The Exchanges

Q So you have, you know, King, as we said there, incredibly iterative, data-driven culture, data-driven company to the extreme. What's the biggest difference between King and Revolut?

A I think that's what it is. I remember, you know, so I came from a place at King where we were We were later from a maturity perspective, you know, like, I mean, we had launched many games and so on. The business was doing fantastic. And we were more in an optimization place where I remember at some point we were running like maybe a thousand AP tests in parallel and so on. It's like, like it was, it was a machine, you know, the, the machine was really optimizing everything. And I remember when I arrived at Revolut, there was not much AP testing. And for me, I was like shocked at it. I was like, what? I mean, like the, you got to do that. And then what I realized is that You know, there's a good reason as well why you don't want to do A-B testing. The first thing, um, product, I mean, uh, product managers, they, they, uh, they do A-B testing to get an idea of what the right product is. If you're a good product manager, you don't need an A-B test. You should be able to know what is the right product. The second reason is because at Revolut, we wanted to do, like, 10 X. Everything is 10 X and so on. You don't need an A-B test to measure a 10 X. You know, you do it, you see it goes up, it's good, you know. And the third reason is that A-B testing increases your type to market. Because if you A-B test, you first need to do your A-B test, then you see which one wins, and then you lau…

AI assessment note: “we were more in an optimization place... at Revolut, there was not much AP testing”

Q So if testing and speed of iteration and testing was something that you did very well, what was something that King really messed up?

A It's definitely the most data-driven organization I've ever been in, um, in the sense that really data was driving the organization, not just, so if you take the thing literally, The company was driven by data. I think it was so powerful that, uh, and worked so well that at some point we just went way too far in the sense that, for example, when we were launching games, sometimes we were probably looking more at the data than the game itself. You know, it's like you can have meetings and then you just look at the numbers and you're like, uh, you, is that wrong? Well, I mean, that's the big difference, for example, with Revolut. I think we're a lot more product driven. Which means we spend more time looking at the product than looking at the data.

AI assessment note: “we just went way too far in the sense that... looking more at the data”

Q So you have, you know, King, as we said there, incredibly iterative, data-driven culture, data-driven company to the extreme. What's the biggest difference between King and Revolut?

A I think that's what it is. I remember, you know, so I came from a place at King where we were We were later from a maturity perspective, you know, like, I mean, we had launched many games and so on. The business was doing fantastic. And we were more in an optimization place where I remember at some point we were running like maybe a thousand AP tests in parallel and so on. It's like, like it was, it was a machine, you know, the, the machine was really optimizing everything. And I remember when I arrived at Revolut, there was not much AP testing. And for me, I was like shocked at it. I was like, what? I mean, like the, you got to do that. And then what I realized is that You know, there's a good reason as well why you don't want to do A-B testing. The first thing, um, product, I mean, uh, product managers, they, they, uh, they do A-B testing to get an idea of what the right product is. If you're a good product manager, you don't need an A-B test. You should be able to know what is the right product. The second reason is because at Revolut, we wanted to do, like, 10 X. Everything is 10 X and so on. You don't need an A-B test to measure a 10 X. You know, you do it, you see it goes up, it's good, you know. And the third reason is that A-B testing increases your type to market. Because if you A-B test, you first need to do your A-B test, then you see which one wins, and then you lau…

AI assessment note: “at King where we were... optimizing everything. And I remember when I arrived at Revolut”

Q But, but the hard thing is, do you know whether it's being on a billboard or being on an F-One car that works? Because the attribution is limited. How do you know it works?

A So, because we, I, um, uh, you have multiple ways to look, to look at it. So what we do in airport, which is pretty cool, is that, so we do the advertising on jet bridges and so on, you know, if you, if you've been to Rome airport and so on, for example, we've launched something really cool there. But next to it, we put vending machines. So you have vending machines in the airport to collect your Revolut card. So you just go on the vending machine, you press the button, and you just get the card, you know. So it's anonymous, I mean, unlinked card, you just take it, you scan the, the, the, the QR code, creates the account, and then the, the card is automatically linked. And that, then, I can really link all my activities and get a huge ROI from what I do.

