How do you digitize implicit knowledge? - Michal Hudeček (Levebee)
Michal Hudeček - CEO of Levebee
Segment: K-12
Business model: B2C & B2B
Location: Prague, Czech Republic.
Levebee makes learning easier, more effective and fun for kids in areas such as math, reading and languages. His background combines deep technical knowledge with business skills proven by multiple exits.
The full transcript of the podcast can be found below for those who prefer reading rather than listening. The podcast is hosted by Frank Albert Coates.
Episode summary
Levebee is an app that helps all kids learn math, including those with special education needs or from disadvantaged backgrounds. The app digitizes math interventions, including learning materials and pedagogy.
Starting from the early math concepts is crucial for effectively teaching all students, especially those with special education needs.
Levebee has more than 500 schools on an annual subscription and is financially sustainable.
The biggest challenge in the go-to-market strategy is the fragmentation of educational markets and the need to persuade administrators. The main challenge for Levebee is understanding and digitizing the implicit knowledge and decision-making processes of teachers.
The goal of Levebee is to build a personalized digital math teacher that combines technology, research, and pedagogy.
Support from Vodafone Foundation and participation in the Impact EdTech Accelerator have been beneficial for Levebee.
Understanding pedagogical research and focusing on the educational value is important for EdTech startups. Identifying the real problem and pain points in education helps startups make informed decisions and prioritize effectively.
Book recommendations:
Transcript of interview
FC: Do you want to start off with just giving us a 30 second elevator pitch of what Levebee is?
MH: Sure. So, Levebee is an app that is helping all kids to learn maths and by all kids we actually mean all kids, including kids with special education needs or from disadvantaged backgrounds. And the way we do it is that we are digitizing what's happening during math interventions, which means we are not only digitizing the learning materials, but also the pedagogy behind the whole process.
FC: Great. Yeahs, that was below 30 seconds. So I think it qualifies! Thank you. So yeah, let's get a bit deeper. And if you want to sort of go back to when you started in EdTech and why did you do it and why did you end up creating this company?
MH: Yeah, my journey to EdTech was not a straightforward path, let's say, let's put it that way. I actually started coding eShops and CRM systems when I was 15 and I started my own projects back then. I was even studying computer science high school. But at some point I realized that, I could learn more about technology by myself.
But, when it comes to business, it's better to learn from someone else's mistakes, which is why I decided to study at the University of Economics. And this is where I met my co founder who brought me into EdTech and we eventually launched together Levebee. So it was through him that I learned about EdTech.
And I actually realized that it's a perfect match for me. It combines all the passions that I find very interesting. It combines the technology. It combines the business and marketing, but it also allows me to do a bit more of the research and the science and education, which is another thing.
The backup plan I could have if everything fails is I might be a science teacher. That's what I think would be fun for me. So, these three things are all combined together and we have a lot of discussions with my other co founder, Mrs. Renata Wolfova. She's a special education teacher, so we exchange a lot of pedagogical books and research.
And my wife is a teacher as well, so this serves as a topic that we discuss a lot during the dinners and walks and vacations. So, since then I think I've been hooked.
FC: Nice, I think those are all good reasons to be in this field, and maybe particularly why K-12, maybe, and not sort of the other areas of education?
MH: I think it's very important to start from the very beginning. One of the EdTech startups we were doing before was also about reading. But it was for effective reading for adults. And we learned that, yes, you can only fix so much when the students are 20, but you should really look back when it's all starting.
And the same with Maths. Yes, you can start, you can have an app for teaching fractions, but you can never do that completely right for the kids with special education needs, for example, because one of the barriers might be that they don't know the previous concepts in maths. So you actually have to start from the early math concepts from the kindergarten to be able to eventually learn fractions.
And this allows us to actually go back sometimes even years back, in the curriculum to fix the gaps. And without that it's really for certain kids impossible to learn fractions. So I think it works better if you start from the very beginning and then work your way up.
FC: That makes a lot of sense, of course. And can you say something about the company and sort of what stage you're in terms of, I don't know, fundraising and market expansion and things like that?
MH: Yes, we have now more than 500 schools on an annual subscription. We are getting close to 400, 000 students in the app. And the company is financially sustainable. We are not looking for funds right now. We might be in the future when we decide to expand or like find out the specific strategy to expand to big countries. But currently we are not forced to. So that's the stage we are in.
