Founder of Flo Health: Dmitry Gurski

Being in Imperial, most students end up in McKinsey when they leave. And you said a while ago that most smart people go to McKinsey and founders are crazy. Can you elaborate?

It's very simple. If you're smart, you should make a proper evaluation of your chances of success. And then you start making such evaluations. For example, I will start my company. And then you start getting numbers. Maybe one startup from, I don't know, 50 would get a seed round. Then just one company from 1,000 companies which got a seed round would become a unicorn. And even in Y Combinator, we discussed, they graduated 6,000 companies.

And if you exclude the fake unicorns of 2021, maybe they got 50, 60 real unicorns. It means that even in the case of Y Combinator, a real success rate is maybe 1%. And then you start calculating your bets and thinking, should I put several years or many years of my life and half of my health to a chance of success, which is measured in 0, 0, 0, 0, 0, 0.1? No. But to go to McKinsey and get a nice salary and make beautiful slides about nothing.

So is your advice for people to go there?

If they're smart, of course, because it's like a definite road. You're just making your slides and then you're talking about using very smart words about nothing. Okay, there you have it, guys. I know why CEOs pay to McKinsey. It's very simple. You always know all answers on your own. But you just need external justification of unpopular decisions. And then you say, look, McKinsey says that we must make this layoff or close this project or start this project.

Well, let's follow. And then it's not your responsibility and you pay for that McKinsey. It's the single reason why McKinsey exists, honestly. Just because they're taking responsibility from your shoulders. Is there an event for McKinsey here, guys? Nobody would raise their hands. But with the ChatGPT Deep Research, I think McKinsey will have difficult times. Now I use the ChatGPT Deep Research a lot and I don't need any consultant anymore.

I'm sure they're using it, too.

Of course, but they use the same ChatGPT and ask for half a million and they may use it for free and get results not in six months, but immediately.

Dmitry, we see you today after massive success with Flow. I heard one of your first businesses was selling mushrooms. Can we go there?

Not like the same kind of mushrooms as you might buy on Comden. Just normal mushrooms. You may buy them on some food markets here. But they were buying these mushrooms from us in Belarus and pay like one buck. And they sell here for 100 bucks per kilogram. Maybe, you know, just like small yellow mushrooms. It was like a very popular way to earn money for kids in Belarus 25 years back. It was a good way, but a very risky business, almost a venture business.

Because if there is rain, then you may get mushrooms. No rain, no mushrooms. And then when the rain is good, there are many mushrooms and the price goes down. And you should run to a wood very fast before the market reacts. Because if you just bring mushrooms first, the price is the same. But then closer to evening, all kids would bring mushrooms to market and the price would go down. It's almost like a stock market or venture capital.

But how do you go from that to actually starting your first tech business?

Before tech business, we had book publishing. And before book publishing, I was writing books on my own, computer books. And before that, I was writing articles. And before that, I was making some simple work on the Internet, like websites. It was 2000, 2009. I'm properly old. I started to work like 25 years back. And then it was a moment when the App Store was opened in 2009. And we understood that it's almost the same opportunity as it was when the Internet had launched in the beginning of the 90s.

And we decided to jump to this opportunity. And we started our first company, which was making support apps. It was called Support.com. We owned the domain Support.com. And this company was acquired in 2012. And probably it was a very huge mistake at the time to sell it because the price was not especially high. It was more like acquiring by a playtick. It was a big company, which was earning money from Internet casino to get mobile engineers, but not even support apps.

And they paid maybe $16 million or so at that time, or $12 million. I don't remember. Something insignificant, really, by standards of exits. But then all our competitors, several years after, they exited by price in hundreds of millions. For example, MyFitnessPal got half a billion. Runtastic got $150 million. All of them exited much, much more expensively. That means that if you just had waited a bit, you would have exited much more expensively.

But in any case, we got some experience and money. And we started using this experience and this money. Other companies, including later Flow, quite successfully. We had six exits after that, including exits to Facebook and Google. And we now own not just Flow, but several scout companies, like, for example, Simple with $150 million in revenue. It's a nutrition app.

What were your learnings from those exits?

