0:00:02 - (Matt Widdoes): Welcome to the Growth@Scale podcast. I'm your host, Matt Widdoes and CEO of MAVAN.com. Today we're joined by Eric Seufert. Seufert is an analyst in mobile development, tech marketing, data analysis and consumer expert. He's also the general partner at Hercules Capital, a pre seed stage venture capital fund dedicated to investing in the future of mobile tech. Eric also runs the popular digital economy trade blog Mobile Dev Memo and built and maintains Thesis, an open source Python library for marketing cohort analysis.

0:00:29 - (Matt Widdoes): Throughout his career, Eric has worked at transformative consumer technology companies within the mobile ecosystem like Skype, Wooga, and Rovio Entertainment, the developer behind the popular Angry Birds franchise where he was the vice president of user acquisition. Welcome to the podcast today, Eric.

0:00:44 - (Eric Seufert): Hey Matt, thanks for having me.

0:00:45 - (Matt Widdoes): Yeah, I'm super excited to dive in. So for people who don't know you, tell us, who are you? Where have you been? What do you do? In your own words?

0:00:51 - (Eric Seufert): Yeah, so I spent the kind of first stretch of my career as an operator, primarily in Europe. I was. I went to graduate school at University College London. I went on a full scholarship that was granted by the European Commission, and I stayed. I stayed in Europe. My first job, post grad school was at Skype. I was there through the acquisition by Microsoft. Then I saw I was based in Tallinn, Estonia there, which was Skype's largest office and where it was founded, I saw that a number of companies just across the Baltic Sea and Helsinki were doing interesting things with the freemium model in gaming. So I moved to Helsinki and I worked at a number of mobile gaming startups.

0:01:31 - (Eric Seufert): Then I moved to Berlin, worked at a company called Wooga, which was a casual games developer which ultimately sold to Playtika. Went back to Helsinki to work at Rovio as the vice president of user acquisition. Rovio ultimately went public on the Helsinki Stock Exchange, the Helsinki Nasdaq. Just recently we sold to Sega. But after Rovio, I left and right. When I left Rovio, I started a company that built an analytics platform for mobile marketing teams. I sold that to a San Francisco based mobile gaming company, moved to San Francisco, worked there for a bit.

0:02:04 - (Eric Seufert): Then I kind of struck out on my own when my son was born, started investing actively as an angel. From that, I started an investing investment syndicate through Mobile Dev Memo, which is my blog, and the community that kind of sits adjacent to the blog. We did ten deals in nine months. I realized that running a syndicate was kind of a thankless job that involved a lot of uncompensated work. So I started a dedicated venture fund called Heracles Capital.

0:02:29 - (Eric Seufert): That fund, one is basically fully deployed at this point, but I was investing in the kind of where I am investing in the mobile ecosystem across ad tech, publishing, tech, and content. I wrote the book Freemium Economics, which was published in 2014. So basically a decade ago, and I published Mobile Dev Memo, which as you mentioned, is sort of trade blog dedicated to the digital economy. So I work with early stage startups through the fund.

0:02:55 - (Eric Seufert): I work with later stage companies as an advisor slash consultants, and then I write about the digital ecosystem, particularly as it pertains to mobile.

0:03:05 - (Matt Widdoes): Let's talk about freemium economics. So for anybody who hasn't gotten the book, I'm sure you can get it on Amazon. But let's get into the concept of freemium and some of the strategies that are used for leveraging that model. I'm just curious at a really high level, maybe for people who don't know what it is, it's probably worth a quick one, but I think most people these days, especially our listeners, will know what that is.

0:03:25 - (Matt Widdoes): But also the critical components of a successful freemium model and how series A to C startups can leverage that to drive growth and retention.

0:03:34 - (Eric Seufert): Yeah, so I became intrigued with the model when I was at Skype. Skype was, I mean, a lot of people probably don't know what Skype is. People below a certain age probably never heard of it. Skype was when Skype was acquired by Microsoft, they kind of mothballed it in favor of Teams. And so I don't know. I mean, Skype still exists and I use it sometimes to do VoIP, if I need to call some, some foreign number, but it's much less popular than it once was. But for those who are unaware of what Skype was, is a free product, right? You could download it on your desktop.

0:04:05 - (Eric Seufert): This was kind of before mobile was ubiquitous. Smartphones were ubiquitous. But you download on your desktop and you could make phone calls over the Internet to somebody else who had Skype. Now it's just totally free to use. So you could use it as a messaging platform, but you could also just call people. If they had Skype installed on their desktop, you could call them. Skype actually was a tremendously popular messaging platform. So I remember a lot of companies I used. Skype was the internal messaging tool.

0:04:31 - (Eric Seufert): It was pre slack. And the way that it was Freeman was, it was just totally free to use and except that if you wanted to make a call to an actual number phone number or if you wanted to have your own dedicated phone number that people could call. Then it was, it was a paid like product. It was a paid feature, right? So I thought, okay, this is really interesting. And as smartphones started proliferating, and this is like 2010 when I was there. So this is kind of post-App Store, post-iPhone, but still, people still kind of, like, primarily use dumb phones or that Nokia was very popular back then.

0:04:59 - (Eric Seufert): IPhones were not the primary smartphone device in a lot of markets as they are now. As they started to proliferate, I thought, well, I could see a world where the total addressable market for any product that is distributed through the app store or whatever approaches the entire population of the earth. It could make a lot of sense if the product was free and everybody had access to it, and there was no sort of hurdle to clear in terms of, like, disposable income, discretionary income, that you could give your product away for free and then just sort of monetize people on the basis of their proclivity to engage, or they're sort of like the magnitude of their engagement with the product, right? So that people, and instead of gating the product behind a fixed price point, you could actually index the amount of money that people gave you with their enjoyment of the product, right, with the joy that it brought to them or the utility that they received from it.

0:05:56 - (Eric Seufert): And so instead of just having one single price point, you could charge for a bunch of different things and essentially allow people to pay you in a way that mapped to the value that they received from the product. So gaming, what I saw people doing with gaming, and this was kind of like the, at this time, this was like the dawn of free to play gaming in the west. Gaming was doing that very, very sort of expertly, right? And it started on kind of like Facebook Canvas, but then it was moving into smartphone, you know, dedicated smartphone apps where the game was free, anybody could download it. Gaming is obviously like a very popular medium of entertainment.

