hire big teams and lots of interns again. “I think companies can be smaller. You can probably use robots for like, a third of your job. So you don't need a marketing intern to post on Twitter or sit and write emails. I think companies will just shrink in size and move faster. And it will lead to there being more companies.”
David is currently very happy at Multiverse. He says he has smart colleagues and he’s learning a lot.
“It’s a different type of job that I have right now. It's fun to be the CEO, because it's your thing and you can do whatever you want,” he says. “But my day job was terribly boring. I would be doing all the things that nobody else did. Running payroll, writing sales emails, taking the long train ride to give a talk to high school teachers, and doing support chat… all of that stuff that we couldn't really hire anybody for. And now at Multiverse, I spend most of my time on interesting product strategy. My only job is to think about what's the best thing to build in which order. Which, in a way, was the only thing I liked about my old day-to-day.”
Summary
We step back to reflect on the conversation. What are the important takeaways?
Just start. Then learn. The reality is likely to be different from what you imagined.
Make sure that you are solving a real problem. If it’s a new market, make sure someone is going to buy your product.
Product risk is easier than market risk. Solving an existing problem in a new way is likely to be easier than convincing people to solve a new problem.
Listen to your customers. Understand how they are trying to stretch your product and why.
Be flexible but have focus. Build so that you can accommodate new opportunities but retain focus - it’s hard to be something for everyone.
“It’s essential to build something that people want to buy. I talked to so many founders, especially in EdTech, who put blinders on. They don't want to look out, they think it’s a great product. But often it feels like nobody's gonna buy it,” he concludes.
“It's heartbreaking as a founder. I wish someone had said ‘David, weird allocation rules for group peer review is not going to change the appetite in the market. People just don't care that much about peer review tools.’ Focusing on the market is definitely the most important learning.”
This case study appears in my programme on Finding Product-Market Fit in EdTech. October’s cohort is now closed. Sign up for an early bird discount to join January’s cohort.