AI assessment note: “next to it, we put vending machines... then, I can really link all my activities”

Q People say the most important thing for Revolut to prove out is its ability to be the primary account, and the percent of primary accounts is still low. How do you think about focusing on that versus being everywhere to everyone?

A The strategy at Revolut, it's, it's very simple, you know. Obviously, we want people to use us as a primary bank, but we don't need to do, we don't need everyone to do it now, you know. It's a journey, you know, that's, that's, that's the big difference with, let's say, any other banks. When you decide to move, uh, from HSBC to Barclays, well, you make, uh, You make a conscious decision that I'm moving there, while Revolut, we're more, we're like a snack, you know? So, so you, you want to take a bite, you know, so you try Revolut, and then maybe because you go on holiday, you know, and you need a car to go on holiday, ok, fine. Then you come back from holiday and say, oh, well, you know, maybe I want to do a bit of trading and stuff, you know, I'll buy some ETFs and so on, so then you buy some ETFs. So, oh, then, I mean, there's a bit of savings, ok, maybe I can put a bit of savings, and then, but I'm going to use it every, and so on. And then, And then you go little by little, and then you reach a point, maybe, I mean, for some people it can take six months, for some people it can take three years, and so on, and you say, you know what, I actually have everything I need here, you know, and then I'm moving there, you know. So, I'm not saying, I mean.

AI assessment note: “we want people to use us as a primary bank, but we don't need”

Q People say the most important thing for Revolut to prove out is its ability to be the primary account, and the percent of primary accounts is still low. How do you think about focusing on that versus being everywhere to everyone?

A The strategy at Revolut, it's, it's very simple, you know. Obviously, we want people to use us as a primary bank, but we don't need to do, we don't need everyone to do it now, you know. It's a journey, you know, that's, that's, that's the big difference with, let's say, any other banks. When you decide to move, uh, from HSBC to Barclays, well, you make, uh, You make a conscious decision that I'm moving there, while Revolut, we're more, we're like a snack, you know? So, so you, you want to take a bite, you know, so you try Revolut, and then maybe because you go on holiday, you know, and you need a car to go on holiday, ok, fine. Then you come back from holiday and say, oh, well, you know, maybe I want to do a bit of trading and stuff, you know, I'll buy some ETFs and so on, so then you buy some ETFs. So, oh, then, I mean, there's a bit of savings, ok, maybe I can put a bit of savings, and then, but I'm going to use it every, and so on. And then, And then you go little by little, and then you reach a point, maybe, I mean, for some people it can take six months, for some people it can take three years, and so on, and you say, you know what, I actually have everything I need here, you know, and then I'm moving there, you know. So, I'm not saying, I mean.

AI assessment note: “we want people to use us as a primary bank, but we don't need”

Q So if testing and speed of iteration and testing was something that you did very well, what was something that King really messed up?

A It's definitely the most data-driven organization I've ever been in, um, in the sense that really data was driving the organization, not just, so if you take the thing literally, The company was driven by data. I think it was so powerful that, uh, and worked so well that at some point we just went way too far in the sense that, for example, when we were launching games, sometimes we were probably looking more at the data than the game itself. You know, it's like you can have meetings and then you just look at the numbers and you're like, uh, you, is that wrong? Well, I mean, that's the big difference, for example, with Revolut. I think we're a lot more product driven. Which means we spend more time looking at the product than looking at the data.

AI assessment note: “we were probably looking more at the data than the game itself”

Q Your product velocity is insane. How do you have such high product velocity?