0:06:32 - (Eric Seufert): It remains, you know, so it's the biggest single sort of form of entertainment by revenue. And then just like allowing people to make what are called microtransactions as needed or as they, as they found utility in those microtransactions. And so you got to this dynamic where you had the value of a user could fall across a distribution. So there wasn't like, well, everyone is worth this much because everyone pays us once, and it's this amount of money. But people's value to the developer of the product fell across the distribution. And with games, you saw that this distribution can be very long tail. Right? So you had people that spent a lot of money.

0:07:08 - (Eric Seufert): Thousand dollars, $10,000. And when you got into that sort of situation where people could potentially spend very large amounts of money, and again, ideally, and this works best when that level of spend is commensurate with utility they extract from the game. This isn't meant to be manipulative. You're not meant to coerce people into spending this. If you give them the opportunity to enjoy the product to that extent, then they're probably more willing to pay versus just getting coerced or tricked or something.

0:07:37 - (Eric Seufert): But if you create that dynamic, well, then you can make a lot more money than you otherwise would have given a fixed price point, because with a fixed price point, you disqualify a lot of people. There's just a lot of people who will never pay for something or who can't pay for something. But if everybody could potentially pay for this product. Cause it's priced at zero, then you create this dynamic where this very long tail of what's called LTV and the economics of that could be much more powerful than just a fixed price point product, because there's other benefits to having a lot of people use your product than just whatever the monetization potential is. You get network effects, you get virality effects.

0:08:14 - (Eric Seufert): And so that model was kind of nascent. In the west. It existed for a long time in Japan with free to play games, but in the west, it was fairly new, and I thought it was really intriguing. Having just come out of a master's degree in economics, I was really intrigued by this model. And so I sort of just spent a lot of time thinking about it, studying it, wrote a book called Freemium Economics, which is ten years old, but I think it more or less holds up.

0:08:38 - (Eric Seufert): It was a very high concept. It was very big picture. And now the freemium model, especially in mobile, is. Is totally pervasive. I mean, it's the dominant model. It's almost like the only model. For the most part. There's a handful of scaled companies that are scaled products that are using a fixed price point. For the most part. The vast, vast majority, the great preponderance of products that exist as apps in the mobile ecosystem are freemium.

0:09:06 - (Eric Seufert): It's dominated. It's dominated the mobile ecosystem. But that wasn't always true. I remember when my book came out, there was a big question mark around whether this model had staying power, whether people ultimately would prefer that their products be monetized in this way versus just an all you can eat kind of thing with paying one fixed price point and getting as much engagement as you want or whether they would want to tie engagement to microtransactions or purchases for specific things.

0:09:33 - (Eric Seufert): Well, I guess that question has been answered now, but it very much was an open question at the time.

0:09:37 - (Matt Widdoes): Yeah. And what are kind of core components of that? And, you know, we see it a ton in gaming, to your point. And that network effect of having, you know, effectively, you know, if you have a billion users or a hundred million users and the fact that, you know, 98% of them don't really pay, which is fine. And even in a product like Skype, you, if you can get everybody using that for VoIP, it's like, okay, we'll figure out the money later and people will need to upgrade later. Or I use it free for five years, and then it becomes the main platform of a business with thousands of people or something, how that allows people to kind of grow with you and gain market share and kind of create a moat potentially around the product because if somebody else were to cut, you know, if your minimum is even $5 a month and somebody else comes out with something just as great but is free to use, you're going to have some bleed. There any kind of core components that make freemium models, you know, successful generally?

0:10:32 - (Eric Seufert): Yeah. Well, I mean, that's the, you know, the problem with the race to the bottom is at some point you reach the bottom. And so, like, yeah, freemium sort of represents the end point there because you're. You're probably not going to go negative price point, which means paying the user. Right. Although that was kind of the theoretical model behind a lot of crypto.

0:10:47 - (Matt Widdoes): PayPal did some of that, too. Early days with people with $20 and stuff like that.

0:10:51 - (Eric Seufert): Right? Yeah. So, okay, there's a couple things. Like, for one, and so there's a. There's multiple ways to implement the freemium model, but I think for the most part, like a purist would say, freemium is monetized via what are essentially like one-off transactions that, in aggregate could represent a lot of money from any single user. Right. Again, commensurate with their value that they derive from the product.

0:11:15 - (Eric Seufert): I don't think it's truly freemium if the product's monetized with a subscription versus just a free version of the product, that's more limited. I don't see that as freemium. To me, freemium is, you essentially have some base set of functionality, and the paid components of that improve the experience beyond that or add something substantive to it, but that the core product experience is complete even at the freemium price point.

0:11:42 - (Matt Widdoes): Takes as an example, right?

0:11:44 - (Eric Seufert): Yeah. So I think in order for a product to be successful with freemium and Skype was not a successful company, by the way. I mean, they were going to go public. You know, the history here is, I mean, not. Not that many people are familiar with it because, I mean, I think it was an important company, but it wasn't that big. It's gonna go public. But it was losing a lot of money, and it made, you know, very few people converted. Right. To the paid product. And there was just that one product. I think in order to be a successful freemium product, you have to have a very wide catalog of things to offer to the user that are paid, that are paid upgrades or paid add ons or paid experience improvements, right.

0:12:19 - (Eric Seufert): Because you need to be able to facilitate that sort of, like, maximum level of utility. Right. For any given user. And so, like. And I wrote. I wrote in the book in 2014, this is probably more possible now, but, you know, the sort of, like, abstract theoretical maximum would be like total personalization. Like, everything could be personalized to the taste of the user. Well, back in 2014, I mean, that that was sort of unrealistic.

0:12:42 - (Eric Seufert): But now with generative AI, there's probably more viable to do that, right? You can personalize the experience across every possible dimension, from the aesthetics to whatever the in product items that they're purchasing. All of these could be personalized to the taste because they could be sort of generated in real time with generative AI. But you need a very broad and diverse set of items that the person could potentially purchase because that facilitates that long tail of the LTV, that facilitates the person being able to spend a lot of money if they get the value out of that.