It's very significant to understand when to sell. Because you may sell too early, but then you may sell too late. I remember that in 2016, we got very significant success with an app called Prisma. This app was making filters based on your pictures. And it was run on machine learning. It was the first app of this kind at that time. Extremely, extremely viral. It was one of the most downloadable in the App Store at that time.

And it was bid from, probably, Google at that time to acquire this company. And the price was rather decent. But then the founders and we got a bit greedy and asked for an insane price. They declined. And then this viral moment ended. And we still own this product 10 years after. And it has a bit of a miserable existence. We're getting some money, but not growing, not dying. Some kind of a zombie state.

And nobody wants to buy it. It means that it's very significant to understand the right moment. Because too early, you may miss. But then it may be even more dangerous to wait. Because then you may just get, like, your basket will be full of rotted apples. And it's really, really difficult to sell rotted apples. Sell when your apples are still fresh.

And how did you think about Flow? What was the idea generation?

We always mostly were working with health apps. Probably because of our own interests and maybe even education. Because my education in university was very close to medicine. It was the design of drugs, pharmaceutical chemistry. And the idea of Flow was based on our previous experience with other health apps. But also it was an observation that this market was quite big if you measure by number of apps and installs.

But also quite shallow, quite underserved if you measure by depth and quality of products. And we had a very simple idea to bring to this market. We called it a super app, but a product with deeper depth. Because the market was full of hundreds of what we called pin calendars. And nobody had the ambition to create something deeper than a pin calendar at that moment. And that was ultimately our idea.

Was it a super app idea from the beginning?

Yeah, yeah. It was the initial idea, yeah. Okay. But we didn't expect that it would take 10 years and hundreds of millions to implement it.

I heard that the first two iterations of Flow failed. Was this true?

It was not. We managed to create a successful version of Flow from a short attempt. The first attempt failed mostly because of a conflict in a team. And most of early-stage startups die because of conflicts between team members. Because when people have extreme tension, extreme pressures, they just crack. And then we just repeated the same. And we started to create a more sophisticated version of a product and a simpler version of a product.

And our bet was about a more sophisticated version with many functions. But then this simple version of a product, which was just a more sophisticated product, but with fewer features. It got traction, and we understood that simplicity is a key in the case of consumer products. Then we doubled down on this product and we killed another one.

But how did you know early on that it was an idea worth solving? I believe you spoke about retention somewhere. What kind of retention early on gives you that?

I think we didn't understand the fact that retention would be so high when we planned that. And it was quite a shock. But then when we saw that retention was so high, we understood that it would become very huge inevitably. The main observation was that if the market supports so many apps, and it was maybe in the top ten of health categories at the time, five apps were peer-to-peer trackers, there's a place for a more sophisticated product.

And like products we were making for other segments. And it was more of a logic like that. But then retention defined everything. Because without retention, we would not get our huge amount. And without our huge amount, we would not get so much word-of-mouth traffic. And without word-of-mouth traffic, we would not grow organically so fast. At the end of the day, everything was because of this huge, totally addressable market.

How early did you raise your first round?

In 2016.

And when did you start the flow?

2015.

How difficult was it?

Well, not especially difficult, because we had exits before, and it was more kind of bad, oh, these people know what they're doing. And because of that, they invested just... Probably they didn't believe too much in this product, but probably they believed that we would figure out something. But in the case of early-stage companies, you always mostly make bets on the founder, because then it's always very different than you planned at the beginning.

And all these initial projections are pretty much meaningless, and everything will be different. And because of that, it's pretty much about the right market and the right person. And everything else is pretty much meaningless at the early stage.

I believe Max Levchin was running a competitor at the time?

He was not running, he started competitor in 2012. And it's a pretty interesting story, because it was a company called Glow. And it was started by Max Levchin three years before Flow. They got almost immediately 30 million from Anderson Horowitz and Founders Fund, it was in Silicon Valley, and then they failed. Now they're 100 times probably smaller than Flow. And why? Mostly wrong product strategy, because they had a bet on fertility market and B2C segment of fertility market, but it's a problem of LTV to CAC, because usually sustainability of business is defined just like your ratio between your full value you're getting from customers, LTV, and a cost of acquisition.

And to make business sustainable, LTV to CAC should be quite big, and early stage very big. And in case of fertility markets, the problem is that people don't pay much out of pocket, and it means that LTV is not high, but customer acquisition cost is very high, because it's difficult to find this cohort of users, and it's very competitive to buy such users, because businesses with B2B business models are betting a lot to get them.

And it's an inevitably weak model, and they failed this segment, the B2C segment of fertility market, and pretty much all companies failed that, exactly because of that, and there are no big players in this segment.

I think you have all of these questions from investors early on. What if Apple gets in to compete with you, and what if Google gets competing with you? I think in 2021, you had Apple getting into the market with a similar product to Flow, and I've seen you posting that there was zero difference in the rate of growth of Flow after that. Why did this happen?

Investors have been talking that Apple Health would kill Flow for five years, and Flow is growing, growing, growing, and nothing happened. And even when we were raising our round from General Atlantic last year, it was one of the main reasons of decline. But I'm not so much worried, because Apple, they're not building a deep product, they're just building a simple feature. And Apple started maybe 700 different apps, and it became monopolized maybe in one or two from 700.

And mostly because of absence of deep focus and specialization, this advantage to be pre-installed is not enough if you're not bringing enough good product.

Did you see any advantage in your team when competitors like them launched? Did people actually work harder?

Yeah, good competition, of course, is very significant, because when you don't have a good competition, a team may become very complacent. And good competition, of course, adds some drive, and you may use it as a natural way to energize a team, because especially big companies, they naturally become quite slow, and let's call not lazy, but a bit in work-life balance, and then you need some motivation, external motivation to drive it. And it's a bit difficult without some external pressure and external crisis.

If we go back to the first fundraising you have, and the last race you have, what are the real changes in the fundraising process?

Not especially, pretty much the same. You need to meet 100 or 200 of them, get 95, 97, 99 no's, and to get one yes, and it's pretty much inevitable.

Why is it inevitable?

Because you need to find a firm which understands your business models, then you need to find a partner without traumas about similar products before. Then you need a firm and partner who are really investing in this moment. Then this person should believe in you so much to bring to an investment committee. There are so many different elements that it's almost like dating. What's the way to be successful in dating?

Just go on a date, go on a date, go on a date, like one day. But it's absolutely the same process. It means that you're just getting like a no, no, no, no, no, no, no, no, and then you get the yes. But it's a very extremely soul-breaking experience, especially for the first time, even when you know rules, even when you know in advance that you will get like 100 no's to get one yes. Nothing changes as difficult as it's been.

But initially, our first round was 300,000. Our last round was 230 million. But I can't say that the process is too much different. Due diligence is very different because at this late stage, due diligence is extremely heavy. It may be like thousands of hours of work and millions in expenditures to finish. When you're raising hundreds of millions, but for the rest, for the process, it's pretty much the same.

We had the Revolut earlier. We have George that worked in the Revolut earlier. Then one of the discussions we had was they've been running these experiments for roughly 40 to 50 experiments at any given time when it comes to product. I think you have a much more extreme approach there, which is, as we spoke earlier, you have 300 experiments running at any given point when it comes to creating a new product and a new segment. Can you elaborate on that?

Experimentation is really significant, and such optimization may be really powerful, especially in monetization. For example, when we started our onboarding with Paywall four years back, conversion to trial was maybe 2%, and now it's almost 20%. It's just like 10 times increased just because of thousands of experiments and optimizations. The same with other parts of product. There's always a diminishing return.

I think maybe in comparison with Revolut last year, we lost a bit of a good balance between optimization and making new stuff. We probably over-squeezed lemons, and we didn't grow enough new lemons. Now we are refocusing to build more new stuff rather than to optimize it. This experimentation may be much more powerful than people may think.

What's the biggest surprise?

How small changes may impact everything. For example, you may change one word on onboarding, or one button on Paywall, and conversions would move 20%, 30%. It just means that people act pretty much emotionally, irrationally, because, of course, you can't find a rational explanation why you change one word, and then your conversions change 10%, 20%. It also means that you can't really predict what would work, especially in monetization. You just need to make proper hypotheses and experiment. I have had thousands of experiments, and I still can't predict what would work.

What about the new products you mentioned? How do you change experiments to actually create new product lines?

For instance, sometimes you may see that something works, and then just a kind of side effect. And then you double down on this side effect, and you may grow a new big part of business just because of this random side effect. You just need to recognize that there is some promise in something, and more often it's quite random, and something you don't really expect. Could you build a sophisticated product based on just research and analytics?

In B2C, I would doubt, because in B2C, user experience is really significant, and even more significant than how powerful are features. It may be achieved by really more evolution, by making iterations, hundreds of iterations, thousands of iterations, rather than by projecting.

What's the toughest moment for Flow that you had as of today?

Hopefully, all of them in the past. Toughest? We had many of them. Maybe the most traumatic for the team was a moment when the Ukrainian war started, and at that moment we had 70% of our employees in Belarus, and we had to ask them to make a choice or to be relocated in three days to Europe, or to lose the job. And people had to move, and 90% of them moved in several days, and many of them have never come back.

For example, I have never come back since 2020 because of political risks. It was quite traumatic. But also people showed a huge trust to the team, because could you imagine that you need to make a decision to move your family in two days, and just start your life in another place from zero, and you're not a student. You have, like, everything established. I moved out of Belarus when I was 27 years old.

And I had everything established at the end. It was, like, just a decision which was done in two days, and we have never come back after that. It was tough, but then, like, I think we changed for better and we became stronger.

If you just arrived with a billion dollar valuation, how do you go to the $10 billion?

It's not so easy to imagine, not so difficult to imagine how we need for that, because there are benchmarks at the market. Like, in our business models, there is, for example, Duolingo. And Duolingo now costs $17 billion. And why Duolingo costs $17 billion? Because they're still growing well, like 40% at this stage, and they have maybe $700, $800 million revenue and good profitability. It means, like, it's very, like, answer, like, OK, $500 million in revenue, maybe 25% EBITDA margin, and growth, like, 30%, and it will be a $10 billion company.

And it means that probably, like, to double our revenue and to increase profitability and to stay at 30% revenue growth, and it will be a $10 billion company or so.

What are your thoughts on AI in health?

AI in health is a big, big topic. The main, I recently, maybe it's changed, but maybe one month back, I checked, like, a website of FDA. Is it any GNI solution which would be approved as a medical device? And zero. It means that, like, two or three years, even after, like, the first version of ChatGPT, the first version of ChatGPT, like, oh, yeah, three. Still, like, zero companies got FDA approval for any GNI solutions.

There are, like, some, like, AI solutions which were approved before, mostly in recognition of, like, different, like, in radiology, like, in different aspects of radiology. But in GNI, like, nothing. And the reason that in medicine, this regulatory process is, and safe process, extremely heavy and extremely slow. And, of course, for a reason, because it's about the safety of people. But I would say that in health, the main difficulty is not to launch a product or to use GNI.

The main problem is to pass regulations. And at this moment, kind of, like, pretty much all companies, they all exist in the gray zone, or they didn't manage to do that. And probably even regulators, they don't know really how to regulate such companies. I think it will be, like, a big process to figure out how they will do that.

I think the other topic that is interesting is that there is a lot of wearables, especially women's wearables, that are growing a lot lately, would Flow get into hardware at any point?

I think no, because in any case, the market will be concentrated around, like, several brands. And because you just don't have, like, enough fingers and hands to have many devices. And now the market is super concentrated. For example, in the United States, among our users, like, 70% is Apple Health, then maybe 5%, maybe 10% or so, like, Fitbit, now probably less. It's declining. And then, like, maybe 5% over, and everything else much smaller.

Because it's so concentrated market that it's, like, no sense to do that. What's necessary is to have, like, a proper integration with such devices.

You spoke about fast decision-making being more important than no decision-making in the past. Can you elaborate?

It seems like what the problem is, like, prolongation of this idea about smart people. But my observation and all the research about that, like, what made, like, CEOs successful. And it's interesting that it's not, like, a level of education or IQ or pedigree. It's decisiveness. It's the signal factor which predicts the success of a CEO and probably entrepreneur. And it's, in fact, that the most educated and smartest CEOs, they show, like, worse results than people who are less smart.

Because smart people, they tend to overthink and tend to overanalyze. But the problem is that your overanalyzing can't help. Because when you're running a company or when you're starting, like, a startup, you're pretty much, like, working with something that is not here. It's in the future. And the future is unpredictable. It's unpredictable for McKinsey. It's unpredictable for a CEO. It's unpredictable for any person.

Because if it were predictable, then, like, a CEO would take McKinsey and they would just, like, stamp, like, unicorns. Like, they have money. They, like, what's the problem? They have, like, they may hire the best people. Why not to stamp, like, 100 unicorns per year if they have everything? But they can't. And nobody can. Exactly because it's, like, lemmings run. It's pretty much lemmings run because it's much more about, like, not, like, that some lemmings are smarter, but some lemmings, they have better stamina, and some lemmings are lucky.

And then, like, you just, like, maybe, like, people like a CEO, they just have, like, some, like, they're staying, maybe, like, at some milestone and looking, like, which lemmings survived, like, first mile. And then they just bet on this, like, the strongest and luckiest lemmings. But at the beginning of the journey, like, it's very difficult to understand, like, what lemmings would win because, like, 90% about luck.

Being in Imperial, what should students work on?

Huh? What should students work on? I think it's even... I think we established McKinsey. I think it's not even significant, like, what you're doing as a student because then you will change your occupation many times, especially nowadays. You should expect that you will change your profession many times. That means that it's absolutely, like, you may do everything what you want at this stage and then be ready to change your occupation, like, many times.

And I still don't know, like, what I will do, like, for the next 20 or 30 years. Hopefully something. And I think I will change, like, several, like, fields of occupation. And I start, like, I had, like, education in chemistry and have never worked in chemistry. But I think, like, it may become useful at some moment. For example, if I start working with, like, AI, like, a deep AI for health. And you just don't know, like, all those, like, dots and lines, they will be connected with time.

And it means, like, you just must move in. Like, if I have any regrets, it's, like, just about not trying. Because, like, even, like, I even don't think that failures, and we had many of them, like, were really failures. Because then, like, they, like, changed our directions. And they changed our perception of reality. And then we made, like, right steps. And without that right steps would not be done.

For example, even focus on Flow was a result of calamity that in 2014, like, a Crimea war started. And, like, economy in the region crashed. And at the time, we had product projects, which were more focused on the Russian-speaking market. It's, like, crashed. And we decided that we need to, like, just, like, to bring all our resources, like, to one place and focus on something international. And we made a bet on Flow.

But without this kind of, like, Crimea war, and without this crash of our old business, probably we would not be so focused on Flow. And probably Flow would not happen. And then, like, was it, like, fail at the time? Probably. But was it was inevitable requirement, like, to be very successful then? Yeah. And then, like, 25 years after, you will think about your step very differently. Because all dots will be connected to lines.

And everything will get some sense. And because of my advice, like, do everything, like, you feel right. Because, like, it's, like, just journey make sense. And everything will change.

I think for my last question, I wanted to ask you the consensus, if you'd like, or the general advice is don't work with family. And you work with your brother and your wife. So what do you advise people?

And tomorrow will be 20 years of my marriage, of my wedding.

Congrats.

It's, like, this marriage, even, like, and we have worked together for 20 years. And, like, our marriage survived that. And I think, like, there are no universal advice. But it's really significant, probably, to divide fields of responsibility. Because most of all problems happen when these fields of responsibility are not divided. And people are fighting because there is no, like, a real, like, kind of hierarchy and authority when you work with family.

And we have, like, we divided our responsibilities, like, different projects with my brother and my wife. She's not intervening, like, in my decisions at work. And I don't have any say at home. Like, and because of that, like, and because of that, like, it's balanced.

Super. Who has questions, guys? Who has a microphone?

Thank you. Hello. Can you hear me?

Yeah.

Thank you again for this talk. I was looking forward to it a lot. I wanted to ask, Dmitry, you recently spoke about how people can get things wrong in terms of building for consumers, or specifically B2C, especially having a mindset of, like, SaaS, for example. Where do founders, and I know you mentioned how investors may also see consumers a bit wrong. Where do they go wrong, and how do you pretty much win in the world of consumers or B2C?

I think investors overall don't know how to evaluate the consumer well, most of them, because most of investors, they have grown investing in SaaS, and they're applying, like, a principle they learned in SaaS, like, trying to evaluate the consumer. And of course, they're making, like, a conclusion that, oh, because retention is low, it's shitty. But again, they just don't know how to evaluate, like, what's right and what's wrong, because the nature of retention in consumer is very different.

Because in consumer, like, people, like, come and go, they stop subscription for Netflix, they start subscription for Netflix, they reinstall it. But if you just apply, like, principle of SaaS, that it should be, like, 40% of retention and low churn, then you would inevitably conclude that each consumer business is, like, kind of like a bad business. And, like, there are, like, other methods should be applied to analyze consumer businesses, for example, not, like, retention and subscription, but net revenue, retention, and cost of acquisition, total addressable market, very significant, like, very, very different analysis in case of SaaS.

And, like, few people may do that right. For example, our investors in General Atlantic, they are among, like, very few investors who really deeply understand consumer, because they invested in companies like Duolingo or Chess.com and other companies, and they understand that deeply, but it's very rare to see understanding of consumer business model as a market, like, among investors, especially at a late stage.

Other questions? I had a follow-up. Go ahead.

Thank you. My other question was, especially in the health sector, how do you overcome or build solutions that have to kind of transcend both culture and borders of different countries, especially with different health systems at each one? How do you, how did Flow, when you were building it, how did you overcome that?

Yeah, I mentioned, like, Glow of Max Levchin, and I think, like, one of the reasons why Glow didn't become more successful is that, like, design of this product was done by liberal, ultra-liberal people from San Francisco, for ultra-liberal people from San Francisco, like, it's, like, bright colors, like, this very bold image, etc., but it was not good even, like, for Mississippi State, not saying about, like, the world, and when you're building, like, for mass market worldwide, at least you should be very moderate and neutral.

Your product should not be bold because it would resonate maybe with your friends, but not with users worldwide, and then, like, localization, real localization is necessary. For example, one of our learnings in our pregnancy mode, we have a comparison of, like, a fetus, a baby with, like, some fruits, and once we figure out that people in different parts of the world, for example, they may not know what's blueberry because for people in Africa or in Brazil, like, a blueberry is something very exotic, and they just don't understand what it is, and sometimes the health systems are different, but design should be very, very, very simple and very neutral, or, like, why Chinese products very rarely work for the West despite, like, they're probably, like, very, very good product people because this, like, Chinese design is so far from Western standards, and they even can't understand how to rebuild this stuff, like, to match it, and because of that, it's, like, almost always, like, fail out of China.

Like, and Western products always fail in China, but because of other reasons, because, like, the moment you have some success in China, like, government would kill you, and it happens with all companies with traction in China.

Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you. Thank you so much.

Privyet, Dmitry. Just two quick questions. The first one was about the experiments. How do you go about building those hypotheses on what experiments to even test in the first place, especially when you said there are about 300 experiments that are coming up? How do you go about thinking that these are the experiments that need to happen? And then the second question was about your family. It was interesting to know that you're working with your family. I was just wondering, how did the VCs perceive that?

For VCs, it was not significant altogether because we're, like, we're not, like, sharing, like, kind of, like, executive positions. Like, my brother, he's co-founder, but he's board director. He doesn't have an executive position. And my wife, she doesn't have an executive position. She's my kind of executive assistant. And because of that, it's okay. There's no, like, any conflict here. And the first question was about?

The experiments and how do you come up with those hypotheses.

It's different. Like, sometimes it's just intuition of product people. Sometimes, like, result of user research. Sometimes we make, like, just, like, I see that something works from competitors. Sometimes something may work in one part of product and we may try another part of product. Sometimes it's just, like, a wild bet. But, like, success rate of our experiments is between 10 to 20%. It means, like, in any case, like, 80% of our experiments fail.

It means, like, we may say that most of the job is done, like, pretty much for nothing. Because, like, it's just wasted. Because, like, it's just, like, fail. But ultimately, you need to do something that would work. And it's just inevitable ways that, like, 80% of things will not work.

I think we have time for maybe one or two questions. Yes, ma'am.

Thank you. That was really interesting. My question is about compliance. Obviously, as a health app, you have to be very mindful of that. How do you deal with it and how did you manage that at the early stages of your venture?

I would say that it's completely impossible to be fully compliant at the early stage. Because now we have 15 lawyers inside our company and my legal budget is measured in millions. And it's still, like, always something. But when you're a small company, it's just impossible. Because even in the United States, it's, like, it's different legislation in different states. Then, like, different legislation worldwide.

There are, like, many nuances. And it's legislation about privacy, legislation about, like, medical device compliance, like other types of legislation. It means that the early stage is pretty much impossible. And advice probably starts from one territory because then it would be really difficult to be compliant, like, for many territories. But then, like, you just should accept that you would not be perfect at the early days because even if you start from the United States and you have 50 different states, each state has its own legislation, own, like, procedures.

Like, it's almost impossible. Of course, like, you must do, like, big things right. For example, you shouldn't bring, like, a medical device to the market without FDA approval. But could you make, like, everything perfect from the prospect of your, for example, privacy legislation and regulatory, different aspects of regulatory? When you're small, no. Like, it's just too difficult and too expensive.

Maybe now with GNI tooling, it will be simpler. Like, but we will see because you may probably, like, just, like, generate that. But it's, like, all this, like, compliance stuff, compliance document. But it's also, like, procedures, controls. Like, it delays when you have, like, a health company. Probably you run, like, twice, three times slower than, like, a company which is not in health because you need to check everything.

In our case, we're checking everything, like, by doctors. We're checking everything by lawyers and lawyers of different kinds. Some of them, like, are checking privacy. Some of them, like, are checking compliance. And then, of course, it's inevitably delayed. But, like, when you're a big company, you just must comply fully. When a small company, probably you have some luxury to be under the radar for some time and not to be, like, fully compliant.

And you just can't be, honestly. It's just impossible. It's just too sophisticated, like, and bureaucrats, they're working, like, long days to create, like, new and new legislations. Really difficult.

And last question.

Hi. Thank you for such awesome insights around compliance. I have this question for you. Like, if you were to build a B2B SaaS today in the world of AI and all the stuff around that, how would you choose your core team and build your core team along with how would you gain the initial traction in terms of, like, reaching out to the potential, paying customers, small businesses, and things like that?

I think now it's, like, a really unique opportunity to build companies, like, having, like, a very limited resource. And because I have been at this, like, in business for more than 20 years, I remember, like, times when, like, clouds didn't exist. And even to start your own website, you had to have, like, physically, like, a server in your place. Like, and you had to have, like, people who were on your server.

Like, it means that, like, everything was much more complicated, even to start, like, a simple website. And then, like, clouds, like, simplified everything. And SaaS business model, like, it's pretty much innovation based, like, just on, like, clouds and innovation in business models. Like, it's a combination of subscriptions with clouds, pretty much. It's, like, how, like, SaaS was born. But still, like, it was necessary to have, like, many people to develop a good product.

And now it's a unique opportunity to develop products having very limited resources. But then, like, what's significant at this moment? It's a really good product intuition and a real deep understanding of needs. And it means that, like, just, like, product talent is becoming much more significant, probably, than engineering talent and product intuition. And my advice would be, like, to really deeply understand what you're doing, or at least to bring, like, a person who understands it really well, like, from the point of view of the customer.

And then to have, like, a good product people. And then, like, engineering is becoming, like, easy and easy. And probably in one or two years, it will become, like...

I'm not sure that engineers will not be necessary, but we'll definitely have many unicorns with just one, two, or three engineers, and the products will be sophisticated.

Super. Dmitry, that's been phenomenal. Thank you so much.

Thank you.