The Exchanges

Q Did the new products make your life easier with growth or harder? Because you suddenly have more to sell, different messaging. How did it change your role?

A I think a lot easier to be fair. I think it's actually the biggest mistake that founders do, that they over optimize their product. Um, so it's quite counterintuitive, right? But it's, it's a bit like if, let's say you're Apple, um, there's only so much revenue you can make with an iPhone because not everybody wants or needs an iPhone. But so now, like, if you have a second product, let's say AirPods, you can sell to one client, not just a thousand dollar phone. You can also sell 200 dollar earphones, right? And then if you, if they also need a laptop, you can sell it for another 2000. And I think, you know, like Revolut was built in a similar way. We basically used FreeFX as an acquisition driver. I mean, the value proposition of the product was so strong. They're like, we didn't really have to pay a lot to get people into our product. Essentially what we needed to do is just accelerate already, um, dynamics that are happening. For example, that friends tell their friends or making sure that people don't, you know, get somewhere stuck through the software. Right. Um, and then basically what we did was like a lot of cross selling. So essentially, okay, you're already in, you use the product. Okay. Now you want to buy cryptos. Cool. Three percent margin. Um, you want to buy stocks. You have a spread there. Um, you want, you need business banking at some point. You make some mone…

AI assessment note: “I think a lot easier to be fair.”

Q What was the reason that they weren't, and what did you change?

A Essentially, I didn't really know how to figure it out, so what I did was I sent an email to 4000 people and was like, hey, why did you stop using Revolut, right? And the interesting thing that came back was that the majority of people didn't stop using Revolut, they just didn't have a use case for it yet. Because back in the days it was known as a travel card, and so what we realized is, like, people don't see Revolut as an everyday banking card. So we started changing the way we market the product, you know, position it in the market, but also making sure that there's no technical issues. Like, for example, people try to, uh, verify their identity at night, and then you use a flash, and so suddenly the software can't detect, uh, your identity code. Or, I don't know, you know, in Italy, we had a weird problem. Um, people don't have MasterCards or Visa, so they can't do, and actually online payments, which we used Traditionally to top up your account. But then also like Italy is a bit backwards in terms of the banking system. So you couldn't make international bank transfers from your bank account. So we had to figure out basically how we get users to actually get money onto the account.

AI assessment note: “majority of people didn't stop using Revolut, they just didn't have a use case”

Q What was the reason that they weren't, and what did you change?

A Essentially, I didn't really know how to figure it out, so what I did was I sent an email to 4000 people and was like, hey, why did you stop using Revolut, right? And the interesting thing that came back was that the majority of people didn't stop using Revolut, they just didn't have a use case for it yet. Because back in the days it was known as a travel card, and so what we realized is, like, people don't see Revolut as an everyday banking card. So we started changing the way we market the product, you know, position it in the market, but also making sure that there's no technical issues. Like, for example, people try to, uh, verify their identity at night, and then you use a flash, and so suddenly the software can't detect, uh, your identity code. Or, I don't know, you know, in Italy, we had a weird problem. Um, people don't have MasterCards or Visa, so they can't do, and actually online payments, which we used Traditionally to top up your account. But then also like Italy is a bit backwards in terms of the banking system. So you couldn't make international bank transfers from your bank account. So we had to figure out basically how we get users to actually get money onto the account.

AI assessment note: “they just didn't have a use case for it yet... So we started changing”

Q Did the new products make your life easier with growth or harder? Because you suddenly have more to sell, different messaging. How did it change your role?

A I think a lot easier to be fair. I think it's actually the biggest mistake that founders do, that they over optimize their product. Um, so it's quite counterintuitive, right? But it's, it's a bit like if, let's say you're Apple, um, there's only so much revenue you can make with an iPhone because not everybody wants or needs an iPhone. But so now, like, if you have a second product, let's say AirPods, you can sell to one client, not just a thousand dollar phone. You can also sell 200 dollar earphones, right? And then if you, if they also need a laptop, you can sell it for another 2000. And I think, you know, like Revolut was built in a similar way. We basically used FreeFX as an acquisition driver. I mean, the value proposition of the product was so strong. They're like, we didn't really have to pay a lot to get people into our product. Essentially what we needed to do is just accelerate already, um, dynamics that are happening. For example, that friends tell their friends or making sure that people don't, you know, get somewhere stuck through the software. Right. Um, and then basically what we did was like a lot of cross selling. So essentially, okay, you're already in, you use the product. Okay. Now you want to buy cryptos. Cool. Three percent margin. Um, you want to buy stocks. You have a spread there. Um, you want, you need business banking at some point. You make some mone…

AI assessment note: “I think a lot easier to be fair. I think it's actually the biggest mistake”

Q I want to start on, you know, people love to throw around PLG. It's like, it's been the kind of hot word in our industry for years now. But you said to me before, there's traditional marketing methods that are outdated. And you were brilliant in, when we chatted, like, just dropping these one-liners that I was like, great, thanks for expanding. Why are traditional marketing methods outdated?

A I think it's just like the market is so saturated, oversaturated, um, that it's just, I mean, basically traditional marketing means like it's not scalable, right? Like essentially it only scales with human output. So essentially if I want to have twice as many people coming to my website, I need to hire twice as many people, which becomes, you know, as I, the company grows, it becomes more and more expensive. And then typically teams would get less efficient. So I need even more people. I think, you know, If you look at growth, which is essentially, you know, you, you build software that does marketing for you, uh, it basically scales exponentially. So, um, I need less and less people in the future to get bigger and bigger results.

AI assessment note: “basically traditional marketing means like it's not scalable, right?”

Q Do you think people do referral platforms or programs well today? We see so many. Invite a friend and get 20 pounds. Do people do it well?

A I don't think so. Um, So the, the interesting thing about referrals is that there isn't really a lot of information on the internet about how it really works. And so it took me about nine months to figure out, you know, how to actually make it work, right? So the common knowledge is like, okay, one person needs to invite more than one person, right? And that gives you an exponential growth. So the, the problem is it, it really depends on how fast that invite happens. So for example, if on average I bring three people in, but it takes three years, It's really slow, the growth, uh, um, yeah, the growth of the company. On the other hand, if you, let's say, get three people in every three minutes, you, you grow a lot faster, right? Um, so one thing we've done at Revolut was essentially engineer how fast or how short it takes for new people to invite their peers and their friends, which actually worked really well to a point where we had to cap how many people actually Sign up.

AI assessment note: “I don't think so. Um, So the, the interesting thing about referrals”

Q What are some lessons on how to make YouTube influencers successful? Everyone kind of has this idea, well, I'm going to do the influencer strategy. I don't think many people know how, and I think they just have an idea that I'm just going to pay them a wedge of cash. Others have the affiliate code. Any lessons from that?

A I think growth is actually quite simple. Um, it doesn't really matter which channel you're using, but it's always basic math. I mean, let's say an influencer, you look at like, okay, on average, how many views will they generate, uh, per video? So you can expect roughly like how many people will see it. Then how many people do you think will click through from the video to actually your website or app store? I would, you know, you can take conservative market estimates, like five percent or something, and then essentially how many people actually convert from, you know, checking out your website to actually becoming a user. And after that, it's basically like you have your typical rates, like what's your subscription rate? And so you can calculate roughly like, okay, if I pay that person, like if I want to achieve a CPA of, let's say, 10 dollars, like what's the maximum I can offer the influencer to do this? Um, I mean, that's basically the starting point. And then essentially what you do is you, um, you test three, four in different niches and see basically what works.

AI assessment note: “calculate roughly like, okay, if I pay that person... what's the maximum I can offer”

Q When we chatted, you said that most companies kind of fail at hiring. I did want to touch on this because this is like a universal process that every founder has to go through. Why do you think most founders or companies fail at hiring?

A Probably two things, like A, you don't test people for enough. So, you know, you made it maybe have a conversation. You don't dig deep into, like, let's say they talk about a certain project or, you know, some past incident, but you don't ask follow-up questions, really to understand context, right? Like, did they contribute towards it or did they lead it? Like, can they explain in detail what was the problem or not? So I think, you know, the best founders, they really, like, they really dig deep in understanding, like, okay, what really happened? Do I believe that that person actually Made that impact? Do they understand enough to make that impact? And then the second thing, which is essentially, if you look at sports, for example, um, you know, Michael Jordan, Cristiano Ronaldo, Kobe Bryant, you don't have to motivate these people. Doesn't matter what they will do. They will always be the best. It's just their attitude. It's the, the mindset, right? And I think, you know, like you want to see the same thing in the people you hire. You want people that are hungry, people that are passionate.

AI assessment note: “Probably two things, like A, you don't test people for enough.”

Q How do you do a referral loop that's so effective?

A Couple of optimizations. Um, essentially we thought about like, okay, if you, let's say I'm here at the podcast, right? Um, and I want to, you know, I, I love Revolut. It helps me, you know, when I travel, I don't have to overspend. And so I'm basically like telling you about the product, right? What do you want to achieve is that, uh, while we're talking here, you can actually sign up and make your first payment. For example, I don't know, order something on delivery. And so we optimize the product to basically make sure that Enough people can go through that whole journey in less than five minutes. And so essentially, because it's so easy, so quick, we had so many people that just went to their friends, to their family, strangers on the street that just basically get people into Revolut.

AI assessment note: “we optimize the product to basically make sure that Enough people can go through that whole journey”

Q I want to start on, you know, people love to throw around PLG. It's like, it's been the kind of hot word in our industry for years now. But you said to me before, there's traditional marketing methods that are outdated. And you were brilliant in, when we chatted, like, just dropping these one-liners that I was like, great, thanks for expanding. Why are traditional marketing methods outdated?

A I think it's just like the market is so saturated, oversaturated, um, that it's just, I mean, basically traditional marketing means like it's not scalable, right? Like essentially it only scales with human output. So essentially if I want to have twice as many people coming to my website, I need to hire twice as many people, which becomes, you know, as I, the company grows, it becomes more and more expensive. And then typically teams would get less efficient. So I need even more people. I think, you know, If you look at growth, which is essentially, you know, you, you build software that does marketing for you, uh, it basically scales exponentially. So, um, I need less and less people in the future to get bigger and bigger results.

AI assessment note: “traditional marketing means like it's not scalable, right? Like essentially it only scales with human output”

Q Do you think people do referral platforms or programs well today? We see so many. Invite a friend and get 20 pounds. Do people do it well?

A I don't think so. Um, So the, the interesting thing about referrals is that there isn't really a lot of information on the internet about how it really works. And so it took me about nine months to figure out, you know, how to actually make it work, right? So the common knowledge is like, okay, one person needs to invite more than one person, right? And that gives you an exponential growth. So the, the problem is it, it really depends on how fast that invite happens. So for example, if on average I bring three people in, but it takes three years, It's really slow, the growth, uh, um, yeah, the growth of the company. On the other hand, if you, let's say, get three people in every three minutes, you, you grow a lot faster, right? Um, so one thing we've done at Revolut was essentially engineer how fast or how short it takes for new people to invite their peers and their friends, which actually worked really well to a point where we had to cap how many people actually Sign up.

AI assessment note: “I don't think so. Um, So the, the interesting thing about referrals”

Q What are some lessons on how to make YouTube influencers successful? Everyone kind of has this idea, well, I'm going to do the influencer strategy. I don't think many people know how, and I think they just have an idea that I'm just going to pay them a wedge of cash. Others have the affiliate code. Any lessons from that?

A I think growth is actually quite simple. Um, it doesn't really matter which channel you're using, but it's always basic math. I mean, let's say an influencer, you look at like, okay, on average, how many views will they generate, uh, per video? So you can expect roughly like how many people will see it. Then how many people do you think will click through from the video to actually your website or app store? I would, you know, you can take conservative market estimates, like five percent or something, and then essentially how many people actually convert from, you know, checking out your website to actually becoming a user. And after that, it's basically like you have your typical rates, like what's your subscription rate? And so you can calculate roughly like, okay, if I pay that person, like if I want to achieve a CPA of, let's say, 10 dollars, like what's the maximum I can offer the influencer to do this? Um, I mean, that's basically the starting point. And then essentially what you do is you, um, you test three, four in different niches and see basically what works.

AI assessment note: “it doesn't really matter which channel you're using, but it's always basic math.”

Q When we chatted, you said that most companies kind of fail at hiring. I did want to touch on this because this is like a universal process that every founder has to go through. Why do you think most founders or companies fail at hiring?

A Probably two things, like A, you don't test people for enough. So, you know, you made it maybe have a conversation. You don't dig deep into, like, let's say they talk about a certain project or, you know, some past incident, but you don't ask follow-up questions, really to understand context, right? Like, did they contribute towards it or did they lead it? Like, can they explain in detail what was the problem or not? So I think, you know, the best founders, they really, like, they really dig deep in understanding, like, okay, what really happened? Do I believe that that person actually Made that impact? Do they understand enough to make that impact? And then the second thing, which is essentially, if you look at sports, for example, um, you know, Michael Jordan, Cristiano Ronaldo, Kobe Bryant, you don't have to motivate these people. Doesn't matter what they will do. They will always be the best. It's just their attitude. It's the, the mindset, right? And I think, you know, like you want to see the same thing in the people you hire. You want people that are hungry, people that are passionate.

AI assessment note: “Probably two things, like A, you don't test people for enough.”

Q What breaks with scale that fast? When you're scaling that fast where it's like, you know, metal cards are going out faster than you can make them literally. What breaks in a company?

A Everything. Everything. Um, you kind of had to re, rebuild everything in-house every few months. Like, I mean, for example, um, I went there when we were like, 80 people. I left when it was two and a half thousand. That's two and a half years. Like, my team went from two people to over a 150 in, like, nine months. Like, everything breaks, you know, like, the way you communicate doesn't work anymore, the way you, I mean, I don't know, like, we had, for example, we ran out, we couldn't upgrade to a bigger database anymore, right? Like, basically, Google didn't have bigger databases, so we had to split the databases and sync them somehow. Um, Yeah, I think it's just like, I can't remember, like, was it seven, eight hundred million transactions that we processed in one single month at the end before I left? And we were just growing every month with hundreds of new, hundreds of millions of new transactions. You know, it was just mad. When we did referrals, for example, when we launched it, we went from 15,000 new users a day Uh, and we only launched it in one market.

AI assessment note: “Everything. Everything. Um, you kind of had to re, rebuild everything in-house”

Q How important do you think it is to have the reviews? You said there about design reviews, but in growth, maybe in particular, how do you think about growth reviews syncing as a team on progress? How often? What should be involved?

A I think in the end, like, I think it's, it's, you need a feedback cycle, right? You need to kind of see like, okay, I try A, is it good? Is it bad? What's good? What's bad? What could I improve? I think the more frequent you have these things, The better because essentially you learn faster. I mean, it's kind of like imagine like every cycle, every time you get feedback is a cycle. Um, every time you basically complete a cycle, you learn. So the more cycles you have, the faster you learn. And so often what managers do is basically a performance review every three months, right? It's not okay. Like basically that means like it's very slow that you develop because if you do something wrong, if you don't think the right way, you know, if you focus on the wrong things, like you get feedback after three months. So I think probably doing it daily is too extreme, but like, I think once a week was kind of like what we found optimal to move really fast, but equally like. For the whole team? Yeah, so we had like in every function we had once a week an update. And so for example.

AI assessment note: “once a week was kind of like what we found optimal to move really fast”

Q How do you do a referral loop that's so effective?

A Couple of optimizations. Um, essentially we thought about like, okay, if you, let's say I'm here at the podcast, right? Um, and I want to, you know, I, I love Revolut. It helps me, you know, when I travel, I don't have to overspend. And so I'm basically like telling you about the product, right? What do you want to achieve is that, uh, while we're talking here, you can actually sign up and make your first payment. For example, I don't know, order something on delivery. And so we optimize the product to basically make sure that Enough people can go through that whole journey in less than five minutes. And so essentially, because it's so easy, so quick, we had so many people that just went to their friends, to their family, strangers on the street that just basically get people into Revolut.

AI assessment note: “we optimize the product to basically make sure that Enough people can go through that whole journey in less than five minutes.”

Q Final one for you. When you look across the landscape today, which company growth strategy have you been most impressed by?

A The most impressive thing I've seen was Nubank. Like, I mean, they have like a 90% retention rate, like over 24 months. It's, it's ridiculous. It's like they, they just figured out something in the Brazilian market that just, yeah, that just, I don't know, it's magic. It's basically like, from what I understood is, um, in Brazil, because of the financial crisis, They made it illegal, um, to not allow installment payments when you pay for, I don't know, let's say a laptop or food, because a lot of people are so poor that they wouldn't be able to afford it otherwise. So essentially everybody has to offer them A. B. It was illegal to put interest on it. So obviously everybody chose to pay in installments because why not? There's no downside. What Nubank did was basically saying, okay, we're offering a credit card. So you can pay with Nubank. You go to the merchant, split it into free payments, but then in Nubank, you can pay it in one go, and you get a five percent discount. And then they took the money to basically lend it to others. Their product, I mean, it's, it's so sticky, it's, it's insane, you know. They didn't have competition. I think they have like a hundred million active users nowadays.

AI assessment note: “The most impressive thing I've seen was Nubank.”

Q We've mentioned data so many times, so final topic before we go quick fire, but you've said before that there's conventional wisdom around data strategy. What is the conventional wisdom, and how do you think about it?

A I think typically companies use, um, event-driven, uh, data infrastructure. Mixpanel, amplitude, didn't really change for like probably 1520 years, um, which usually creates most of the issues in a company. Um, I actually think that if you look at Uber, Airbnb, Meta, they don't use events, or at least not in the early days. Um, we didn't use events at Revolut. I think the problem is often that engineering doesn't really have access to data. It doesn't use data much. So they may be a bit scared of the topic. Um, then the traditionally data people weren't very technical. It was usually companies that had a data team had data analysts, not data scientists or the data engineering. And then essentially like product managers or business people wanted certain insights. So people went to Mixpanel, went to Amplitude, because it was easy, right? Just one line of code. I think then over the years where people, or some companies realized is that actually you don't get a lot of insights, because, you know, you have some payment data in Stripe, right? Um, some people churn, never come back, right? You have payment issues. So even though it, it, it should, you should generate revenue, you don't, right? Because, I don't know, some weird authorization issue in, I don't know, Brazil. Um, you have different campaigns, right? Um, that, that cost you different money, right? And so you want to under…

AI assessment note: “typically companies use, um, event-driven, uh, data infrastructure. Mixpanel, amplitude”

Q Again, I'm really using this for myself. Like, what are the signs that someone is learning quickly, especially when you have a preference for young people where there's just less data points?

A You can use benchmarks, so for example, unis, top unis. If you go to LSE, Stanford, Harvard, you're probably already, you're pre-vetted, so you're quite smart, you're probably quite ambitious. So if you did the uni like in minimum term, plus, I don't know, worked at a tier one company, you probably fit the bill. So that's one way to attract these people. Um, second thing is, let's say, if you're looking for someone more senior, like a product director, if you get promoted in two, three years to product director, You probably are, there's something different about you, right? Because other people, when, when somebody works as a product manager for, let's say, 10 years, you ask yourself, okay, why are they not a director, right? Why are they not leading a team? Um, so we use these proxies to basically guess whether, you know, it's the right fit or not. Um, also like, for example, um, typically in growth, we didn't look for people that studied marketing.

AI assessment note: “You can use benchmarks, so for example, unis, top unis.”

Q What breaks with scale that fast? When you're scaling that fast where it's like, you know, metal cards are going out faster than you can make them literally. What breaks in a company?

A Everything. Everything. Um, you kind of had to re, rebuild everything in-house every few months. Like, I mean, for example, um, I went there when we were like, 80 people. I left when it was two and a half thousand. That's two and a half years. Like, my team went from two people to over a 150 in, like, nine months. Like, everything breaks, you know, like, the way you communicate doesn't work anymore, the way you, I mean, I don't know, like, we had, for example, we ran out, we couldn't upgrade to a bigger database anymore, right? Like, basically, Google didn't have bigger databases, so we had to split the databases and sync them somehow. Um, Yeah, I think it's just like, I can't remember, like, was it seven, eight hundred million transactions that we processed in one single month at the end before I left? And we were just growing every month with hundreds of new, hundreds of millions of new transactions. You know, it was just mad. When we did referrals, for example, when we launched it, we went from 15,000 new users a day Uh, and we only launched it in one market.

AI assessment note: “Everything... we couldn't upgrade to a bigger database anymore... we had to split the databases”

Q That is a change of thought. I wasn't expecting that much. Can I ask you, what are the biggest lessons from being an underdog? You know, you advise Kittel today. It's against the 800 pound gorilla that is Canva, which has an 800 pound gorilla, which is Adobe. What are the biggest lessons in that experience?

A I mean, if you're a big company, you can't, like, the bigger you get, the easier you get disrupted, right? Um, because essentially, Like, let's say a market that's worth ten million is nothing for a company like Convo that makes, I don't know, I think nowadays three billion, something like that. So, you know, like, they don't even go after that market because it's just not big enough, right? So, but it also allows you to basically disrupt from that market and a small, slightly smaller, a bigger market and then a slightly bigger market. I mean, in the end of the day, like, you don't have, you have high conviction that there's need for a software like this, Because you already have proof points. You also have a lot of feedback around, you know, what's not so good with the tools, which allows you to navigate through it. But equally, like it's obviously very hard because you have a huge competitor that is quite strong at what you do. Um, I think like probably the easy way to see it is you want to go in a market where you don't have competition. So if you think about Revolut, um, against Monzo, right? Like the markets we grew the fastest, Wasn't, wasn't the UK. It was Ireland. It was Romania. It was, you know, Poland. So like while we still competed here and there with, you know, different banks, we went in a lot of markets where we just had didn't have competition, which allowed us…

AI assessment note: “you want to go in a market where you don't have competition.”

Q Is there anything that startup founders can learn from that?

A I think in the early days, right, people build a two-sided marketplace where essentially you had to control or grow both supply and demand. I think what a lot of companies nowadays do, they control both supply and demand. So essentially it's a lot easier because you can just, you know, rather than having to convince people, you know, to basically use your software and be super engaged, you can just pay someone for it. They generate content for you, and as you build more content, you generate more traffic, Which then generates you more revenue. So I think like, uh, ultimately, like if I think what's nowadays happening is that because people understand kind of like the tactics to grow their channels, whether it's SEO referrals, um, you know, paid, like they execute a lot faster and you basically like usually bring things first in-house because you have more control. And then as you grow, you may maybe, you know, Shifted more towards user generated, so basically you cut your cost.

AI assessment note: “bring things first in-house because you have more control. And then as you grow”

Q What are your biggest lessons on how to train people effectively? It's hard to do well.

A It sounds very counterintuitive, but essentially spending more time with them. Um, so I have like, uh, Oleg, he's not, he's a good friend nowadays. Uh, he was, uh, the lead product designer at my team at Revolut. Very creative designer, very unstructured, unfocused. And so Revolut expected you like every week you need to show progress, right? So every week we have a design sync with Nick and you just had to show like, hey, we're making progress, right? Like that's basically what you work on the whole week. And Oleg basically like he, he spent so much time just optimizing one screen, right? So basically we had 20, 30 variations of one screen, but we didn't have like a user journey. So When we had the review, um, essentially there wasn't much progress to be seen, right? We didn't really understand how to, you know, how the solution would solve the problem of the customer. And it took him like, I mean, in the second, then the second week happened, right? Same problem. And I was like, okay, this is not gonna work. You know, if, if, you know, I need new user flows in like maximum two, three weeks, because otherwise like engineering doesn't have anything to do.

AI assessment note: “essentially spending more time with them.”

Q Again, I'm really using this for myself. Like, what are the signs that someone is learning quickly, especially when you have a preference for young people where there's just less data points?

A You can use benchmarks, so for example, unis, top unis. If you go to LSE, Stanford, Harvard, you're probably already, you're pre-vetted, so you're quite smart, you're probably quite ambitious. So if you did the uni like in minimum term, plus, I don't know, worked at a tier one company, you probably fit the bill. So that's one way to attract these people. Um, second thing is, let's say, if you're looking for someone more senior, like a product director, if you get promoted in two, three years to product director, You probably are, there's something different about you, right? Because other people, when, when somebody works as a product manager for, let's say, 10 years, you ask yourself, okay, why are they not a director, right? Why are they not leading a team? Um, so we use these proxies to basically guess whether, you know, it's the right fit or not. Um, also like, for example, um, typically in growth, we didn't look for people that studied marketing.

AI assessment note: “we use these proxies to basically guess whether, you know, it's the right fit”

Q Is there anything that startup founders can learn from that?

A I think in the early days, right, people build a two-sided marketplace where essentially you had to control or grow both supply and demand. I think what a lot of companies nowadays do, they control both supply and demand. So essentially it's a lot easier because you can just, you know, rather than having to convince people, you know, to basically use your software and be super engaged, you can just pay someone for it. They generate content for you, and as you build more content, you generate more traffic, Which then generates you more revenue. So I think like, uh, ultimately, like if I think what's nowadays happening is that because people understand kind of like the tactics to grow their channels, whether it's SEO referrals, um, you know, paid, like they execute a lot faster and you basically like usually bring things first in-house because you have more control. And then as you grow, you may maybe, you know, Shifted more towards user generated, so basically you cut your cost.

AI assessment note: “bring things first in-house because you have more control. And then as you grow”

Q How important do you think it is to have the reviews? You said there about design reviews, but in growth, maybe in particular, how do you think about growth reviews syncing as a team on progress? How often? What should be involved?

A I think in the end, like, I think it's, it's, you need a feedback cycle, right? You need to kind of see like, okay, I try A, is it good? Is it bad? What's good? What's bad? What could I improve? I think the more frequent you have these things, The better because essentially you learn faster. I mean, it's kind of like imagine like every cycle, every time you get feedback is a cycle. Um, every time you basically complete a cycle, you learn. So the more cycles you have, the faster you learn. And so often what managers do is basically a performance review every three months, right? It's not okay. Like basically that means like it's very slow that you develop because if you do something wrong, if you don't think the right way, you know, if you focus on the wrong things, like you get feedback after three months. So I think probably doing it daily is too extreme, but like, I think once a week was kind of like what we found optimal to move really fast, but equally like. For the whole team? Yeah, so we had like in every function we had once a week an update. And so for example.

AI assessment note: “once a week was kind of like what we found optimal to move really fast”

Q What are your biggest lessons on how to train people effectively? It's hard to do well.

A It sounds very counterintuitive, but essentially spending more time with them. Um, so I have like, uh, Oleg, he's not, he's a good friend nowadays. Uh, he was, uh, the lead product designer at my team at Revolut. Very creative designer, very unstructured, unfocused. And so Revolut expected you like every week you need to show progress, right? So every week we have a design sync with Nick and you just had to show like, hey, we're making progress, right? Like that's basically what you work on the whole week. And Oleg basically like he, he spent so much time just optimizing one screen, right? So basically we had 20, 30 variations of one screen, but we didn't have like a user journey. So When we had the review, um, essentially there wasn't much progress to be seen, right? We didn't really understand how to, you know, how the solution would solve the problem of the customer. And it took him like, I mean, in the second, then the second week happened, right? Same problem. And I was like, okay, this is not gonna work. You know, if, if, you know, I need new user flows in like maximum two, three weeks, because otherwise like engineering doesn't have anything to do.

AI assessment note: “It sounds very counterintuitive, but essentially spending more time with them.”

Q And that being like, because people thought it was travel cards, I don't know if I have a use case for it, maybe I will do in time. How did you change that, but not lose the original customers who liked the travel cards?

A The interesting thing about Revolut is why it was created in the first place. Nick, uh, was a trader, right, like trading millions or hundreds of millions, uh, of currencies. Um, so it's basically like, I don't know, I need to sell a hundred million dollars into euros, right? And he would trade it for free, um, against other big banks. And when he went to Vegas and took some money to play poker, um, they took away around 10 to 15% because of the currency exchange. And because he was like, look, I mean, I trade hundreds of millions of dollars without paying any fees. He was like, hey, maybe there's an opportunity, right? So he was a bit pissed about, That banks basically ripped him off, um, and then started Revolut. And I think, you know, like, this was kind of like, for us, it was kind of like the, the engine that drove the company so fast, because we, we were always based on a multi-currency product, which allowed us to add cryptocurrencies, uh, commodities, but also, like, create new innovative products, like, for example, you're spending in pounds, but saving in euros.

AI assessment note: “The interesting thing about Revolut is why it was created in the first place.”

Q And that being like, because people thought it was travel cards, I don't know if I have a use case for it, maybe I will do in time. How did you change that, but not lose the original customers who liked the travel cards?

A The interesting thing about Revolut is why it was created in the first place. Nick, uh, was a trader, right, like trading millions or hundreds of millions, uh, of currencies. Um, so it's basically like, I don't know, I need to sell a hundred million dollars into euros, right? And he would trade it for free, um, against other big banks. And when he went to Vegas and took some money to play poker, um, they took away around 10 to 15% because of the currency exchange. And because he was like, look, I mean, I trade hundreds of millions of dollars without paying any fees. He was like, hey, maybe there's an opportunity, right? So he was a bit pissed about, That banks basically ripped him off, um, and then started Revolut. And I think, you know, like, this was kind of like, for us, it was kind of like the, the engine that drove the company so fast, because we, we were always based on a multi-currency product, which allowed us to add cryptocurrencies, uh, commodities, but also, like, create new innovative products, like, for example, you're spending in pounds, but saving in euros.

AI assessment note: “The interesting thing about Revolut is why it was created in the first place.”