Language Learning #3 — Toby Mather from Lingumi
Biography of the speaker
Toby Mather graduated from the University of Oxford with a Bachelor of Arts in Russian and Italian Language and Literature. During his undergraduate experience, he interned at Memrise, headed the Cherwell Newspaper at the University of Oxford as Broadcasting Director and worked in the innovation space as an Associate for Emerge Venture Lab. After attending Entrepreneur First, an international Talent investor programme, Toby founded Lingumi, an English language learning programme for pre K-12 children.
The interview was recorded virtually and transcribed by Zanichelli Venture.
Question 1: Lingumi was founded back in 2016 and has been growing extensively ever since, leading to a position in the top 100 most promising EdTech startups in Europe. Can you tell us more about your company?
TM: Lingumi is a teaching platform for pre K-12 pupils. We are building the first layer in the global education stack based on the premise that if you are a parent, you expect your child to have their first teacher when they are 6 or 7 years-old. The expectation finds its roots dating back to Socrates’ time and even back then the main scaling mechanism to get a teacher was the classroom. But as I was developing Lingumi, I wondered: why do we need to wait until children are 6 or 7 years-old? Probably because that’s the earliest a student can sit still in a classroom to learn. At Lingumi, we believe that parents should expect their child to have their first teacher when they are 2–3 years-old, when the child is in the optimum neurobiological state to learn. Technology is facilitating the process and the scaling mechanism can now change with our new learning paradigm. In the language learning space, entrepreneurs need to solve many problems: cost of access, quality of supplies and teachers, access to both teachers and children and more. We believe that live teaching is not the way forward for pre-schoolers because 1) it’s very expensive, 2) if you scale up your teacher supply, the quality of teaching would inevitably go down and 3) psycholinguistically and psychologically a live learning experience can be quite intimidating for younger children. In our learning approach, we take the most extraordinary teachers and we study what they do when they teach (e.g. the words the use, the games they play, the way they deliver their classes); then, we move all of this to an asynchronous learning setting. The best comparison to our model is the company Peloton: they use the most extraordinary fitness coaches and customers can do one of their classes with their friends at any time you like. This is an evolution of the way we currently learn and it is really exciting to me. If you want to transfer this to our product, Lingumi is trying to build the best pre-school in the world. When you open the Lingumi app, you can learn anything in the best way possible. We are currently focussing on English but later we will expand to other subjects.
ZV: Very interesting. It is definitely true that children of younger ages can learn anything much faster, so leveraging this aspect can bring huge benefits to language learning. How did you come up with the idea for Lingumi?
TM: In my undergraduate degree I studied Italian and Russian languages at The University of Oxford. I then spent 1 year in Russia and 1 year in Italy teaching English: two very eye-opening experiences. As I entered the English classrooms, it felt like it was day 1 of English for every child, even if they had probably already spent 3 years of learning English. Students should have been building on existing foundational knowledge and that was simply not the case. After those two years, I went back to university and I started to research topics around language acquisition and psycholinguistics to set the foundation for Lingumi: a self-led gamified platform with a well-structured pedagogical paradigm behind it. We also noticed that the real magic ingredient in a classroom environment was the relationship between the teacher and student, rather than the content. Based on this finding, we are building a social learning experience at Lingumi, where young children can build a meaningful relationship with their teacher. We occasionally hold live events for our Lingumi users and it is very heart-warming to see the young learners coming together to see and talk to their Lingumi teachers. From the dark jungle of self-led learning, we are creating new learning experiences where the students feel connected to, and empowered by, their teachers.
ZV: I really resonate with your point about building a meaningful relationship with your teacher — it really does make a difference when you are learning new content.
TM: I have probably had 30 teachers in my entire life but there is one who really changed my life: their love for English literature, grammar and punctuation was palpable and infectious. It might be hard to find that one teacher if you don’t live in a big city, but technology may just help you with that!
ZV: What made that teacher so magic to you?
TM: They really loved what they were doing. As a simple — and frankly quite geeky — example, they taught me how to use the semi-colon. Most English people are so afraid of misusing it but they loved it so much. The possibilities of a semi-colon are so interesting, ranging from poetry to classic literature and grammar. Unfortunately, asynchronous teaching is quite exhausting. Maintaining a high level of passion and sharing it with every children in the room is an undoubtedly huge challenge; however, I am a firm believer that even the most inaccessible topics can become extraordinary if taught in the right way: for example, Roberto Benigni, an Italian actor, is famous for his reading of the “Divine Comedy’’ by Dante Alighieri; he delivers the content in an inspiring and captivating manner that makes an otherwise inaccessible literary content very pleasant.
ZV: I love how the human interaction is at the centre of any of your analysis; at the end of the day, we learn languages to speak to other people, and not to robots. On another note, we found your publication “The Deliberate CEO” quite insightful; you share very interesting insights to aspiring entrepreneurs and I feel there is so much to learn. Would you be able to share with our readers a key lesson you learnt as the CEO and founder of Lingumi?
TM: I would definitely think more carefully about my early steps. As the founding CEO, you feel like you want to do so many things once you get through your first fundraise. I would definitely make the transition from “selling to prove market-fit” to “selling to scale up” a lot smoother. I would probably choose a metric that meaningfully represents your business, whether that is acquisition cost, churn rate or customer lifetime value and I would be absolutely ruthless with it: I wouldn’t set any marketing budget before fulfilling that metric objective.
Question 2: On a recent HolonIQ Executive Brief, we notice that there has been a growing interest in the use of technology language learning. In what way do you think that the current advancement of tech can help people learn languages better? How is Lingumi leveraging the technology?