0:13:15 - (Eric Seufert): And you need a very broad, total, addressable market. I mean, this only works if a lot of people could use the product. If it's a niche thing, you're not going to get the absolute numbers that you need to unlock the promise of this model. If it's a niche thing, you just give it away for free. Well, then the absolute number of people that pay anything is going to be probably too small to support a business.

0:13:39 - (Eric Seufert): You need a very large, total, addressable market. You need the ability to service people as their personal tastes dictate in a way that unlocks the ability to spend a lot of money. And then you need all of the infrastructure that allows for that. I make the point in the book that building a successful freemium product is as much a function of astute product design as it is analytics and data management and data science, because you need to be able to parse signal from people to determine, like, how to drive in the direction that allows them to unlock a lot of value.

0:14:17 - (Eric Seufert): A very straightforward example of that is just a recommend recommender. Let's say I do have a recommendation system. Let's say I do have a product catalog of a million items. Well, I can't just sort them alphabetically or by color because I need to push a person into the thing that is most relevant for them. And it's very much in the same way, like an ad system does, that you need to do that within the product.

0:14:38 - (Eric Seufert): You need to surface the things that are the most relevant to the person to support them, purchasing them, or even having visibility into them. So you need that sort of infrastructure that's more of, like an operational thing. But the value in having the large TAM and a lot of people using the product, even if it's just for free, is you can, a, you benefit from viral effects, right. More sort of, like, there's more of a possibility of benefiting from viral effects than if you have a very small product, and there's just sort of, like, a larger ceiling to that. And then, B, you can monetize with ads, for instance. Right. And that. That is different than, you know, like, the freemium implementation of, well, matching the amount that's paid to the amount of utility, because that just tracks with engagement. Right. Ads are one to one with engagement. The idea with freemium is that it's actually, you're delivering.

0:15:25 - (Eric Seufert): You know, people can have different, differing levels of monetization based on, you know, the same amount of engagement. Someone who just gets more value out of this product is more specifically relevant to them, is going to pay more than someone who isn't, even if they engage the same amount because there's only so many hours in the day. The idea here is what you ultimately want to achieve with freemium is you want to maximize consumer surplus. You want to make sure that everybody is paying the maximum amount that they possibly could based on their utility in the product.

0:15:53 - (Eric Seufert): If you have a fixed price point, some people would have been willing to pay a lot more than that for the product because they get that much more value out of it. Right. And what you want to make sure of is that you sort of match in success, like, in this sort of, like, theoretical ideal here, is that everyone is paying exactly an amount that maps to the value that they extract from the product.

0:16:12 - (Matt Widdoes): Yeah. Which means you need to have a product that's segmented well enough to be able to walk people up and allow people who maybe are just using it for the most basic case at the time being a chance to get real value from just the core fundamental product, which is very different than offering somebody a free trial or something like that, because ultimately you're going to reach that same gate of subscription.

0:16:34 - (Matt Widdoes): I'm curious on the, if we take a click further down into diving into launch outside of business model and look at popular mobile gaming launches like Angry Birds as an example, which you led, any kind of key factors that jump off the page there for either a new startup founder or anybody really launching a mobile app, I mean, it doesn't even have to be in gaming, but kind of key components that people should be thinking through as they're going to market there.

0:17:02 - (Eric Seufert): Yeah, let me just correct something that I said. I said, you want to maximize consumer surplus. Sorry, I misspoke. You want to minimize consumer surplus. Consumer surplus is what the user gets in value above and beyond what they paid. You want to minimize that, right. You want to match that. Like they're paying exactly what they're paying at a level that sort of like meets the value that they derive. You want to minimize that.

0:17:22 - (Matt Widdoes): Equilibrium is perfect, basically. Yeah.

0:17:24 - (Eric Seufert): Right. So I said maximize. I should have said minimize.

0:17:26 - (Matt Widdoes): Minimize.

0:17:26 - (Eric Seufert): Said the wrong thing. Yeah. So I think there's apps for which it is easy to take the wrong lesson about product launches. Right. So a lot of these, like, viral apps that take off, you know, like wildfire, are actually usually end up being case studies in a misguided app launch. And there's reasons for that fundamentally. And I talked about virality and I said that's a good thing. If you can generate virality, if you can catalyze virality, that's free discovery to a point.

0:17:57 - (Eric Seufert): To a point. Because the problem with virality is it's actually almost exists as an inverse measure of depth of the product. And so the more likely that something is or the more capable something is of going viral, the less depth it has as a product to generate long term engagement. Why? Because as a fundamental feature of something, a fundamental feature of most things that go viral is that it's very easy to understand quickly, it's very easy to grasp almost instantaneously. And so almost by definition, that means there's not a lot of nuance to it, there's not a lot of depth, and there's not a lot of things to learn, and therefore there's not a lot of surface area for delivering value. It's easily assumed.

0:18:38 - (Eric Seufert): It's easily consumed. Right. It's junk. It's junk food. And so, like, you look at a lot of the apps that go viral very quickly, and people say, oh, what a successful launch. Well, they end up kind of petering out a lot of the times. What you want to do is you want to have, like, sustained growth, and a lot of times that's supported with user acquisition. Right. And so, like, there's an as. Like, almost like, as a corollary to the freemium, you know, the sort of factors of success that I described with the freemium model is the ability to do sustained user acquisition for that product.

0:19:05 - (Eric Seufert): That tends to mean it's got sort of, like, depth of the economy, supports just spending money on it, but also that there are. You've got the big TAM, you've got the recognition of a value proposition. It's very difficult to do user acquisition for something that could go viral, has the capability of going viral. It's generally very hard to do user acquisition for those things because what problem are you advertising that you're going to solve if it's that easy to grasp and you could solve it with this very superficial product that has no depth, was it really a problem or is it just a junk food snack?