A I think because we are able to, um, so the organization is very, very flat, right? Um, and, uh, we are making sure that, um, every team is, um, as self-sufficient as possible, you know? And that's what really builds the agility, you know? If you're, uh, if you're self-sufficient in building what you have to build, you just go, you know, with, uh, with, with it. And we run more quarterly cycles and so on, and we have A very high, a very high pace, but because there's no interdependencies between the different things that we build, at the end of the day, we are splitting the companies as many startups as possible, you know. I know, like, big organizations, we say, oh, you know, those are silos, people are not collaborating, et cetera, et cetera. You know, you need to always find the right balance between coordinating things while giving the right amount of focus, and I think that's what we We are, we are doing pretty well at Revolut where every team has a huge amount of autonomy and a focus to really go very fast in what they do.

AI assessment note: “organization is very, very flat”

Q You mentioned kind of growth engine when you started that on, like, what enables Revolut to continue growing so fast. What actually is a growth engine? I know it sounds strange, but we often hear it What does that actually mean?

A So a growth engine is where, um, uh, where you have multiple components that all play together. I'll give you, I'll give you a very simple example. I look at it quite, um, uh, the famous full funnel, you know, that, that everyone says. So you have the upper funnel, the mid funnel, and the lower funnel. And the growth engine is when you manage to get those three elements working together. You know. So at the end of the day, the growth engine is when you make sure that you have the lower funnel that works like a fishing net, grabs all the conversions that ultimately mid and upper funnel are feeding you with, you know. So, and that's where all of that works in pers-, in perfect synchronization, so that at the end of the day, it's all been grabbed by the, by the, by the lower funnel. So that is, that is the engine. When then you find the right balance, Between those different elements that optimize the, the overall, uh, the overall, the overall thing.

AI assessment note: “a growth engine is where you have multiple components that all play together”

Q You said the brand's getting bigger. It definitely is. Does CAC get lower over time as brand amplifies word of mouth increases? Or does it get higher as you saturate target market and you move to the fringes of your ICP?

A Well, I think it's actually both. So, so, because it's exactly that. It's exactly that. It's really those two, you have those two factors playing against each other, and it's about which one is stronger than the other. And then, um, and then, unfortunately, it really depends on the market, you know. So some markets, the brand is becoming so big that actually you start getting more organic, and it's becoming like super organic, and then the, the, the, the, the growth energy is feeding itself, And then it just goes. While others, and it's linked as well to the amount of competition that you have in the market, where actually because you have, um, a diminishing return curve, where it suddenly gets more expensive to get the next one, and the next one, and the next one. So, I think it really depends on the, on the competitiveness of the market.

AI assessment note: “Well, I think it's actually both.”

Q You mentioned kind of growth engine when you started that on, like, what enables Revolut to continue growing so fast. What actually is a growth engine? I know it sounds strange, but we often hear it What does that actually mean?

A So a growth engine is where, um, uh, where you have multiple components that all play together. I'll give you, I'll give you a very simple example. I look at it quite, um, uh, the famous full funnel, you know, that, that everyone says. So you have the upper funnel, the mid funnel, and the lower funnel. And the growth engine is when you manage to get those three elements working together. You know. So at the end of the day, the growth engine is when you make sure that you have the lower funnel that works like a fishing net, grabs all the conversions that ultimately mid and upper funnel are feeding you with, you know. So, and that's where all of that works in pers-, in perfect synchronization, so that at the end of the day, it's all been grabbed by the, by the, by the lower funnel. So that is, that is the engine. When then you find the right balance, Between those different elements that optimize the, the overall, uh, the overall, the overall thing.

AI assessment note: “a growth engine is where, um, uh, where you have multiple components that all play together.”

Q How have you changed as an operator working with him?

A So, well, I've taken his approach of managing, uh, of managing my teams, you know, what, what we, what we said about... Beating them weakly. Yeah, I mean, like, like the whole problem solving thing, you know, the whole, we're all here to solve problems. I'm not here to hear, I mean, I'm not going in meetings to hear how good you're doing. Which is what most meetings are about in corporate, in the corporate world, you know, you get, you bring the exec so that you do a nice presentation about how good your latest campaign was, and then you move to the next one, and then at the end, I mean, the diary of an exec is just that everyone is like, oh wow, we're amazing, and so on, and that's really not, that is, that's, yeah, of course it's important to see what's working well, but I really want to know where the problems are, and trying, we can find a solution collectively, you know, into, into solving it, you know. I think that's, that's the thing that is, is very, very powerful.

AI assessment note: “I've taken his approach of managing, uh, of managing my teams”

Q Your product velocity is insane. How do you have such high product velocity?

A I think because we are able to, um, so the organization is very, very flat, right? Um, and, uh, we are making sure that, um, every team is, um, as self-sufficient as possible, you know? And that's what really builds the agility, you know? If you're, uh, if you're self-sufficient in building what you have to build, you just go, you know, with, uh, with, with it. And we run more quarterly cycles and so on, and we have A very high, a very high pace, but because there's no interdependencies between the different things that we build, at the end of the day, we are splitting the companies as many startups as possible, you know. I know, like, big organizations, we say, oh, you know, those are silos, people are not collaborating, et cetera, et cetera. You know, you need to always find the right balance between coordinating things while giving the right amount of focus, and I think that's what we We are, we are doing pretty well at Revolut where every team has a huge amount of autonomy and a focus to really go very fast in what they do.

AI assessment note: “the organization is very, very flat... every team is as self-sufficient as possible”

Q How have you seen that change over time? Because when you were doing that at King in the early days versus today, when it's so much more polluted, crowded, expansive, how has it changed?

A Well, it has changed a little bit, but not so much. Fundamentally, it is still the same game. It's still a bidding war, you know, uh, now it's true. It's getting a little bit more complex because you're losing some attribution and so on. So you need to be a little bit smarter in your measurement framework. So But you just need to, I mean, you reinvent yourself every other year, you have to reinvent the whole measurement framework and so on, but at the end of the day, the game is still the same game, you know, it is the same, it's the same beating war, you just have impressions, you need to go to the highest beater, and that's it, you know, that's what the game is.

AI assessment note: “It's getting a little bit more complex because you're losing some attribution and so on.”

Q Now we have many mutual friends, mostly also hailing from your start at King. So I want to start with that. When you look at the insanely successful growth of Candy Crush, what did King do well that was instrumental to the success and growth?

A I think first of all, everyone thinks there was, um, a huge amount of luck. I think the idea that what King did really well, I think they've tested extensively. Everything. So it was a volume game, you know, at some point. And I think they understood, they understood that super quickly saying, you need to try as much as you can. At some point, something will pick up, you know? And I think they did that super, super, super well. And I think then there was this super fast pace. I remember we were launching a game, like, I mean, every quarter, pretty much we were launching a game and not every, all of them were successful, but just like, The pace at which we were producing games, releasing the games, and so on, I think we did that super, super well.

AI assessment note: “they've tested extensively. Everything. So it was a volume game”

Q Can I ask, what's in the mid funnel?

A This is where, I mean, at least the way I, I see, I look at it, it's the place where, You can, um, you can run some level of attribution, but you have a strong halo because of the high reach that you get, and you can develop as well some deeper content. So performance marketing is basically banners and or super short videos, you know, like six seconds and so on. So you don't really build anything in terms of, let's say, product awareness, brand, Uh, consideration like Abitian. So you don't build any of that. So what you say is that you, you do, you do play the full right away. You play the full funnel because you have a big branding impact, but you are revenue generator right away, you know, which is the exact opposite as the formula one thing where there's no revenue. There's nothing, you know, it's like it's pure brand and so on. But, and that's what I call a bit mid funnel. It's activities that help you do a bit of everything. Like suddenly, You do some brand, but at the same time you do conversion and you drive and you drive revenue, you know, so you kind of take it all.

AI assessment note: “It's activities that help you do a bit of everything.”

Q Can I ask, what's in the mid funnel?

A This is where, I mean, at least the way I, I see, I look at it, it's the place where, You can, um, you can run some level of attribution, but you have a strong halo because of the high reach that you get, and you can develop as well some deeper content. So performance marketing is basically banners and or super short videos, you know, like six seconds and so on. So you don't really build anything in terms of, let's say, product awareness, brand, Uh, consideration like Abitian. So you don't build any of that. So what you say is that you, you do, you do play the full right away. You play the full funnel because you have a big branding impact, but you are revenue generator right away, you know, which is the exact opposite as the formula one thing where there's no revenue. There's nothing, you know, it's like it's pure brand and so on. But, and that's what I call a bit mid funnel. It's activities that help you do a bit of everything. Like suddenly, You do some brand, but at the same time you do conversion and you drive and you drive revenue, you know, so you kind of take it all.

AI assessment note: “It's activities that help you do a bit of everything.”

Q What's the hardest thing about your role today, Anton?

A It's a bit what I was saying about the living animal of the growth engine, you know, because, so, I mean, we've built an engine that is working today, but it doesn't mean that it's gonna work forever, you know, so you need to progressively evolve it Pieces by pieces. That's very difficult. You know, when you know something works and you need to keep iterating so that you make sure it keeps working and you keep being one step ahead of everyone, you know, and that's, I think that's difficult. You know, it's a, it's always this idea, you know, of the car getting eaten from the back, you know, and you, and, and, and you, and you realize really late that actually, I mean, you're gone, you know, and I think staying on top of it, even though it's working really well, never become pleasant. About the success we have. We rarely look back at Revolut. Like, um, we, we're never happy about our success, you know. It's very rare we reach our targets, you know. We never reach our targets, you know, because our targets are much greater than what we, than what we're achieving. So I think it's...

AI assessment note: “when you know something works and you need to keep iterating so that you make sure it keeps working”

Q How have you seen that change over time? Because when you were doing that at King in the early days versus today, when it's so much more polluted, crowded, expansive, how has it changed?

A Well, it has changed a little bit, but not so much. Fundamentally, it is still the same game. It's still a bidding war, you know, uh, now it's true. It's getting a little bit more complex because you're losing some attribution and so on. So you need to be a little bit smarter in your measurement framework. So But you just need to, I mean, you reinvent yourself every other year, you have to reinvent the whole measurement framework and so on, but at the end of the day, the game is still the same game, you know, it is the same, it's the same beating war, you just have impressions, you need to go to the highest beater, and that's it, you know, that's what the game is.

AI assessment note: “it has changed a little bit, but not so much. Fundamentally, it is still the same game.”

Q You said the brand's getting bigger. It definitely is. Does CAC get lower over time as brand amplifies word of mouth increases? Or does it get higher as you saturate target market and you move to the fringes of your ICP?

A Well, I think it's actually both. So, so, because it's exactly that. It's exactly that. It's really those two, you have those two factors playing against each other, and it's about which one is stronger than the other. And then, um, and then, unfortunately, it really depends on the market, you know. So some markets, the brand is becoming so big that actually you start getting more organic, and it's becoming like super organic, and then the, the, the, the, the growth energy is feeding itself, And then it just goes. While others, and it's linked as well to the amount of competition that you have in the market, where actually because you have, um, a diminishing return curve, where it suddenly gets more expensive to get the next one, and the next one, and the next one. So, I think it really depends on the, on the competitiveness of the market.

AI assessment note: “Well, I think it's actually both. So, so, because it's exactly that.”

Q But, but the hard thing is, do you know whether it's being on a billboard or being on an F-One car that works? Because the attribution is limited. How do you know it works?