TM: I would not consider myself a tech expert — Lingumi is more a teaching platform than a language learning app (though we did start with language because this is what I taught in the past). Within language learning there are lots of tech stack layers that are being commoditised: voice technology, low latency cloud delivery, spaced repetition algorithms and more. These are all becoming basic table stakes. The trends that excite the most are those where the human element is the richest because this is what we learn languages for. I really admire a couple of founders that are doing this with their companies. Ed Cooke from Memrise is incorporating videos of native speakers using casual language in the street and I found it incredible. This is how I first started learning my first few words in Mandarin. I still vividly remember the guy who taught me my first Mandarin sentence, sitting in a street in Beijing while playing chess. At Busuu, the team is making co-learning part of the experience: one can get their language corrected by other learners, creating a community around learning. From a tech perspective, there will be undoubtedly new advancements in speech recognition, video delivery and all the other tech aspects, but we will soon reach a point where these will all become commodities. The real question will be: how do we leverage such technologies to provide the best learning experience that is distinctly human?
ZV: It really comes across clearly that tech should serve and accompany something that is fundamentally human.
TM: People talk a lot about AI, ML and learning personalisation in the language learning space and that is very useful, but what is it really serving?
ZV: We might indeed run the risk of just getting the content right to get some “game points” as opposed to actually learning the language. Do you see this as a risk for language learners?
TM: Yes, I absolutely do. The jump from an app such as Duolingo to the real world is incredible. I can give you an example: let’s say I am in Italy and I walk up to a hotel desk. The level 1 of language learning would be saying simple sentences such as welcome or hello. In level 2 you would probably be required to understand and answer some questions. In level 3 you might have to deal with something that went wrong, something unexpected. Nobody has attempted to recreate a language environment like this so far but it would be a great example of how technology can help you get hooked into the game simulation of a real-life experience.
ZV: That’s a very interesting example. Technology can also be used to shape your short-term results so that learners don’t feel discouraged on their way to mastering their language.
TM: I agree with this and I have experienced it myself recently while I reactivated my chess.com account after watching “The Queen’s Gambit” on Netflix. On chess.com one can play lessons or matches. In the former, there is a staged environment for specific challenges and I was surprised by how good I was at solving individual scenarios. But when I tried my first match in the easiest computer setting, I got absolutely destroyed and I lost in 7 moves. All the safety of the paddling pool just goes away when you get into the actual swimming pool. What is really nice about chess.com, though, is that you can use the “hint” button that highlights the moves you made and lets you trace your steps back to understand where, when and why you made errors. You can then repeat your sequence learning from your mistakes!
Question 3: Many learners now turn to language learning platforms or applications to improve their language skills; at Lingumi, your market focus is on pre K-12. Thinking about the future of this education space, do you see language learning applications incorporated into the more traditional learning paths in a B2B model (e.g. Lingumi for early school, high school, etc.) or will they be more suitable for a direct B2C consumer use?
TM: Possibly every configuration you mentioned. There is currently no tech-specific business because every business is a tech business. What is interesting now is the handover from offline to online as you move across multiple providers — the integration across platforms will become a lot more important that the actual services provided, as Ben Thompson, the author of “Stratechery”, a newsletter on tech and media trends, elegantly puts it.
ZV: I can see why integration will play a much bigger role down the line. How do you think the platform integrations might affect the learning experience?
TM: If I were to continue to learn Italian, I would probably be able to share my language level with my teacher or I could perhaps take a language entry test to identify my proficiency level. Teachers might be able to gather a lot more data through a few conversations than computers. It is also difficult to let a computer know what my level is right away without spending a considerable amount of time onboarding onto the platform. The handover approach in learning a language is complicated and highly dependent on the company; it might discourage students quite quickly in their learning journey if it is not properly addressed.
ZV: Knowing your starting level can definitely help to advance your language knowledge. Though, it can be argued that a technology-based, and perhaps AI-powered, platform might help creators collect more data on the overall performance of the student, depicting a better idea of their proficiency level.
TM: That is true, but it only solves part of the problem. For example I am currently using Anki to do vocabulary exercises; before that I used Memrise and Quizlet. I couldn’t unfortunately pass the history of what I had learnt previously, which meant I had to restart from scratch. Also, the new platform does not know at what point of my decay learning cycle I am, making it more difficult to provide better content for me. These are undoubtedly niche problems and I am curious to see how these will be tackled in the near future.
Question 4: What is the key to keeping little children engaged and motivated, and how can you know and prove to parents/early schools admission boards that the teaching has been effective?
TM: Well it is as simple as collecting basic metrics such as how many students come back to the class the following day or how long students spend one each section. In early ages I think it is key to build a habit. Children are curious, they want to learn and discover. I was really excited to attend my favourite teacher’s classes. For children, learning language is joyful, playful and interactive — these are simply requirements and not choices. We can then think a lot about attention span, start and end of lessons, how not to exhaust the children, the number of words to learn and a lot more. We are very far from getting it right but we are moving in the right direction.
ZV: Thank you so much for all your insights, Toby. Is there anything else you would like to add as a final note?
TM: There is currently a lot of enthusiasm around tech. We are probably underestimating how education will change massively in the next 10 years. It is a continuously rising sea level, and not a tide. Technology and even COVID-19 has accelerated a major shift in the learning paradigm and the developers, such as companies like us, are just beginning to create the solutions that will be the major delivery of that change. Education will be completely different in 10 years’ time.
Key takeaways
- The relationship teacher-student is an invaluable resource for any learning process.
- Asynchronous learning constitutes an economically viable and feasible solution for scaling up in the pre K-12 language learning space.
- While it is clear that technology is changing the way we learn and interact, we must always keep a human-centred approach in designing new products and services.
Closing statement
We thank Toby Mather for taking the time to speak with us; if you want to read more about Lingumi, visit their website. Have a look at our other interviews on the Language Learning space: