The Exchanges

Q How do you think about true activation? Like, what does an activated user mean for you? It's someone who lands and converts, it's someone who lands and converts and pays. What is true activation?

A So I live in a world where not everybody who uses your product is a paid customer. So I also live in a world where every single customer has other value besides just paying to you directly. So for example, at Lovable, we have a ton of free users. A lot of our costs are within our freemium, not with our paid marketing, not with employees, not with marketing or sales budgets. It's with Freemium, because we view Freemium as actually marketing channel. What I mean by that is that a free user, when they come in and they get delighted by what they accomplish for free, they go and do marketing on our behalf. They go talk across socials. They go refer to our friends. We very carefully measure. We have such things called lovable score, where we measure how often people actually refer us to somebody else, and we keep a really close eye on it, because That is an earned channel that you cannot buy, that you cannot compete with, and that we own. So free user holds value to us. So even if we acquire through paid marketing channels, let's say for users, they hold value to us. It's not all about conversion, but back to your activation question to me, it's purely engagement based. That has nothing to do with, um, monetization whatsoever. So what is it that your aha moment is with the product? What is it steps for you to set up to get to that aha moment? And, um, what are the first habit loops t…

AI assessment note: “back to your activation question to me, it's purely engagement based.”

Q When we think about video, what would you do with video? How would you think about what would work, what doesn't work, what you see that's cool, what's not cool? How would you approach video in a world of unlimited, no constraints?

A So I think AI generated video is the future. I was just playing, uh, with video platform where just I uploaded my, uh, just still photo and it just created a full thing of me moving my hands, doing the full ad with like back and forth transitions with, with an app screen. I think that there's still going to be a place for manual video creation. And those are always going to be like artisan videos where it's just like handcrafted hand, Hand created. It's like a craft that it's like, think about like taking a picture versus painting the picture, uh, with the, with watercolors. It's like, there's still going to be some differences there, but I think for advertising needs, AI is where videos and even any creatives are going to be at AI models on video creation and just visual creation are improving in such a rapid way and they can create these striking images And they can animate you in the way that you would never be able to record properly in one shot. Uh, that I think it's the future. I'm like playing around even with creating animation for myself when I'm reading my own blog, as opposed to like me recording it. So let me just put in my screenshot, give it the blocks transcript and let it do it for me. So it looks like I'm reading it out loud and like annotating it, but it's not me. It's just AI doing all of the work. So I think the future of advertising is going to be almost pu…

AI assessment note: “So I think AI generated video is the future.”

Q How do you prevent morale from reducing when numbers are down in very transient worlds? When it's summer, people are outside more, they're not unlovable as much, say. How do you prevent morale seeping when numbers are not there?

A So I think step number one is try to have pre-mortems about what would happen if your numbers are start to go down. So you have the action plan at the earliest time of noticing those numbers, so you're not reacting to it. You already have a go-to-war plan, so to speak, when something goes wrong. So this should be happening before any big release, before any big launch, before end of the year. You should look at this, do pre-mortems on the best, on the, on the, Best path forward is the worst things happen to you. Competitor enters. And so like, I really believe in having those conversations early on. Number two, knowing the key predictive indicators that will lead you into the decline, not just looking at revenue, because when moment revenue starts declining, it's already too fucking late. So like, you need to know like beforehand, what will I see? What will lead? What will warn me of that decline? Just so you at least have a couple of months, maybe six months to react before actual revenue impact is going to get hit. So those two things I think are important. And I think number three, we just have to accept the fact that some people don't do good with negativity of the business. They're not going to be the ones that are stand up and try to reverse it out. And you need to look at your culture and say, there are going to be people that are passionate about what they do, that are …

AI assessment note: “try to have pre-mortems about what would happen if your numbers are start to go down”

Q Why do most families go wrong on the onboarding process for various themes? And I'm sure you've got some thoughts on this one.

A I'll tell you number one mistake. Not having data team ready for growth team. So hiring growth teaming before you have data analysts or data literacy in your company before you're data informed or hiring growth and not attaching any data analysts into their for you. Because if you do that, what is growth person or growth hire supposed to do? Use their intuition? Then that's not really a growth mindset. They're just gonna pick an area and most likely have a strikeout versus Creating something that is more sustainable. So data, even though it's a different department, a lot of times they did this fall into growth on because that's the most department that uses data the most, but not equipping them with data team and data analysts is the biggest problem that you can create. And then there's number two, expecting wins right away. You hire your growth, your growth person, your growth team, and then the three months you expect the forecast lift where you put assumptions that they didn't drive forecast lifts. It's the worst thing that you can do to growth team in terms of really messing up their onboarding, in terms of them creating that predictable, sustainable system.

AI assessment note: “Not having data team ready for growth team... expecting wins right away.”

Q You said there about also the second one, like expectation setting and not pushing people into the wrong expectations. In terms of helping me again, you're my advisor, I'm the CEO. What should I ask the new head of growth to do in his first few months? How do I not set those expectations that could mislead him and his actions? Help me.

A Yeah, so there's three main stages that a growth team should go through. The main tool of any growth team is experimentation. I am able to figure out in a scientific matter which correlations are causations or which are the levers that I can pull to drive impact to my growth system and the growth model and accelerate the growth. So the stage number one is just to be able to run experiments. It sounds simple, but it's not depending on your infrastructure and experimentation tools. A lack that you have. And it's also very much depends on the metrics that you able to track and explain what's happening in experiments. So number one, learn how to walk, learn how to actually set up, run quick tests that don't take six months, by the way, to go out to production. So you're able to experiment stage. Number two, you were able to learn from those experiments. One of the most important things that growth team can add to your entire data and just company culture is the third data set of data. You have quantitative data set, which is transactional data set of what's happening in your product, which is your Google analytics data, your mixed panel data, your segment data, whatever it is. Then you have a qualitative data set. Those are all of the user interviews, maybe surveys that you're sending or people who share their emotions or their views about the product or how they see the value of t…

AI assessment note: “there's three main stages that a growth team should go through.”

Q numbers matter to you? If we go, you, you acquire a customer, you're like, is that good? Because we've got no idea how much they're going to pay over time. You look at lovable, you know, and bluntly you can spend a lot or you can spend little. It's nice in that way. But you don't know how much they're going to spend. How do you determine what numbers matter?

A So if I'm doing a paid marketing, the payback period is the only thing I look at. How quickly do I recuperate my money on average from any given campaign? And I try to really try to do at least the last click attribution perspective, especially if it's performance marketing, a pretty tight tie in between those numbers. If your conversion windows are really long, So not everybody converts within 24 hours, let's say, and if it takes you six months, a year to convert, again, very dangerous to invest into paid marketing, because then you don't even know if you're going to recoup that money, so unless your conversion window is under three months, I would say don't start with paid marketing as your growth lever, because you will never be able to make it spin fast enough, because you don't, you have such a long conversion windows. If you have shorter conversion windows, I think it's very important To also look at any of the signals beforehand, how people are activating, how many of activated people are actually converting to paid, because there's going to be many KPIs that you can look into predict conversion or even predict retention that are not monetization based. So really understanding that behavior in your user base becomes the number one concern for you to be able to even know, do I go into this or do I not? And again, paid marketing to me is, uh, Is a cheap way out, and it's n…

AI assessment note: “if I'm doing a paid marketing, the payback period is the only thing I look at.”

Q Growth span marketing. Would you do? Oh fuck. I didn't sponsor a real Madrid shirt. Would you do an F one car? Would you pump money into Anton's brand and give him a media team to do daily vlogs and videos of what he does in a day? And.

A Okay. A couple of tactical things come to mind. I would still do out of home advertisement. I think that is coming back in spades because that's where eyeballs are at. That's where you can actually like capture attention. I would go very heavily into video and talking, not just visual ads, because then you can explain things so much easier. Uh, so I would go after Amazon prime videos, Spotify is, um, and all of those of the world, because there's just like not a lot of competition. I will go off to a creator economy so hard. I would have every single newsletter, every single podcast, have my ad on it and just box out everybody else. And then I would send every single one of my customers my own t-shirt with a hat and just have them being my work walking billboards all over the place and they would be top quality. The things that they actually would want to wear and they would feel like there is like absolutely kick ass thing that is a cool thing to have. So I would just deploy as much as possible at the end of my customers and people where they're already spending eyes and eyeballs at, ah, to have my brand all around it.

AI assessment note: “A couple of tactical things come to mind. I would still do out of home”

Q job. Um, when we look at that, we have so many growth leaders, marketing leaders who are like, okay, we have a channel that works. Do we double down? How do we think about channel makeup when we're thinking about growth? What lessons do you have on when to concentrate, when to diversify, how many to diversify to? Was it right to diversify from socials from Anton? Love your thoughts.

A Yes. So when I say diversify from Anton, that's not to say that Anton is not a meaningful channel anymore. I'm just saying we have other channels running in parallel. So he's not the only one that he was at the beginning. And that's a really big difference because he's still growing his channel and it's still very important. We're just not, it's not, we're not relying on it as the only single failure point. Second of all, I would say that if I was any company, I would really heavily invest in social, but not in social in the traditional sense of You, uh, posting from your company account or hiring an intern who is your social media manager that is posting some puns on Twitter here and there. I really mean where you start building in public and you encourage your employees to build in public and grow their social followings and really reach your customer through your own team's voices. I think that is the channel that so many companies underestimate how important How important and how powerful it is, how much trust it builds, how much it unites people behind you in your mission, because they see the people that build the company and they want to connect to other people. We're with our little lizard brains still want that connection. Like we earn for it. And then the world where we're bombarded with marketing messages, everywhere we turn, be able to see and be able to connect is …

AI assessment note: “we're not relying on it as the only single failure point”

Q Does that mean how you hire? Like, if you're hiring for a growth role today,

A I think that there's still a path to hire. Hey, there is a generalist and then there is a specialist. Uh, so somebody had kind of like jack of all trades at the kitchen sink of all of the problem solving. Those people are very powerful, especially very early in company stages. As you start to grow and you actually say, okay, these are the channels that are really working for me. I really need to hire specialists that truly can fine tune every single channel and get like the squeeze the last five, 10% out of it. So you still need specialists and It's like a CEO specialist. That's still a thing. It's very much still a thing. And there's need for that. However, everybody now is also expected to do everybody else's job in some capacity, at least on the average level. So the level of the capabilities that all of us do has increased. So for example, I now ship code to production. I have never done that before, but I also go and I ship my own apps and run my own campaigns completely without anybody's support. I build apps. I go out there, I create a noise around it. I drive engagement to it. It's part of like all, all of the growth works to me. I write copy. I make prototypes too, for our engineering team that, that is designing it. Sometimes design shits all over it. Sometimes they say, good job. You can go through it. But all I'm saying is like, I'm doing all of this now myself vers…

AI assessment note: “that is expectation to me for every single employee”

Q Can we seed it? Can we, like, pay for UGC? You see UGC marketplaces, you see, like, head of UGC. Is there a way where we can artificially UGC community?

A So you can see it, but I actually think that what you need to find is your early super users and your early super adopters. Those people that are really excited about your products and you should pull them in to be your community managers. You should pull them in to be your community ambassadors. You should deploy them to be your biggest advocates and for them to start bringing other people around them because of their excitement for the product. But you do need to identify them very early and build community around Them. That is the right way to do it because then they will bring the positivity into it as opposed to it being just a support outlet.

AI assessment note: “So you can seed it, but I actually think that what you need to find”

Q You said there about also the second one, like expectation setting and not pushing people into the wrong expectations. In terms of helping me again, you're my advisor, I'm the CEO. What should I ask the new head of growth to do in his first few months? How do I not set those expectations that could mislead him and his actions? Help me.

A Yeah, so there's three main stages that a growth team should go through. The main tool of any growth team is experimentation. I am able to figure out in a scientific matter which correlations are causations or which are the levers that I can pull to drive impact to my growth system and the growth model and accelerate the growth. So the stage number one is just to be able to run experiments. It sounds simple, but it's not depending on your infrastructure and experimentation tools. A lack that you have. And it's also very much depends on the metrics that you able to track and explain what's happening in experiments. So number one, learn how to walk, learn how to actually set up, run quick tests that don't take six months, by the way, to go out to production. So you're able to experiment stage. Number two, you were able to learn from those experiments. One of the most important things that growth team can add to your entire data and just company culture is the third data set of data. You have quantitative data set, which is transactional data set of what's happening in your product, which is your Google analytics data, your mixed panel data, your segment data, whatever it is. Then you have a qualitative data set. Those are all of the user interviews, maybe surveys that you're sending or people who share their emotions or their views about the product or how they see the value of t…

AI assessment note: “three main stages that a growth team should go through”

Q Do you think that's a bit wooey, Elena, honestly?

A Listen, I'm like, I'm like terrified of the fact that I'm becoming a brand marketer because I'm just like, because like it's becoming so much more because of brand. People will follow the brand because of its promise, because of its relation, because functionality is going to be available everywhere. So I don't know what else can differentiate you besides creating an absolutely delightful experience. Stressing about the most littlest pieces of friction in your product, being really fast on innovation because you still cannot fall behind, but at least being like on the forefront, even if you're not the leader, but it's that relationship that you develop with your customer is everything. How lovable are you? Because if you're not lovable, I think the end is near for a lot of companies that are only purely focused only on just Pure functional ability of the products.

AI assessment note: “terrified of the fact that I'm becoming a brand marketer”

Q A lot of startups today have subsidized LLM costs, very much like AWS did and Google Cloud did with credits for years, and Anthropic and OpenAI have done, ah, now very much with this generation of startups. Should they account for them in costs or should they let themselves be artificially impressed by their margins despite this clear steroid injection of free for this time period?

A Why would OpenAI sell the LLM under cost? Yes, it's like market grab for sure, but I fundamentally believe that they think that LLM's costs are gonna have to come down to the point that they're not gonna have to raise the prices, but they're gonna start getting the margins to actually make those prices make sense for them and potentially lower it even further. Because LLM is gonna be like a commoditized feature Very soon. And it doesn't feel so right now, but it will be like access to internet or access to the cloud. It's commoditized. It's cheap. So I think that I believe that they're also believe that it's going to come down quite a bit once we get efficiency gains out of how LLMs are constructed. And, um, this is where, when it becomes completely commoditized, you cannot monetize on it anymore. You have to start moving towards monetization of outcomes. And again, Whoever evolves their monetization models first there will win.

AI assessment note: “Why would OpenAI sell the LLM under cost? Yes, it's like market grab”

Q Do you think that's a bit wooey, Elena, honestly?

A Listen, I'm like, I'm like terrified of the fact that I'm becoming a brand marketer because I'm just like, because like it's becoming so much more because of brand. People will follow the brand because of its promise, because of its relation, because functionality is going to be available everywhere. So I don't know what else can differentiate you besides creating an absolutely delightful experience. Stressing about the most littlest pieces of friction in your product, being really fast on innovation because you still cannot fall behind, but at least being like on the forefront, even if you're not the leader, but it's that relationship that you develop with your customer is everything. How lovable are you? Because if you're not lovable, I think the end is near for a lot of companies that are only purely focused only on just Pure functional ability of the products.

AI assessment note: “terrified of the fact that I'm becoming a brand marketer”

Q Do you try and get attribution from it? Do you try and do, uh, like referral codes, discount codes, whatever it is to get attribution? Or you like, you know what, it's in the, it's in the brand bucket. Fine.

A I think discount codes are always nice, uh, just to give that audience a little something. I truly believe in, I wouldn't say like a pure discounting strategy, but using your product, uh, for getting your product away for free for people to try as the biggest portion of your marketing budget, I think always has to be the case. So what I mean by that, let me just explain real quickly. Yes, you can do discount codes. You can just give free offers. You can make your freemium even bigger, but If you look at your overall marketing spend, your free giveaways, whether it's part freemium, whether it's discount codes or whatnot, has to be bigger than your paid marketing spend. That, I truly believe, is the right ratio for every single company, because if you're not using your product to acquire customers in some sort of offers way, I think that you're just giving too much money away to Googles and Netas of the world and not putting enough pressure on your product to wow people, but Discounts are good.

AI assessment note: “I think discount codes are always nice, uh, just to give that audience a little”

Q A lot of startups today have subsidized LLM costs, very much like AWS did and Google Cloud did with credits for years, and Anthropic and OpenAI have done, ah, now very much with this generation of startups. Should they account for them in costs or should they let themselves be artificially impressed by their margins despite this clear steroid injection of free for this time period?

A Why would OpenAI sell the LLM under cost? Yes, it's like market grab for sure, but I fundamentally believe that they think that LLM's costs are gonna have to come down to the point that they're not gonna have to raise the prices, but they're gonna start getting the margins to actually make those prices make sense for them and potentially lower it even further. Because LLM is gonna be like a commoditized feature Very soon. And it doesn't feel so right now, but it will be like access to internet or access to the cloud. It's commoditized. It's cheap. So I think that I believe that they're also believe that it's going to come down quite a bit once we get efficiency gains out of how LLMs are constructed. And, um, this is where, when it becomes completely commoditized, you cannot monetize on it anymore. You have to start moving towards monetization of outcomes. And again, Whoever evolves their monetization models first there will win.

AI assessment note: “Why would OpenAI sell the LLM under cost? Yes, it's like market grab”