A So, because we, I, um, uh, you have multiple ways to look, to look at it. So what we do in airport, which is pretty cool, is that, so we do the advertising on jet bridges and so on, you know, if you, if you've been to Rome airport and so on, for example, we've launched something really cool there. But next to it, we put vending machines. So you have vending machines in the airport to collect your Revolut card. So you just go on the vending machine, you press the button, and you just get the card, you know. So it's anonymous, I mean, unlinked card, you just take it, you scan the, the, the, the QR code, creates the account, and then the, the card is automatically linked. And that, then, I can really link all my activities and get a huge ROI from what I do.

AI assessment note: “I can really link all my activities and get a huge ROI from what I do”

Q What's the hardest thing about your role today, Anton?

A It's a bit what I was saying about the living animal of the growth engine, you know, because, so, I mean, we've built an engine that is working today, but it doesn't mean that it's gonna work forever, you know, so you need to progressively evolve it Pieces by pieces. That's very difficult. You know, when you know something works and you need to keep iterating so that you make sure it keeps working and you keep being one step ahead of everyone, you know, and that's, I think that's difficult. You know, it's a, it's always this idea, you know, of the car getting eaten from the back, you know, and you, and, and, and you, and you realize really late that actually, I mean, you're gone, you know, and I think staying on top of it, even though it's working really well, never become pleasant. About the success we have. We rarely look back at Revolut. Like, um, we, we're never happy about our success, you know. It's very rare we reach our targets, you know. We never reach our targets, you know, because our targets are much greater than what we, than what we're achieving. So I think it's...

AI assessment note: “you need to progressively evolve it Pieces by pieces. That's very difficult.”

Q How have you changed as an operator working with him?

A So, well, I've taken his approach of managing, uh, of managing my teams, you know, what, what we, what we said about... Beating them weakly. Yeah, I mean, like, like the whole problem solving thing, you know, the whole, we're all here to solve problems. I'm not here to hear, I mean, I'm not going in meetings to hear how good you're doing. Which is what most meetings are about in corporate, in the corporate world, you know, you get, you bring the exec so that you do a nice presentation about how good your latest campaign was, and then you move to the next one, and then at the end, I mean, the diary of an exec is just that everyone is like, oh wow, we're amazing, and so on, and that's really not, that is, that's, yeah, of course it's important to see what's working well, but I really want to know where the problems are, and trying, we can find a solution collectively, you know, into, into solving it, you know. I think that's, that's the thing that is, is very, very powerful.

AI assessment note: “I've taken his approach of managing, uh, of managing my teams”

Q Can we just start there? I spoke to so many people, investors in the business, and they're like, Harry, it's meant to slow down. They're getting faster and better. How?

A I mean, the thing with, with what we build, So we build an engine, which is a combination of many, many things, meaning there's no, because you have two types of growth. You have the growth that are made of, let's say, launches, let's say, marketing campaigns and, uh, like some coup, you know, like you make a big marketing coup, you know, and then you get the spike and so on. There's none of that at Revolut, you know. I mean, if anyone, I mean, you're probably gonna ask me, what is the one thing, you know, that really worked at Revolut? I can't tell you, you know. It's a combination of so many things in the growth engine, and because they all add up, when you look at the growth of Revolut, it's just very steady. And it's steady, and it's accelerating, um, week after week, you know, we just monitor every week, the acceleration keeps on going, just because the engine gets more and more optimized, you know, as we are progressing, and, and, and for now, we definitely don't see the limit, even in our mature markets, you know, which is quite crazy.

AI assessment note: “just because the engine gets more and more optimized, you know, as we are progressing”

Q Can we just start there? I spoke to so many people, investors in the business, and they're like, Harry, it's meant to slow down. They're getting faster and better. How?