0:19:38 - (Eric Seufert): Right. And so that depth tends to be inversely correlated with the capability of going viral. That's not to say that you can't have virality because people talk like, hey, this thing solved a problem for me. You should check it out. But when you see an app shoot up to the top of the top downloaded charts, well, then that, you know, there's other reasons that could happen, but, like, apropos of nothing, it was just a viral app.

0:19:58 - (Eric Seufert): Well, then that probably means that it's also going to plummet just as quickly. Right. So I think there's just. There's examples of that happening that I think people take the wrong lessons from. Like, they want to try to emulate that, and you shouldn't, right. Because that's not a. First of all, you just don't. You can't control that. So just even if you approach the product line from that. From that perspective, do you really want to push something out into the ether that you have no control of? Because what happens, you know, probably it's not done like, no product's ever really done, right? Wouldn't you want to just sort of iterate and grow in, like, a sustained way that you could turn off or on or scale up or scale down if something goes viral, it's out of your hands.

0:20:37 - (Eric Seufert): And take the example of Clubhouse. That was an incomplete product. It went viral, though, and they actually tried to instigate the virality. That was a mistake. What they should have done was control the distribution, perfect the app, improve the app, iterate on the app, build an experience that promoted long term engagement, and not push this gimmick out into the world that everyone was going to try once but not return to. Because if you don't have retention, you don't have a business.

0:21:01 - (Eric Seufert): A sustainable business is something you can control with marketing. User acquisition is a really powerful form of distribution, right? Because you can turn it off or on. Virality is something wild and unmanageable. It's something that takes the app out of or the product out of your hands and allows just sort of the whims of word of mouth to control your fate. Right now, if you show somebody an unrefined, immature product, they confront the immaturities of it, they confront the flaws with it, and they churn.

0:21:33 - (Eric Seufert): You will probably never be able to get another opportunity to serve them. Right. That's just. That's just the way that works. Like, I tried it, eh, it was okay. Clubhouse was okay. I didn't really see the point. It was kind of interesting to hear from celebrity x and be, you know, in this live stream with them, but I don't know that I would use it every day. Like, are you ever going to go back? To my mind, I wrote this piece on Mobile Dev Memo a while back, like, how to launch a consumer app.

0:22:00 - (Eric Seufert): Well, how to launch a consumer app is you push it out, you get some users, you examine the retention, you find the reasons that people churn, and you address those and you iterate on those. It's just this very iterative, systematic process. Get some users in, track them as a cohort, see where they churn, improve those flaws, improve the reason that they churn. Consider which users I brought in in the first place.

0:22:28 - (Eric Seufert): There's a very strong sort of feedback loop between the marketing and the product analytics to determine, am I bringing in the right users? Is that being proven out with their retention? Maybe. It's actually a targeting issue in a lot of ways. You're like, it's both. It's a product problem and it's a targeting problem, and you're sort of like ping ponging the marketing data against the product data all the time to get to the intersection that provides the greatest possible TAM of relevant users such that they retain to a degree that supports sustained and growing user acquisition.

0:22:58 - (Eric Seufert): That to me is a systematic, methodical product launch. It's not fast, it's really slow, and it can be tedious, but that's the way you have to do it. You talked about Angry Birds. With Angry Birds, there just was no possibility of doing that. First of all, there was a lot of earned media because, hey, there's another Angry Birds game. People discovered it in soft launch. We had it named something else. It wasn't engineers, too, but it obviously was a sequel. And so there's just all this news. There's news media that we couldn't control and it was like one of the, you know, and still probably is it. When, when we launched this in 2017, it was one of the most iconic brands in gaming.

0:23:30 - (Eric Seufert): And so there was just going to be no way of controlling the distribution of that product. But, you know, and that's, you know, that's a nice problem to have in some ways, with a sort of a game or any sort of product that's starting from a position of just zero recognition. You have what is really the luxury of taking this sort of slow, methodical, iterative approach to expose more users to it, examine why they churn if they do tune the marketing to make sure that you're getting more of the users that are good and fewer of the users that are bad for whom this product is not relevant, and then make those iterations just sort of like in the cyclical manner that is the way to approach a product launch. But that's boring.

0:24:08 - (Eric Seufert): It's not exciting and it costs money and it requires, like, deep expertise in product analytics and user acquisition. And so if you tell that to, like, some three person startup team, you know, they're, they're, they're going to get disappointed and crestfallen if you tell them, hey, if you just hit the right notes, this thing could go viral and take over the world. And all of a sudden you're getting term sheets that's much more enticing. Unfortunately, those tend to not make great businesses.

0:24:33 - (Matt Widdoes): Well, and it takes a lot of patience, too. I mean, I think people inherently understand how interconnected these things are and inherently kind of understand. Yeah, we should be putting, we should be testing. So many people want to focus on testing and user acquisition as an example. And then you ask, what have you changed in the product lately? And they're like, oh, the product's done. It's like, well, the product's really never done. Just like the marketing isn't.

0:24:54 - (Matt Widdoes): And I think to your point, it reminds me of one of my favorite quotes from an old advertising professor I had, which he said, nothing will kill a bad product faster than good advertising. And so to that point that you make on the kind of unwieldy virality that can come from a kind of surface level product, is that one, like you said, they come in really quickly. Clubhouse is a great example of that.

0:25:17 - (Matt Widdoes): And there's a lot of heat, and everyone's using it, and then all of a sudden it peters out. And consumers kind of have a switch. Like, they've either considered it or they haven't. They've heard of it or they haven't. And you kind of only get one chance to move that to the on position because once it goes off to your point, someone's like, oh, are you using Clubhouse? And they're like, I don't use it anymore.

0:25:36 - (Matt Widdoes): They don't take the time to go back and look in two years at how much you've improved the product or not.

0:25:41 - (Eric Seufert): Right.

0:25:41 - (Matt Widdoes): And so you do want to make sure that it's fully optimized, um, taking a measured approach, making sure you have it buttoned up, making sure you have enough, enough depth. We had a game launch at Zynga once called Princess Bride Slots, and it was Princess brand or Princess Bride IP. And what the product teams didn't anticipate is that the heaviest users of these slots apps from the portfolio were able to essentially complete the entire game, if you will, in less than a month.

