Discover more from Big Bets
FutureLearn with Matt Walton (Big Bets 4)
This is a newsletter about big bets, explored through conversations with the product leaders who worked on them. For this one, I caught up with Matt Walton, the former CPO of FutureLearn, an edtech startup that began life as a big bet by the Open University.
Jumping on the bandwagon
‘2012 was the Year of the MOOC, according to the New York Times,’ Matt tells me. I’d asked him how the Open University came up with the idea for FutureLearn, and he explains that they were jumping on the latest education bandwagon. Coursera, EdX and Udacity had all just launched, generating a ton of buzz, and US universities had started to sign up and create courses on these platforms.
In some ways it was a typical story of internet disruption coming to any industry: the Open University had been doing distance learning for 40 years, but suddenly new players had sprung up to capture this demand online. They faced the classic innovator’s dilemma. So they felt a strategic need to do something in response, building on their existing mission to transform access to higher education. And rather than trying to win as a single provider of online education, they took a bet on building a platform, where other universities could create content.
‘It was a very open brief at the beginning,’ says Matt. ‘To be honest, there was more of an idea that we should do something in this space, rather than a clear vision of what the thing should be.’ And hanging over this endeavour were various failed attempts to create a shared platform for UK universities, in a sector that has been historically slow to embrace change. There was a huge degree of scepticism.
Setting up the bet
One of the hardest questions for any bet is how to create the right conditions for the team working on it. Unusually, the OU’s Vice Chancellor Martin Bean actually had a tech background, having previously worked for Microsoft. He decided to set it up as an independent company, at arms length from the OU, so that it could operate like a startup. And he signed up 12 other leading UK universities to take this bet with him.
Martin hired a CEO to work on it, Simon Nelson, also an unusual choice with a background leading online at the BBC, who in turn hired a leadership team, including Matt. To begin with, there were differences of opinion about how the product should be built. ‘Promise me you won’t write a line of code,’ said Martin to Matt in their first meeting. His assumption was that they’d be able to use their existing IT team to customise Moodle, their current online learning platform.
Matt had different ideas. ‘If you want to do something disruptive, you don’t want to start with an existing team, delivering on technology that’s already 10 years old,’ he warned. This argument won out, so they got the green light to hire some new developers and were given a small office in the basement of an existing OU building in Hawley Crescent, Camden. Matt remembers saying that all they needed to be successful was smart people in a room with laptops, whiteboards and squeaky pens, just going for it. ‘In hindsight, I wish I’d specified natural light,’ he laughs.
Starting with a prototype
These first conversations took place in January. Their next move was to create a partner forum with their growing partnership of now 16 universities. Matt describes the first meeting as ‘rambunctious’, with everyone voicing a different opinion on how the product should work.
In early March, the small FutureLearn team of four, built a prototype in three weeks, to bring their vision to life and help ground a lot of those first conversations that were abstract and theoretical. ‘Having something to show is so important for getting people to talk about the same thing, and getting them excited,’ says Matt. The prototype also demonstrated their main hypothesis: that the key to successful online learning was making it social.
The early pioneers of MOOCs believed they could make online learning effective by ‘bringing back the deadline’. Open educational resources that enable you to learn at any time mean that most people never get around to it. A deadline can turn these resources into a course, an event. The FutureLearn team combined this insight with one of the main lessons of the web itself: making things social makes them better. ‘I was interested in this idea that you created the conditions for a really engaging social experience,’ observed Matt. This approach was underpinned by the research of their academic advisor, Mike Sharples, which showed that education is more effective in a social context.
The prototype was simple: a to-do list, an activity feed, and a progress page, along with activities like videos, articles and quizzes. It had a social element baked into it, with learners able to engage with each other on every page. Technically it was a mess: ‘we hacked it together in Ruby on Rails,’ says Matt. ‘The backend was horrible: actually one lesson from the prototype was how much attention we needed to pay to the Course Creator tool.’ But it brought the project to life for the developers who’d be working on it, for the executives at the OU and, crucially, the university partners.
Their finely-balanced strategy for managing the stakeholders of this bet was to a) remain at arms length, but b) regularly build trust. Every two weeks, he took the train out to Milton Keynes (where the OU is headquartered) to do a sprint review for the decision makers and advisers. ‘It was important to show the progress we’d made, and also help them understand what it took to build this. If we’d gone away for 6 months and come back with what we’d made, they might have said ‘is that it?’ But thanks to the frequent engagement, they saw the complexity and felt part of the journey.’
The same was true for the partners. Only 6 weeks after the first partner forum, Matt and team were able to come back to the group with a working prototype. They had to keep them on side, because FutureLearn was asking them to create courses for a product that didn’t yet exist.
Beta and live
After the prototype phase, the team decided that they had to be running a course in some way in 3 months’ time. 3 months is Matt’s magic number. In his own words, ‘it’s long enough to do something of value, but not so long that you can go far wrong’.
This ended up being a 2-week course that they ran for 150 friends and family, which they bootstrapped by taking a course that the OU had already run online on their old platform, called Moons of the Solar System. 3 weeks after that they ran another course for 1500 OU alumni. This course was created by one of their most engaged university partners, on the subject of building a brand. This course went on to become the first that FutureLearn ever ran with the public.
After meeting that first milestone, they set a second one, another 3 months: they wanted to launch a beta in September. By this point the team was 12. Big enough to get things done quickly, but small enough that communication was easy, which is important in the MVP phase when you don’t yet know what you’re building.
They had learned a lot about what worked and what didn’t from these first two tests - learners really wanted a ‘like’ button, for example - and rapidly made changes before launching their first course to the public in October. From that point, they ran one course a week in the run-up to Christmas, still in beta.
Did anyone take these courses? ‘Oh yes,’ says Matt. ‘We were oversubscribed and had to cap them at 20,000 learners each, because we knew we weren’t offering a great experience yet, and we also didn’t know how much traffic our site could support.’ This seems pretty impressive for a beta with no marketing budget. Instead, they attracted the first users by riding the wave of interest in MOOCs from the press, which saw Matt being asked to demo the platform live on Newsnight, for example. They also traded on the established brands of their university partners.
The results of the beta were excellent in terms of overall engagement, and they seemed to validate the hypothesis around social learning. The FutureLearn team scrutinised all their social metrics, such as how many comments each user made, and discovered that 40% of learners had engaged socially. This was huge compared to the same metric from Coursera and the other MOOC platforms, which seemed to hover around 10%. Crucially, they also found that people who were social were 6 times more likely to successfully complete the course. Matt points out the challenge of establishing causation for this correlation, but says it gave the team enough positive signal to keep pursuing their product vision.
In January, they introduced a wider portfolio of courses, and started running them uncapped. They also began to repeat courses they’d run already. They even moved out of the basement, and into a well-lit office in the British Library. By this point their team was 35 people.
The product continued to get good take-up, and score well on learning and social metrics. They started to think about monetisation, and settled on selling certificates to people who’d completed a course. Later, they bundled this with access to assessments and ongoing access. A year after that they launched their first full degree on the platform.
I ask when it stopped feeling like a bet. ‘When we moved back into Hawley Crescent in early 2015,’ he says - referring to the office whose basement they’d started out in. ‘But this time we took over the whole building’. They now had 120 employees.
This moment also marked a turning point in becoming much more commercially focused so that they could show a return to their university partners. ‘It had been easy to secure the original funding for courses, rustled up as a strategic one-off,’ Matt says, ‘but now we had to really help people understand the ongoing benefit of doing this.’
Following three rounds of investment from the Open University, in 2019 they took external investment for the first time, from SEEK. Australia’s top employment marketplace invested £50m for a 50% share of the business, valuing the company at £100m. By this point FutureLearn had over 14m registered learners, 160 course creating partners and 200 staff. The bet had well and truly flown the nest.
Especially in traditional institutions, the more distance a bet has from the core org, the more likely it is to succeed
Keep validating your product with increasingly large groups of users. 3 months is a great length of time between these milestones.
Having something to show regularly (a prototype, a sprint review) is the best way to keep stakeholders onside
Forwarded this email? Subscribe here. Enjoyed this email? Forward it to a friend. And hit reply if you have any feedback or bets you’d like me to cover.