A I mean, the thing with, with what we build, So we build an engine, which is a combination of many, many things, meaning there's no, because you have two types of growth. You have the growth that are made of, let's say, launches, let's say, marketing campaigns and, uh, like some coup, you know, like you make a big marketing coup, you know, and then you get the spike and so on. There's none of that at Revolut, you know. I mean, if anyone, I mean, you're probably gonna ask me, what is the one thing, you know, that really worked at Revolut? I can't tell you, you know. It's a combination of so many things in the growth engine, and because they all add up, when you look at the growth of Revolut, it's just very steady. And it's steady, and it's accelerating, um, week after week, you know, we just monitor every week, the acceleration keeps on going, just because the engine gets more and more optimized, you know, as we are progressing, and, and, and for now, we definitely don't see the limit, even in our mature markets, you know, which is quite crazy.

AI assessment note: “It's a combination of so many things in the growth engine, and because they all add up”

Q What do you mean you cracked the TV measurement framework?

A So we were able to have some, some good level of attribution of, between TV and the downloads, and to get, to be able to really know, um, how much we could attribute to our TV activities. And, and, and because we managed to get that, then we were one of the very first companies To actually make, uh, to make that work, which this is a very good example. The other good example, uh, is, uh, iOS, performance marketing on iOS after scan, and you have companies that know how this works, and therefore they can measure it, and therefore they can scale, and you have all the other ones that I mean, are really struggling and so on. At the end of the day, they spend maybe two, three times more on Android versus iOS, you know, which is a nonsense, you know.

AI assessment note: “we were able to have some, some good level of attribution of, between TV and the downloads”

Q What did you try that didn't work as planned and you cut? And what did you learn?

A Well, so, I mean, on the marketing side, um, we've been, uh, uh, there's been some, uh, some campaigns, for example, that we've been running in some, in some countries where we've been using some, uh, creatives that were quite, um, generic, and after two weeks we saw it was a disaster, and we just cut it off, you know. We don't wait for the campaign to, uh, to finish. It's like, you know what, this is just not happening. This is just, this is, there's no way. This will, this will pick up, and you just cut it, you know, while maybe in other places you would just have continued and say, you know what, we're here, and then we'll do the retrospective at the end of it, say, no, I mean, it's not working, it's not working, you just cut it, you know, and fine, your TV campaign that was supposed to run, let's say, for three months, well, after two weeks, it's just not working, you just cut it, and then you move on, you know, it's fine.

AI assessment note: “creatives that were quite, um, generic, and after two weeks we saw it was a disaster”

Q What did you try that didn't work as planned and you cut? And what did you learn?

A Well, so, I mean, on the marketing side, um, we've been, uh, uh, there's been some, uh, some campaigns, for example, that we've been running in some, in some countries where we've been using some, uh, creatives that were quite, um, generic, and after two weeks we saw it was a disaster, and we just cut it off, you know. We don't wait for the campaign to, uh, to finish. It's like, you know what, this is just not happening. This is just, this is, there's no way. This will, this will pick up, and you just cut it, you know, while maybe in other places you would just have continued and say, you know what, we're here, and then we'll do the retrospective at the end of it, say, no, I mean, it's not working, it's not working, you just cut it, you know, and fine, your TV campaign that was supposed to run, let's say, for three months, well, after two weeks, it's just not working, you just cut it, and then you move on, you know, it's fine.

AI assessment note: “campaigns... using some, uh, creatives that were quite, um, generic, and after two weeks”

Q What do you mean you cracked the TV measurement framework?

A So we were able to have some, some good level of attribution of, between TV and the downloads, and to get, to be able to really know, um, how much we could attribute to our TV activities. And, and, and because we managed to get that, then we were one of the very first companies To actually make, uh, to make that work, which this is a very good example. The other good example, uh, is, uh, iOS, performance marketing on iOS after scan, and you have companies that know how this works, and therefore they can measure it, and therefore they can scale, and you have all the other ones that I mean, are really struggling and so on. At the end of the day, they spend maybe two, three times more on Android versus iOS, you know, which is a nonsense, you know.

AI assessment note: “The other good example, uh, is, uh, iOS, performance marketing on iOS”