0:26:10 - (Matt Widdoes): And so then what happens? Well, they all bounce, and then it's like, wait, hold on. This is meant to be a, we're going to have more things coming soon. Like, wait, it's like, no, they consumed it already. And so you, you need a good, in an environment like that, you need probably a good two to three months tale of when stuff's going to launch. But even within an early stage app, I think trying to go too big, too wide, too early, and potentially burning your reputation or the kind of place that you have in consumers minds as either not being ready or just not for them can be very dangerous. And that's the exact opposite of what investors would want. And really what any founding team would want is to just be like, okay, so everybody says our app sucks now and we've reached a billion people. Okay, well, that's kind of a really, that's even worse than going slow. And I don't think people fully appreciate the risks of that.

0:26:59 - (Eric Seufert): Right. Investors don't either. And so a lot of times, what investors want is essentially a product that will suffer that fate. Right? I mean, because like investors tend to be very sensitive about, or businesses that are supported through paid acquisition because they just see that as like a drag on growth. Right. But the reality is, you know, they don't tend to not. They tend to not. Well, a lot of them have actually been operators, but they tend to not understand like how the most successful products are scale.

0:27:25 - (Eric Seufert): And it's not through just permanent, infinite viral growth, because that's not possible. Right. But there's other considerations here too, and that's why, like, the intersection of the product analytics and the marketing analytics are so important to manage properly. Right. So I wrote this piece a while back called The Growth Trap. And, you know, if you're a product that's just seen a lot of organic adoption, and you've been doing that sort of like ping-pong project between, well, we're gonna optimize the product based on the marketing data.

0:27:55 - (Eric Seufert): Well, the marketing data is just nothing. It's whatever. It's the cohorts of users that have discovered the product organically. Right. And so what you've. But if you've sort of like adapted the product to that, to sort of service, super service those users, you might. What you might be doing is making the product less appealing to the much larger group of users that would have to be reached through user acquisition.

0:28:19 - (Eric Seufert): Almost by definition, for a product that is starting from a baseline of zero and has no known brand attached to it and won't be discovered outside of active search by the user. The group of users that found your product, the group of users that sought it out, are going to be some super high intent subset of the very most relevant people for your product. If you optimize a product for those people, you could actually be alienating the much larger group of people that would enjoy it and would contribute money to it, but wouldn't necessarily go and seek it out. They're not power users. They're not potential power users. You have to think about not just the existing user base, but the entirety of the TAM.

0:28:58 - (Eric Seufert): And how will you reach, how will you onboard the biggest percentage of that? Is it through their own or like self directed organic discovery? Probably not. Probably. That's a very small percentage of the TAM. Probably the biggest. The vast majority of the TAM is, are people that you'd have to go and recruit and convince and sell with an ad. And if you've optimized a product for the people that would seek it out on their own, well, at some point, and it's probably very early on, you're going to run out of them. And you've just tuned a product that, to the point that it alienates all the other people, the people that are vastly more voluminous. Right.

0:29:33 - (Eric Seufert): And so, you know, that that's a, that's a real tension, though, because it's like, well, we're going to spend money on marketing before we've tuned the product. Aren't we going to waste a lot of that spend? We're going to be reaching people and they're going to come in and they're going to churn out because it's not, you know, perfectly tuned. Well, yes, but it's an investment. It's an investment to get the data to tune it so that, well, they're maybe gone forever. And that's why you go slowly and methodically. They might be gone forever, but there's a lot more of them out there. And then the next cohort that we bring in this product will be a little bit more relevant for them, a little bit more sticky, the retention will be a little bit higher, we'll gain some more data and we'll improve the product on that basis. And hey, by the way, the organic people might be bouncing earlier, they might be churning earlier, but that doesn't matter because that's not a sustainable business to be built for them. There's a sustainable business to be built for all the other people if we can retain them.

0:30:17 - (Eric Seufert): That's what I call the growth trap. It's this vicious cycle of making the product more relevant for the people that will actively go and seek it out, who are a tiny minority of the total addressable market for the product and creating barriers for all the rest of the people that could potentially use a product like this because they become alienated, because the product is so tuned for like the highest intent people.

0:30:41 - (Matt Widdoes): Well, I think so many people think of user acquisition as a singular function which is for the purpose of scale and growth, versus one of the most important tools, especially for an early stage company, is using user acquisition for discrete product testing. So testing our messaging and like what's resonating upper funnel. I don't care what happens later, we're just seeing what gets the most energy because you may find that some feature you test in paid acquisition is the number one thing is the highest converting from click, but your product is really weak there.

0:31:15 - (Matt Widdoes): And if you have a competitor and they're stronger than you in that camp, it should kind of incentivize you or cause you to kind of consider investing more into that feature. Or we want to understand how this resonates with a particular segment of consumers. You can't get that organically, right? We can go if you want to know, like, how do pregnant women view this product? Okay, well, there's kind of only. There's a very fast and easy way to kind of get that in.

0:31:40 - (Matt Widdoes): Pregnant is maybe a little hard to target on. But you get the premise.

0:31:43 - (Eric Seufert): I get the.

0:31:44 - (Matt Widdoes): So, yeah, and so I think so often people just view it as this mechanism for scale. And in some ways I can understand, especially people who are more like product-led, growth-driven, where they're like, I don't want to be beholden to that, but both can coexist. It's not binary. It's not like we only grow through paid or we don't at all, and we turn our backs entirely to it and we see it as this cost mechanism versus a revenue driver. And so I think those are all really great points.

0:32:10 - (Matt Widdoes): I'm curious on that. You mentioned data and analytics in that and any tools nowadays, I mean, that landscape is changing so fast and it always has. But kind of any big data or analytics tools that early stage companies should adopt from your perspective on just getting actionable insights?

0:32:27 - (Eric Seufert): Yeah, I mean, I think the way that this space has evolved is actually like a really. I would say that this space has evolved in a direction that is really productive from the perspective of optimizing the product. Right. So the way that the space used to exist, or like this sort of like, approach that companies used to take with this is they wanted to own the end to end system for aggregating data, collecting data. Aggregating data and then providing insights into it that could be actioned upon.