We have been around, for the company itself around nine years now, but we started as a reading app in Czech for young kids, but we expanded with reading to Slovakia and Poland. And now for the last two years, we are working on maths, which allowed us to expand to even more countries.
We now have pilots in 15 countries running and trying to learn what's the best way to go to market in other countries.
FC: And I know this is a big area we've discussed it before, going to market. So what are sort of the biggest challenges when you are looking at your go to market strategy and deciding where and when to go?
MH: Especially in Europe, the biggest challenge is the fragmentation of the educational markets. And they are much more different than definitely I imagined before. I think most people imagine that, first of all, like the organizational, administrational things like who can buy software or who cannot. We come from the Czech Republic, so, the schools here are very independent.
I think the independency index is like one of the highest in the world. It's on one side a problem for the government to enforce any kind of policies, which is troublesome in some areas, but for us, it allowed us to go from the bottom. So we could really persuade the teachers and show them; if the app is helping the kids, then they can just go to the headmaster and they will buy it.
But in other countries, in most of the other countries, to be true, this is not the case. You have to go through the municipalities or administrations. The schools don't really buy software themselves. It's usually somehow centralized. And that means for us that we are talking to completely different people.
We have to persuade the administrators, not the teachers. So that's definitely been a challenge. There are cultural differences as well. What do you think is a good teaching? What's the prevailing philosophy? Whether you should be teaching more explicitly or more like discovery based learning in some countries still is a thing.
Differences are actually much, much bigger. And yes, there is also language. Recently we learned that even language, even though we are working on math, which you would say is kind of universal, the language influences a lot what kids struggle with. So for example, in English, of course, you have 11, 12, 13, 14, which is suddenly two different names for the numbers or when it comes to the terminology for geometry.
You have in Czech, if you have the triangle that has all the same sides, the name, official name you learn at school is all-the-same-sides triangle. So if you have a test and they ask you, for example, PISA test, it could be, it is actually in PISA test. So they ask you which triangle is all the same size, you immediately know what to look for.
But in English, they have Latin names for them. So, it's much more difficult. You actually have to know what the Latin name means to be able to find it. So there are definitely some language dependencies as well we are discovering and have to adjust the app accordingly.
FC: And if we move on to the team, could you say something about how you found the team? I heard about your co founder, but like how did you find the team, the rest of the team, and any sort of specific things you would say about how you work together?
MH: Sure. Actually, the initial team was found and organized by my co founder Michal Zwinger. He actually found Renata Wolfova, the doctor, the special education teacher that is now the chief of methodology. He had also assembled the first programmers and graphic designers. So that's definitely to his credit. But recently, after the COVID, we have closed the office completely because no one wanted to go there anymore. Which turned out to be a blessing because we were able to find people outside Prague as well. And that really opened a lot of opportunities for us. We could find much, much higher quality people if we didn't require them to come physically to Prague. And Prague for some reason is a headquarter for a lot of development centers of big corporations so it's hard to find people here.
The unemployment rate is just zero or like minus something. So that helped a lot. So now it's me Dr. Wolfova, two programmers and two marketers, one for the Czech market, one for the international market, and one customer support. So these are seven of us that are full time, but in total there is about 20 of us, more than 20 of us, but that includes the teachers, translators from the other countries that are helping us with the pilots.
FC: And if you look at the business and the idea, you know, what will make you succeed and what would you say is your secret sauce?
MH: What we are trying to build here is this mythical, personalized digital math teacher. I don't think the technology is there yet, but whoever builds this, I think will come from the math interventions background, which is what we have as well. Because that's where the knowledge is. That's the most personalized treatment you can get when it comes to maths.
Especially when you have some disorders or some special education needs. Also I think you will need to be able to build fast, but in the long run. Which is important to add because it's easy to build fast in the short, short term when you can take shortcuts. But as I said, we are around for more than 10 years, actually. So not yet, but it will be 10 years in a few months.
And we are still able to add new features quickly, even with such an old and huge code base. So I think that's very important as well. We are mostly working on new features, not debugging whatever we break. And, the third kind of ingredient is having the channel to the customers.
So, as I said, the technology is not truly there for everything, even though there's a big hype about AI, but I don't think it's still there, but it might come. And when it comes, we need to be able to, first of all, develop the features quickly, but then we need to be able to get them to the users. Which is why I think it's important to build the brand and build the customer base, even now when we still are not fully personalized AI teacher, but it's important to be able to get it to the users quickly. This is why we try to have the direct contact with the teachers and the students and to build the brand, to be trusted and have the expertise to build on.
