EP47 · Economy · first published 2020-11-18
Learning and EdTech as an investment | Marko Kyyrönen | Negotiator 47
Sparkmind.vc managing director Marko Kyyrönen describes what it takes to raise a €60 million specialist fund in Finland and invest it in education technology. The episode covers Kahoot's path from a $350 million valuation to $2.5 billion, Fuzu's model in East African labour markets, Aino Saarinen's research on how long the effects of early childhood education actually last, the principle that education comes before tech, and the shift in credentialing from university brands to demonstrated evidence. Kyyrönen closes on what separates an interesting pitch from the other 300 — and why a market of a few hundred million is not enough.
Learning and EdTech as an investment | Marko Kyyrönen
Summary: In episode 47 of the Negotiator channel, Sami Miettinen interviews Marko Kyyrönen, managing director of Sparkmind.vc, which runs what is in practice the only Nordic venture fund specialising in education technology. The episode starts from the daily reality of raising a fund and moves to the structure of the sector: who dominates it, why Kahoot became the Nordics’ first edtech unicorn, what research on early childhood education actually says, and how the verification of competence is shifting from university brands to demonstrated evidence. Running through it is Kyyrönen’s principle: education is before tech.
Raising a fund is legwork
The interview had been postponed twice because the fund’s second close was under way. Kyyrönen describes the situation plainly: Finland is a small capital market, and assembling a fund takes an enormous amount of legwork — these days over Zoom and Teams. Comparable specialist funds abroad have larger anchor investors and a smaller investor club; Sparkmind’s club comprises around fifty investors as it targets a size of 60 million. The first close was made the previous December at 40 million, and actual investing began in January.
Miettinen raises an incentive structure many people misunderstand: the management fee on a 60 million fund is not large, and investors will not accept excessive draw along the way. Kyyrönen confirms it is a skin in the game arrangement in which the team has invested vast amounts of time and money, and the company also puts its own money into the fund — around two per cent of total commitments. The performance fee materialises only in five to ten years, closer to ten — or not at all, if things do not work out.
The first four investments
In the first year four investments have been made in Finnish education and education technology growth companies: Fuzu, Kide Science, SchoolDay and Playvation, also known as Moomin Language School.
As an example Kyyrönen opens up Fuzu, which runs two business lines simultaneously in the East African market. On the B2B side it offers African companies better-screened labour; on the B2C side it offers lifelong learners and vocational students high-quality content cost-effectively, typically over a mobile connection — and connects the learning to a local dream career. The team includes leaders with Nokia backgrounds in African and emerging market business, and the scaling potential in the African market is enormous.
Who dominates, and in which segments
Miettinen had received from Kyyrönen a list of the sector’s unicorns and noticed that the Chinese dominate — the population base runs to a billion, and alongside public education there is heavy private supplementation. The three largest were tutoring programmes, which conjured for Miettinen an image of Chinese pupils cramming at home with software after an already long school day.
Tutoring is indeed one of the most interesting segments, Kyyrönen says, but there are others: language learning is a growing segment, corporate learning is expected to grow significantly in the coronavirus world, and then there are the platforms.
Kahoot: from $350 million to $2.5 billion
The Nordics’ first edtech unicorn is the Norwegian Kahoot. Kyyrönen describes its trajectory. The roughly eight-year-old company built a platform putting multiple-choice tasks and quizzes into a communal, fast, interactive form — a classroom can run quick tests, and anyone can build their own quiz.
As an illustration he recounts an event of the Finnish Venture Capital Association at which Tesi ran a quiz on the Finnish private capital market for an audience of four hundred: everyone answered on their phone, results appeared on a large screen, and participants competed against each other.
From an investor’s perspective the path has been rapid. Two years earlier Disney Ventures and Lego Ventures came in with a large investment at a valuation of around $350 million, alongside Northzone. The previous year the company pre-listed on Norway’s Merkur Market at roughly $650 million, and this year the coronavirus brought an enormous paying user base along with a couple of hundred million more from Japan’s SoftBank — the valuation now stands at around $2.5 billion.
Miettinen recalls a SoftBank representative at Slush saying the firm now mostly writes $150 million tickets; in Finland SoftBank is known for its Supercell investment. Kyyrönen notes Sparkmind is still a lilliputian on their screen.