0:32:55 - (Eric Seufert): And so if you think about like an amplitude, it's like, well, we own the data warehouse, we do the aggregations, and we provide you with the dashboards. And the way that this space has evolved is to create components of all those different pieces and let people mix and match. And that's a much better way of doing it because it's much more flexible. If I think about how people collect data or aggregate data now, BigQuery, Redshift, Snowflake, but then you can attach whatever front end tools you want to that it's Looker or it's Tableau, but the way Google Sheets, it's just much more flexible and it produces much better results because companies are able to track what they need to track, first of all, to identify what data is meaningful to them, what data supports, insightful analysis, and then to build the tools on top of that, that surface, those insights, the best.

0:33:48 - (Eric Seufert): I think the data analytics space is in a sort of much more helpful, it has a sort of much more helpful and just general business model now than it did, say, ten years ago, which is a good thing. But then that also entails a lot more choice and a lot more sort of usually more work on the part of the developer. If you want something plug and play more customizable, it won't be customizable. So I suggest to companies, like, okay, well, if you're using Amplitude, at some point you won't be, like in success.

0:34:18 - (Eric Seufert): And so it's probably better to just prepare for that future. Now maybe you use it and then you transition off of it or something, but you're going to need the data aggregated in a way that is more flexible at some point. And so it's actually really, that's part of the lock in. Right. They own the data and the data is aggregated in a way that maybe doesn't make sense to you. And like the way that you think about like an event catalog instrumentation, right. Like how you think about like the atomic unit of an event, right.

0:34:43 - (Eric Seufert): It may, they may implement that in a way that's unintuitive to create lock in, right. So I think that to me is, you know, when I've seen companies try to roll your own from the very beginning, a lot of times that's just sort of like unnecessary. It's a distraction, it's like unnecessary, you know, resource, it's heavy investment. But, but at the very least you want to think about that, like what is, what is the atomic unit of an event for my product?

0:35:03 - (Eric Seufert): And how can I think about that? Maybe I can instrument it myself. You probably wouldn't want a redundancy, but then map that backwards into an action that someone takes in your product, what things are meaningful. And that's just really just a product design exercise. But starting there, I think really makes sense. Then at some point you implement that yourself and you have your own data warehouse or whatever, but at the very least you understand what drives outcomes.

0:35:28 - (Eric Seufert): What drives outcomes for my business. And then how do those sort of get strung together into a user experience? That to me is the sort of critical exercise, the critical project from day zero in building a product is thinking about what drives business outcomes for me. And how do I break that down into atomic units? That can be indicative of someone engaging in a healthy, productive way with the product or not.

0:35:53 - (Matt Widdoes): Yeah. And how do those stack? And then what are the early indicators of somebody taking some down funnel event, like a revenue event? And how do we use predictive analytics to kind of read that sooner in the user experience and optimize against that? I think so. Oftentimes people are, they have product features that they tout a lot, that they market it, it's on their homepage, it's whatever. And then you speak to the users and they're like, either one, I didn't, I didn't know that existed. So it didn't have the right surfacing or, yeah, it's not that cool. Like they think it's really cool, but maybe they're ahead of the time, maybe they're behind. Maybe they just are, again, creating something for some really niche part of their audience. Where, to your point, previously, 90% of the users maybe don't care about it, but all of the internal attention is going against that kind of granularity on the data and really using that to understand what there's, like, the product you think you have and the product you have and the product you have is dictated by the market and what people say about it.

0:36:47 - (Matt Widdoes): No matter how much you want something to be true, it just, it is what it is. And we've all seen features where people are like, this is really cool, this is really cool, and nobody cares. I think of, I can't even remember the name of it, but it was like Google social component. What was that? Do you remember that?

0:37:00 - (Eric Seufert): Circles or something?

0:37:01 - (Matt Widdoes): But it's like nobody wanted. Yeah, nobody wanted that. And it was like. But that was at some point the future of Google internally, I'm curious, like, you've put a lot of effort and time and have had a number of cycles over the last decade on big data projects. I'm curious. And you run Thesis, which is this huge repository of Python scripts. How could early stage founders best utilize these kind of cohort analysis tools like these to improve their performance?

0:37:28 - (Matt Widdoes): Any thoughts there?

0:37:29 - (Eric Seufert): Well, so I think one thing that people don't do, which is like, I think really critical, and it sort of pertains to this consideration of the marketing data intersects with the product data, and I need to sort of find the optimal intersection of that in order to be successful, is cohort aging, right? So a lot of times people look at, like, dnu maybe, but they're looking really just at DAU, and they look at DAU as this monolith, right? It's not, right. DAU is composed of cohorts of certain ages, right? And at some point, you know, in success, the cohorts are compounding the vast majority of the DAU at any given point in time is going to be not new users. Right. I mean, it would be very bad if the, you know, if it was all churned or.

0:38:11 - (Matt Widdoes): Yeah, everyone else churned. And you are just sustained by the latest batch of the latest, right.

0:38:16 - (Eric Seufert): Yeah. If on a sort of like, you know, kind of consistent basis, your DN, your DAU is primarily DNU, which means everyone's turning out.

0:38:24 - (Matt Widdoes): Right.

0:38:26 - (Eric Seufert): If you're compounding, which is the ultimate goal of product growth, is to have the cohort stick around and retain such they compound, well, then at some point, just mathematically, the majority of users will be old. You have to balance the needs of the old users against the needs of the new users. And if you just look at DaU and you look at overall take rates of different product features and stuff, you're going to be overemphasizing the use cases of the older users. Well, now maybe the new users are being alienated because their needs are different.

0:38:55 - (Eric Seufert): They have less knowledge of the product. Right. Because they were the hardest to reach. You know, they took the longest to onboard. The people that onboarded from day one of the launch of the product. Well, they probably actively sought out your product and they, they had high intent. Well, the people that you're just now reaching two years later, unless they just like aged into relevance or something like, yeah, it's like, well, I only found it now because, you know, you had to show ads for two years in order to finally reach me and to understand that, you know, that sort of like interesting component or the demographic component. So it's just understanding.

0:39:29 - (Eric Seufert): Okay, well, now you're stuck with the conundrum of like, well, who do I service? Well, ideally everyone. Ideally everyone gets the optimal experience. It's not like you just forsake the older users in favor of the new users because the older users, almost by definition, are going to churn earlier. Right. What you want to do is give everyone the best experience. So that is a segmentation exercise. Like, okay, well, new users are getting an onboarding that has been optimized. For the new users, the onboarding is what they'll see. The old users aren't seeing the onboarding anymore. So if I optimize the onboarding two years ago, never touched it again.

0:40:02 - (Eric Seufert): The new users, who probably have a different profile, different makeup, different level of intent are seeing something that was optimized for people that just looked very different. So that's something that I think most people understand. It's like, well, I need to constantly be optimizing onboarding. That's very rare to see a company that's one and done on that. But that applies to every step in the user lifecycle, not just the onboarding.

0:40:22 - (Eric Seufert): Right. So that the existing users, the old users, well, they're not going to see the onboarding again, but they're also, but they will be engaging with like the core experience of the product. Well, the new users, maybe that's not the sort of like best fit for them. And so like, how do I figure out, how do I, how do I segment them on the basis of the cohort age? And I think people just generally don't do that. Now, for some products, like a Spotify, I don't think it matters as much, but there are products for which it would.

0:40:46 - (Eric Seufert): Right. And so you just need to be cognizant of that. It's like, well, what is the composition of my DAU? Not just what the number is and number go up, right, because number could go up just because I'm increasing my ad spend and I'm bringing in more DNU every day and they're going to end up churning out, how do I compound and like, so number go up. If the old users are kind of staying flat, they're not churning anymore.

0:41:06 - (Eric Seufert): And I'm bringing in the same number of new users every day, or ideally growing that because I just get better unit, unit economics now than I did before. But you need really understand, like, what the composition of the DAU is. Make sure that you're serving those different cohort ages with the experience. That's, that's, that's the most relevant for them.

0:41:22 - (Matt Widdoes): Yeah. I think that retention decay metric as well as the forward looking predictive revenue curves are also really important because to that point, not all users are created equal. You may have indication of like a new batch, new, you know, a weeks cohort of new users that are acting completely differently in their early indicators of success. Maybe there was a UI/UX change or an onboarding change. Maybe there was a complete change in strategy at the top of the funnel.

0:41:55 - (Matt Widdoes): Maybe there's any number of things that could lead to that. And I think so often people just are super focused on that DAU or the DNU number and just like, all right, cool, users, users, users. But at some point you have to take a step back and understand that daily mix of who's coming back and that should be consistent over time. And you're getting more data, you're getting a new data point every day, every week, every month with your oldest cohorts. But you want to make sure that those are tracking and you should see that over time those are stacking and getting better so that you're retaining users longer, they're reaching revenue faster, and that ultimately they're retaining and more active within the product, meaning that you're providing more utility, presumably, if they're, if they're getting value out of the product.

0:42:36 - (Eric Seufert): Yeah, absolutely.

0:42:37 - (Matt Widdoes): You run a pre-seed stage VC fund now focused on mobile tech generally. I'm curious, key criteria that you look at before you invest into a mobile startup?

0:42:49 - (Eric Seufert): Well, so, I mean, I invest kind of like across a broad spectrum of categories. Right. So, like for ad tech or publishing tech, what I call publishing tech, it's going to be, you know, sort of different standards than just content. Right. Content is an interesting one. I mean, you know, the mobile content space has been very challenged ATT and so, you know, what I'm looking for, like with a gaming company or like just a utility app company is like, have they done this before? Have they been successful here? And like, you know, has this team in particular, has every member of the founding team, you know, been a. Been a part of a success story in like a way where they had meaningful contributions? Right. Because I think there's just a particularly on modernization. Like, that's one thing that I must see in the founding team. It's like someone who's able to explain to me how the product will monetize to a point that supports sustained user acquisition.

0:43:40 - (Eric Seufert): That's something that's generally missing from mobile gaming teams, too. It's like, well, you could have a product manager who can explain the core loop and why players would be motivated and how they'll progress through on some sort of progression journey or the competitive component of the game, but they can't explain how they're going to monetize it or how the economy is going to balance. And that's just so fundamentally important, especially for gaming, because games live or die by their ability to spend money on user acquisition.

0:44:09 - (Eric Seufert): I think for ad tech in particular, I'm really looking for stuff that's thematically consistent with my view of the world, which is that we are seeing the kind of steady drumbeat of privacy interventions, privacy restrictions, and are you building stuff for that world? Are you building stuff for the old world? And a lot of stuff that's built for the old world tries to enable loopholes or workarounds. And I just don't think that's sustainable. I want to see something that brings value to advertisers while recognizing that there's just less signal, less data to use now to make spending decisions, budget decisions, optimization decisions, publishing tech. I mean, that's kind of what I call all of the tools that you integrate into the product to help it retain better, monetize better.

0:44:51 - (Eric Seufert): A lot of that stuff only exists because of the recognition of the change in the privacy landscape. So what I'm really excited about now is anything that manages personalization, like at scale, like how do I give this user the best sort of optimal experience for their own use? Case that I think is just a critical piece of infrastructure in a more restrictive privacy environment, but also anything that utilizes machine learning to just improve retention, improve engagement. I think that kind of stuff is what you need to do to compensate for the lack of fidelity on the advertising side.

0:45:22 - (Eric Seufert): So that's the kind of stuff I'm excited about. Stuff that takes signals that are robust against future privacy restrictions, stuff that utilizes behaviors as an input and not just data that is sourced from cross party contexts. All that stuff I find to be super interesting and I think there's just an immense amount of value to be unlocked there.

0:45:45 - (Matt Widdoes): And curious. Key advice, number one piece of advice, or top few pieces of advice you'd give to somebody starting a venture in consumer tech or mobile.

0:45:53 - (Eric Seufert): Yeah, I mean, consumer is just like consumers, you know, it's a very specific profile of products, right. And, you know, I think there's a couple of different ways to launch a consumer startup. One is to just do the virality route, right. Is to take the wrong example and, you know, to be fair, those, you know, be real air chat clubhouse. You will find investors for that. You know, if they see, they see, you know, a hockey stick graph, they just tend to extrapolate it.

0:46:18 - (Matt Widdoes): Want the next TikTok, you know, indefinitely.

0:46:21 - (Eric Seufert): Right, well, so that's a. But that's a misapprehension. TikTok spent a billion dollars requisition in 2018 in the United States alone.

0:46:28 - (Matt Widdoes): Wow.

0:46:28 - (Eric Seufert): That wasn't a viral story. I mean, there was virality.

0:46:31 - (Matt Widdoes): Yeah, we heard how they launched in France and same thing. It was like all pre-programmed massive media budget. Like, it was made to appear viral, but in its nature, it was paid.

0:46:42 - (Eric Seufert): Yeah, I mean it was, it was absolutely paid. I mean, they were the biggest advertiser across all the scaled social media platforms, which is kind of ironic, but you know that. So I think people take the wrong lesson from that, which is they don't understand the reality of it. But that was not a viral story, that was a user repetition story. Yeah, I mean, so, but I think, like, it's just this very boring, very sort of like, uninspiring, you know, playbook, which is like, iterate, be systematic, understand. It's really an exercise in, like, analytics and understanding, you know, the sort of behaviors of people based on the information they give you through their engagement and iterating and making improvements. Right. And, like, that's not exciting.

0:47:20 - (Eric Seufert): It's boring and seems tedious because it is. It's supposed to be. And sometimes, you know, investors are like, well, you know, I need to see you growing 100% month over month before I get interested. And it's like, well, how could I possibly do that without either just throwing money at this that I don't have any confidence in or going viral, right. So it's, you know, unfortunately, like, self financing from, like, a baseline of zero means you're paying yourself, like, very little or nothing and throwing your savings into this.

0:47:52 - (Eric Seufert): Right. Or you just try to juice the numbers by going viral or whatever and then hope that you can catch up with the engagement later. So those are kind of like the options. You know, I just. I just think, like, one thing to consider is mobile is, and I think will be for our lifetimes, the primary way that human beings engage with technology on a daily basis. Right? Like, it's tempting to get excited about new hardware form factors. Now, there aren't that many. It's basically, you know, but like, the AI, what was it called? Like the Rabbit or whatever, like the AI gadgets. I don't see that, like, the smartphone is it for our lifetimes. And so.

0:48:28 - (Eric Seufert): But there's a smartphone for every person on the planet. And so if, you, you know, if you want to reach essentially everyone that has the ability to use a digital product, well, you can do that with a smartphone. And so, you know, things may get more challenging across any given dimension with privacy. It was, you know, advertising and distribution, but nonetheless, they're out there. The people are out there, the users are out there.

0:48:51 - (Eric Seufert): And so, you know, chasing, like, the latest trend VR or like, AI gadgets or even just AI generally, it's like, well, okay, maybe that works out, and I'm sure it will. And there's already, like, a lot of companies that are very valuable, that are AI first companies, but mobile is not going anywhere. There's always going to be a massive number of people that use a mobile device, a smartphone, every single day. And so I think you'll always be able to build a big business on.

0:49:19 - (Matt Widdoes): It until we get to contact lenses with AR, which is very, again, not brain implants. Brain implants. God-willing right, cool. Well, maybe for our last point here, I'm just curious, when looking to the future, any significant shifts or trends you see right now on mobile that you think are interesting or compelling or kind of emerging?

0:49:41 - (Eric Seufert): Yeah, I mean, a lot of this stuff, I mean, look, mobile is mature, smart. The smartphone ecosystem is mature. Mobile ecosystems is mature. I think where you're seeing a lot of change is being driven by platform policy and regulation. Right. So I think the most, like, probably the most meaningful change that we've seen in the last couple of years is the shift, and really it's like the last two is the shift of monetization off of, away from the App Store and onto the web.

0:50:05 - (Eric Seufert): Right. And then a lot on, like the, you know, a lot of utility apps are now they're driving like the vast majority in some cases are all of their user acquisition spend to web destinations, getting the user to register there, pay there, and then having, and then, you know, instigating the app download there. I think that's a trend that will continue. I think it's really exciting for a number of reasons. One, sidestep the commission.

0:50:26 - (Eric Seufert): You just have more money to spend on marketing. But also the web is just a much better environment for experimentation than the App Store's are. The App Store storefronts are very primitive in terms of the opportunity that they offer for customization testing, dynamic offers, dynamic pricing, dynamic bundles. I mean, if you're selling a digital product, you should have the ability to experiment endlessly.

0:50:48 - (Eric Seufert): You just can't do that in the mobile app stores, right? It's digital. You can do whatever, you could do whatever you want. You could sell whatever you want right now. You pair that with generative AI, then I think there's some really exciting stuff that you could do with web-based storefront, endlessly customizable. You could experiment exhaustively, personalize the product that you're offering right there on the web, and then you just render it in the app and you've got persistent identity across those two things because they had to sign up and get a user id when they onboarded on the web. So that stuff, I think for gaming, it's relevant for gaming, certainly, and there's a lot of gaming companies utilizing that. But I think this kind of enables whole new categories to emerge that are difficult to map to the existing mobile ecosystem. So I am really excited about that.

0:51:32 - (Matt Widdoes): Yeah. And I would imagine, I mean, Apple has been notoriously slow with any sort of innovation on the ad front or App Store front. I mean, they're like, hey, we have video now. It's like, okay, big innovation over the last ten years, but, yeah. And I think Audible, as an example, is way ahead of their time on forcing purchase through web. And for all the points you described, which I would imagine at some point the App Store's will have to kind of innovate or get crushed that way because they're cutting off a huge percentage of their revenue by forgoing that in the App Store. So it'll be interesting to see that play out.

0:52:10 - (Matt Widdoes): Well, thank you so much again for the time, Eric, always great chatting.

0:52:13 - (Eric Seufert): Yeah, likewise, Matt. Thanks for having me.

0:52:14 - (Matt Widdoes): Yeah. Look forward to next time.

0:52:16 - (Matt Widdoes): That was my conversation with Eric Seufert.