FC: So many would argue with you and say that, you know, with the release of these large language models, you know, we're already in sort of next generation of what is possible. So I guess the question is what do you think is still missing for this to be truly useful in education and apply to what you do?
MH: That's actually a question we were asking ourselves, which is why we did our own research and we recorded videos from the math interventions. And we recorded the interactions between the teachers and the student. One teacher and one student, one on one. And, we then transcribed it and then looked really at each line that the student and the teacher said.
And categorized it into certain areas, based on what was the trigger. Actually, how the teacher decided to do that or say that. And, in fact, there is a lot of things that the teacher is monitoring. It's not just the knowledge, whether the student got it right or not, they have to monitor the cognitive load, or they have to monitor the self esteem, motivation, fatigue, it's just sometimes just taking too long and you can no longer work with the child, or the attention.
There are a lot of other variables than just the knowledge that the teacher is monitoring, and simply we don't have the data for that. We are not able right now to measure how tired the child is, or of the self esteem. You know, the teacher, when they see that the child may be on the brink of crying sometimes, which might happen in the special education environment.
We just had a recording last week with a child that was in the fourth grade and didn't really understand concepts from the first grade. But he knew that it is first grade curriculum, so when he failed at that, he was just starting to cry because he knew it's something he was supposed to do.
So, my colleague actually had to intentionally kind of break the methodology and go a little bit up, to try to explain the exchange in addition and subtraction on a little bit higher numbers, even though that's not usually what you should do. But it helped that child to feel a bit better about failing, you know, if he doesn't understand something that is maybe bigger numbers, then it's okay.
So these are kind of decisions that the human has to do. And that's, it's not about that AI couldn't do it, but we don't have a way to measure it right now. I believe that's the biggest, biggest bottleneck. Maybe some technology will come in the future that will allow us to do that, but we are definitely not there and the knowledge part is just I think smaller, smaller part of the whole package of what the teacher actually does during the math interventions.
FC: And probably also in a setting in K-12 or for kids, even much more so than for adults, right?
MH: Sure. And especially if they have some other special education needs. They might have attention disorders or dyslexia or some other, aphasia for example. So it makes it much harder.
FC: And just a final question on this. Do you think there's more needs for this now? How do you sort of see the students evolving and the needs of students in, well, in societies that have changed a lot over just the last couple of years?
MH: I think it definitely is a topic. Like special education or equality in education. That is definitely more and more popular. A lot of countries went through some kind of process of inclusion, which means, like placing,all the kids together in the same classes, which makes it much harder for the teachers to actually teach.
And they have to have some tools that make it easier for them. So I think that trend will definitely continue. And I see a lot of, even from the EU perspective, it's one of the priorities and there are conferences about it and some even funding support for the projects in this area.
So I see a lot more interest in special education than for example, 10 years ago, for sure.
FC: And if you now take a step back and look at the company since you started it, what's been the biggest challenge and how did you solve it?
MH: Yeah, I think the biggest challenge was the part I described in the previous question about getting to know what is actually the role of the teacher. But the tricky part is that the teachers sometimes don't even know it themselves. This is called inexplicit knowledge. It's just they do it kind of automatically.
And if you would ask them before, they wouldn't tell you that they are doing it. And just only after you record it and ask them, why did you say that in this situation? They kind of realize, oh, okay, I was reacting to that. So, for the knowledge itself, there is a lot of literature.
Fortunately, with maths, there is a lot of research you can rely on about what works and how the things might be connected. In general, Math is researched a lot. There are still a lot of gaps, but compared to, let's say, financial literacy, there is a lot of research you can actually base your app on, which is not everyone does it for some reason, but that's another story.
But this part, this kind of inexplicit knowledge that comes from the experience of the teacher, that's a little bit more hidden. And this has been a challenge for us to digitize and put in the app. And I think we are still at the beginning of doing that. So I would still say that has not been solved to our full satisfaction.
FC: Okay, really interesting. And if you look at resources and support that you've been able to get along the way, is there something that you would highlight? Something that really made a difference on your journey as an entrepreneur?
MH: Yeah, I think at the beginning support from Vodafone Foundation helped us a lot. They gave us the first little bit of money for creating the prototype. And without it, that would probably not be possible. More recently we were part of the Impact EdTech Accelerator. Which meant that we were among the top eight startups in Europe in EdTech.
And this allowed us, first of all, to get more funding for starting the math and also it helped us to build the international contacts. So this is the main reason why we were able eventually to run the pilots in 15 countries, because we met a lot of interesting people from various countries in there.
And we were able to get in touch with their friends and friends of their friends. And this is how the pilots were made possible. So definitely I would highlight that. And, regarding the know-how, what's been interesting in English speaking countries, mainly UK; there has been a movement in the last 10 years towards more evidence based education and EdTech tools, of course, as well.
And I think what made the difference was that suddenly the teachers started to write about how to use this like cognitive science research practically in classrooms. And this is something we are definitely capitalizing on and it allowed us to learn how to explain the research to the teachers as well.
That's been definitely helpful. So books like Why don't students like school from Daniel Willingham.
Specifically for maths, Craig Barton, How I Wish I Taught Maths. These are all great examples of, like how to practically apply the research, educational research in classes, which for us means how can we explain to the teachers how the app works and what kind of ideas it is based on. So that's been a big source of inspiration.
And of course we are getting inspiration from other companies as well. There are companies like in US, Renaissance Learning. They are very good at doing the assessments. So we are looking at how they do the research there. And there is Eduten in Finland. Which is a case study in how to sell your country brand. They are based it on, you know, like, Finland's education. And they do a really good job at analytics. They're providing analytics to students, teachers, and even the whole municipalities. Or, Mathific, Australian company that just builds them on quantity. They have a lot of, lot of exercises and are very hard on localization. They are in many languages. Or the Koobits or Khan Academy for explaining things in math. So there is definitely a lot of other companies doing a good job in maths, and we are trying to learn from them too. So, that's interesting to watch for us.
FC: And, if you look forwards in the next 6 to 12 months, what is keeping you up at night?
MH: It's definitely the international expansion. The biggest question is how to replicate the success we have in the Czech Republic and some surrounding countries, and other countries as well. And that's definitely the priority number one. And if we figure out a way how to do this in other countries, there's definitely a way for us to get much, much bigger, maybe even take some funding.
But, we need to figure out a way to do so. So that's definitely my highest priority right now.
FC: Yeah, and a common one for many entrepreneurs inside and outside EdTech, right?
MH: Sure.
FC: And if there's any advice you could give to other EdTech startups, what would it be?
MH: Definitely look at the pedagogical research. For some reason that is still kind of undervalued in startups. There are a lot of startups that still play on the learning styles, which have been debunked a lot of years ago. Or just focusing too much on the fun part. It's like making something fun, making maths fun, making reading fun.
That usually leads to nowhere. I have a list of companies that died trying to do that. When you try to focus only on the fun, you are eventually competing with everything that is fun. So I still believe that the educational value is more important, like finding a way how to help teachers.
It's probably going to work better in the long term. And then there's the other startup general advice, just trying to understand what the real problem is. That helped us a lot to be able to specify the situation. In our case, it is teachers that see that they are trying very hard to teach a student something, let's say multiplication, and the student's trying as well, but it's falling flat like it leads to nowhere. So this is the situation where we want to jump in because now we can help the teacher to find out what the students really doesn't know and they can go back and we have the diagnostic and everything kind of fits into place when we were able to find this specific practical situation, real pain point in the education and like build everything around it.
And that took us time. It's very hard to come up with it from the blank slate. So it takes time, a lot of feedback, a lot of thinking, a lot of trying to present the app to someone, but the clarity helps a lot. Once you get it, it suddenly makes it much easier to make the decisions and decide what the priorities should be and what you should do.
So the understanding of that problem, I think is what would help a lot of startups to speed up.
Track 1: Very good tips. Um, is there anything else you want to share with the ecosystem?
MH: Just join the EdTech Garage for sure. I would love to meet more people there. It's very nice to be speaking with other EdTech entrepreneurs and see how it is in different countries. So as I mentioned that the educational systems are different. This is really a good place to meet people from those educational systems and learn about how it works.
So I've had a lot of interesting conversations for sure there.
FC: Good. Great to hear. And with that, that's my final question. So I just want to thank you so much, Mikael, for coming on the show and sharing all your experience. I need to put my kids now in front of some of your apps to supercharge their math skills and other skills. So yeah, thank you so much for coming on the show.
MH: Yeah. Thank you for having me.