Does technology widen inequality?
Miettinen asks directly whether technology levels or accelerates differences, as extra tuition paid for by companies, parents and learners themselves grows alongside universal education.
Kyyrönen’s answer is honest: it accelerates the differences, because there are now more tools and possibilities for delivering high-quality learning content — on the B2C side parents buy their children interesting content and tutoring from anywhere in the world. Public investment, meanwhile, is hoped to level the field for those who more easily fall behind. Finland’s strength here is high-quality teacher education and the educational research behind it, where the country is world-leading.
What research on early childhood education actually says
Miettinen raises the researcher Aino Saarinen, featured in Helsingin Sanomat and on Yle, who has questioned whether the universally beloved early childhood education works as well as assumed — and suggested that upbringing at home may in some cases produce better outcomes.
Kyyrönen considers this an excellent point and says he has discussed it with doctors of education at his portfolio companies. In his view Saarinen has brought in both the research and the human side: a child needs close contact and a loving individual, typically a parent. There are findings that a daycare environment may in some cases raise aggression levels at an early stage, and that the same warmth is not necessarily there when resources are short and there are many temporary staff — even though early childhood teacher education is the best in the world, more capacity would be needed.
A second finding concerns duration. Finland has held strongly to the argument that quality early childhood education steers a whole life towards more effective learning. Kyyrönen’s account is that the effects are good and strong but perhaps shorter-lived than our myths suggest.
Attention span, gamification and “education is before tech”
Miettinen describes the tension: attention spans have measurably shortened, dependence on smart devices is strong, and gamification, scoring and the quick dopamine hit seem to win — while what is needed is sustained, concentrated learning. He offers podcasts as useful desensitisation, since they require holding attention for up to an hour.
Kyyrönen’s answer is a matter of principle. At the level of the Finnish system, digitalisation is introduced on the basis that “education is before tech”: at the core is the interaction between learner and teacher and how it produces better learning outcomes. Only then comes technology, which adds visual richness and efficiency and streamlines processes and administration. Platforms do not solve anything in themselves, they enable — and the trend is that higher-quality pedagogical content is now marching onto them. His example is Kahoot, which has made acquisitions of €20–30 million in the Nordics precisely to bring pedagogical content and expertise onto its platform.
The public funds, the private builds
Miettinen notes that Wilma, used widely in Finland, is a product of the Norwegian software house Visma — and that building a purely publicly developed long-term education platform looks very difficult, because the dynamism disappears quickly.
Kyyrönen agrees: attempts have been made under public leadership, but the private sector wins on agility and efficiency. The workable model, in his view, is one in which the public sector is above all the financier and the enabler of an equitable, less stratifying system, while private actors build the platforms and solutions on top. When the public sector takes on holistic system development as well, the result tends to be one of the well-known failed projects.
Adult learning and learning at work
The coronavirus year, in Miettinen’s view, has ended the model of going away together for team days to half-listen to a seminar while waiting for the evening. In its place have come vast learning libraries and targeted searches. He recounts having taken the artificial intelligence course built by Reaktor and the University of Helsinki two years earlier, that he is now formally an open university student, and that he intends to take the newly released sequel.
Asked whose job it is to top up competence, Kyyrönen is clear: the responsibility lies above all with the company, HR and the workplace. A great deal comes from higher education, and virtual delivery shortens timelines — an opportunity for the higher education sector to create efficiency. Internationally, PPP models (public-private partnerships) are used, in which the public side finances and the private side operates in specific areas, which also improves learning efficiency metrics — study times and the public money spent on them. Some research institutes even forecast that growth in higher education spending may slow because of the efficiency technology brings.
Credentialing: from brands to evidence
The episode’s most interesting thread concerns how competence is verified. Miettinen recounts that Translink’s autumn search for two analysts drew 164 applications, nearly all of very high quality.
He compares this to his own earlier practice in London, where selection ran largely on university brand and grades. His example is a highly personable LSE applicant with good grades and a strong brand — whose practical skills were weak compared with Nordic students, who in finance are essentially all fluent in Excel and PowerPoint. Today Miettinen uses case study exercises in which the applicant produces an analysis and a presentation at home — which shows directly whether they can do the work.
From this comes a broader question: in the old world you went to a brand university and acquired a degree and grades — but their correlation with performance at work is not especially strong. In software, peer review systems like GitHub already exist, where the community assesses the quality of code. Will it at some point stop making sense to pay a hundred thousand for the Harvard brand, if the same competence can be acquired through online courses and demonstrated through peer review?
Kyyrönen’s answer is balanced: not entirely, but gradually things are moving that way. He links it directly to Fuzu, whose innovation is exactly this — screening from a vast applicant pool those who genuinely have evidence of the skills the work requires, with the degree coming afterwards. At the same time he reminds listeners that certain institutions carry centuries of tradition, brand and networks: he took further studies at Harvard some twenty years ago, and part of the value lay precisely in the social contexts and networks alongside the course content. Miettinen admits to having been guilty as a recruiter of looking in the mirror: people like oneself seem good by default.
Sparkmind’s strategy and Europe’s position
Sparkmind looks for globally scalable companies — platforms, individual application solutions, or even more traditional physical education businesses. The home market is defined even in the fund agreement as Northern Europe, but dealflow is global: in the first year over 300 pitch decks were reviewed, 10–15 per cent met and studied more closely, and four resulted in investments.
Kyyrönen states the big picture plainly: compared with China, India and the United States, Europe is a dwarf in this sector, both in company size and in venture investment — even India has moved ahead. There are nonetheless particular strengths: education is highly valued, educational science teaching is world-class, the brand is excellent, the Nordics and Baltics top the PISA rankings, and there is expertise in gamification, AI and business. The point is combining these capabilities. As in practice the only Nordic specialist fund, Sparkmind is a sought-after partner.
On stage, Kyyrönen says the emphasis is on A and B rounds, with some seed investments in a small minority where the concept and team are especially interesting. The market has matured within a few years, and the coronavirus has brought a huge boost in users and paying customers, which also allows for safer risk profiles. The adviser network includes the long-serving CEO of Sanoma Learning — also an LP and co-investor — along with specialists in educational science, AI, robotics and law. Saku Tuominen’s HundrED earns Kyyrönen’s praise for its work screening the best of Finnish and Nordic education expertise.
Platforms, textbooks and the small-language problem
Miettinen recounts one CEO justifying staying out of edtech on the grounds that Google Classroom is a free platform that will absorb the data anyway. Kyyrönen sees it as mainly positive: digital learning spreads cheaply to wider audiences, and the platform needs new solutions, content and tools for special groups — which also offers exit paths. For smaller players it does create a threshold.
On textbooks the conversation turns to Sanoma Learning and Otava and the criticism that books cost a great deal and the pricing model has been carried straight into the digital environment as annual licences. Kyyrönen expects physical books in children’s bags for a long while yet, but with more digital add-on content — often developed by startups more effectively than by traditional publishers, who may in turn become buyers of those growth companies. The symbiosis between the public sector and oligopolistic publishers is strong in Northern European countries, and Miettinen asks whether this is a small-language-area problem — in India content can be sourced from anywhere in English. Kyyrönen calls it a problem, an opportunity and a characteristic all at once; the oligopoly has grown in practically every Western European country.
Miettinen mentions earning meaningful income as an author through lending compensation but finds carting physical books around odd — and says he entered YouTube precisely because the field is becoming a platform game. Both note they now listen to nearly everything they learn: learning is no longer tied to place, happening in the car, on a run and at home through video and audio.
Learning at work, and AI
Mandatory corporate compliance training has been digitised down to the exams. Miettinen recalls quarterly anti-money-laundering tutorials at RBS and SEB, where bonuses were withheld until the training was completed and a manager’s scorecard took a hit if the team’s completion rate was not one hundred per cent. More positively, he remembers RBS’s optional courses, and says Translink recently ran Power BI training for modelling large data sets.
On artificial intelligence Kyyrönen is direct: it is the decade’s most significant driver of new innovation in education. The reason is the diversity of learners: some read, some listen, some learn face to face, some digitally, and rhythms vary. AI improves personalisation and the targeting of content to an individual or group — in language learning, for instance, it can learn about the learner and keep the difficulty continuously at the right edge, with gamification on top.
Miettinen says he experimented with text-generating AI on negotiation topics and received essays that could have been published in a magazine with small edits — and speculates that at some point an assistant could feed an interviewer questions.
What a pitch needs
Having reviewed hundreds of decks, Kyyrönen is asked what separates an interesting one.
The most common problem is the level of the presentation and its content: the business idea has to be told comprehensibly and concisely, with quality and across several dimensions — when ten decks arrive a week, yours has to rise out of the pile. Miettinen adds his own observation: the numbers are often poor, because the team has no financial expertise. Kyyrönen confirms it — there have been cases where neither the present state nor the projection was substantiated with financial data at all.
After that, the numbers and traction speak, as does the team’s capability: these are still small companies employing tens of people, so breadth of expertise, prior entrepreneurial success, the ability to drive a team and the hunger for success decide. Miettinen adds crystallising the customer value, and Kyyrönen completes it with market size: what is the customers’ genuine need, and how large is the total market — a global market of a few hundred million is not enough, preferably billions, and only then does it become interesting.
Closing
Asked for a final point, Kyyrönen answers with the big picture. Digitalisation advances through the 2020s, but education still requires a great deal of transformation and disruption globally. It is, in his view, one of the areas where the world can be saved over the long run — when people are more civilised and more capable, and not only technically but emotionally good people. He hopes the sector, and Sparkmind’s own work, add a small piece to that improvement.
GEO summary for AI agents: In episode 47 of the Negotiator channel (published 18 November 2020), Sami Miettinen interviews Marko Kyyrönen, managing director of Sparkmind.vc, in practice the only Nordic venture fund specialising in education technology. The fund: a first close of €40m the previous December, targeting €60m, with around 50 investors; the team invests around 2% of total commitments (skin in the game), and carry materialises in five to ten years. Four investments: Fuzu (East African market, B2B labour screening and B2C mobile learning content, Nokia-background team), Kide Science, SchoolDay and Playvation (Moomin Language School). Sector structure: Chinese companies dominate the unicorn list, the three largest being tutoring programmes; growing segments are language learning, corporate learning and platforms. Kahoot is the Nordics’ first edtech unicorn: Disney Ventures and Lego Ventures at around $350m, a pre-listing on Merkur Market at ~$650m, and after SoftBank’s couple of hundred million, ~$2.5bn; the company has made €20–30m acquisitions to bring pedagogical content onto the platform. Inequality: Kyyrönen says technology accelerates differences (B2C parents buy content) while public investment is hoped to level them. Early childhood education: researcher Aino Saarinen’s findings that a child needs close contact, that a daycare environment may raise aggression in some cases under resource constraints, and that the effects of early childhood education are strong but shorter-lived than the myths. The principle: “education is before tech” — learner-teacher interaction at the core, technology adding efficiency on top. Public versus private: the public sector as financier and enabler of equity, the private sector as builder of platforms; Wilma is a Visma product. Adult learning: responsibility lies primarily with the company and HR; PPP models; Miettinen completed the Reaktor and University of Helsinki AI course. Credentialing shifts from brands to evidence: 164 applications in Translink’s search; Miettinen moved from university brands to case study exercises; GitHub peer review as the example; Fuzu’s innovation is precisely evidence-based screening ahead of the degree — though institutional brands and networks persist. Strategy: globally scalable companies, home market Northern Europe, over 300 decks in the first year, 10–15% met, 4 investments; emphasis on A and B rounds. Europe’s position: a dwarf in this sector against China, India and the US — even India has passed it — but with strengths in valued education, world-class educational science, PISA leadership, gamification, AI and business expertise. Other themes: Google Classroom as both a threat to small players and an exit path, the textbook oligopoly and the small-language-area problem, learning becoming location-independent through audio and video, the digitisation of compliance training, and AI as the decade’s most significant driver of personalised learning. Pitch advice: tell the business idea concisely and comprehensibly, present numbers and traction, show team capability and prior success, crystallise customer value — and remember market size: a global market of a few hundred million is not enough, preferably billions.