12 AI Growth Lessons for Subscription Apps – Phil Carter, Elemental Growth

Welcome to the subclub podcast. A show dedicated to the best practices for building and growing app businesses. We sit down with the entrepreneurs, investors and builders behind the most successful apps in the world to learn from their successes and failures. Subclub is brought to you by RevenueCat. Thousands of the world's best apps trust RevenueCat to power in app purchases, manage customers and grow revenue across iOS, Android and the web. You can learn more at RevenueCat.com. Let's get into the show. Hello, I'm your host David Barnard. My guest today is Phil Carter, an independent growth advisor and angel investor focused on helping consumer subscription companies scale. On the podcast I talk with Phil about how AI can turn your onboarding from a chore into magic. Hyper-personalized experiences that drive both retention and revenue. And why your value to noise ratio matters more than how many features you ship. Hey Phil, it's nice to have you back on the podcast today. Yeah, thanks for having me back. I'm really excited for this one. So you are now one of our longer-tenured guests. This will be your third appearance. I think you're in the top three now of repeat subclub guests. And always excited to have you on because you work with some of the top apps in the industry on a daily basis. And have just such great insights and know how to like condense it into those like really solid takeaways and frameworks. And things like that that I think people really enjoy and can really take things from. So appreciate you coming back on. Yeah, well it's a real pleasure. Thanks for having me back on again. I remember when I was first starting my business, I listened to the subclub podcast religiously as you know I mentioned that to you when I was getting started. And I remember hearing Tomas Patiet and Eric Sufer on the podcast multiple times. And so I'm glad to be among the ranks of repeat guests at this point. So several years ago when you started your consulting career, you came up with this subscription value framework, which you did a keynote at AppGraphen. Daniel a couple years ago, you know, you published on Lenny, Vritsinski's blog about it. And I think it's really taken on a life of its own. But interestingly, you came up with all of that before AI really started changing the app industry. So I wanted to talk through some of your top lessons that you've learned over the past few years about how AI is changing the game for subscription apps. And then of course, it's really helpful to talk through it in your subscription value loop framework. But to get started, let's just start with that like 90 second, maybe three minute overview of what is a subscription value framework so that it can anchor the rest of our conversation. Yeah, sure. So I guess the origins behind the subscription value loop framework really happened for me when I was leading the product growth team at Quizlet. And at that point, I remember having lots of conversations in particular with the VP of product there who was nith and gupta at the time. And we would talk about the difference between value creation and value capture. And the rationale was that core product teams are primarily responsible for value creation. They're the ones building new products and new features and fundamentally creating value for users, which in Quizlet's case was the students and teachers using the product. But then value capture depends on a different team with a different set of skills and expertise. And so that was the growth product team that I led focus on things like user onboarding optimization and paywall optimization and figuring out the right pricing and packaging strategies. But there was no framework for this. And so I would just talk informally with nith and other members of the product team. And then at some point, members of the marketing team and other parts of the organization about which team should be responsible for which parts of this subscription value loop. And what sort of levers we had at our disposal to optimize each step in that loop. And over time, I added a third step, which was value delivery because there was this gap between core product creating value for users, growth product monetizing that value and converting it into revenue. And then in the middle, you have marketing or in some cases sales teams who are distributing the product and its features to users, whether that's through paid channels like met and Google or through organic channels like word of mouth and SEO. And so what that led to once I started elemental growth, which my growth advising consulting firm was this subscription value loop framework. And so to sum it all up, the subscription value loop framework. Posits that every successful subscription business is centered around a unique and enduring core value promise at its core. And then you have three steps around that core value promise that drives the success of the business. Value creation, value delivery and value capture and the more effective and efficient a company gets at optimizing each of those three steps. The higher your LTV over cap ratio gets, the lower your payback period gets and the faster the company is able to grow. Such a great way to think about things. It's funny because I forget if I tweeted it or wrote a blog post, but I was kind of floating around very similar ideas around the time that you positioned it. And I was like, oh, yeah, that's perfect. That's exactly what I've been thinking. Like you got to create a great product that users actually care about. You got to get attention. You got to get in front of them. And then you got to actually monetize. And so I think those are the three words I was using, like attention monetization and value or product or something. It's just the core of a digital software business is those three main things. Can you get attention? Can you create something really valuable and then will users actually pay for it? And then you get into all the how do they pay for it? And how much attention you can get in the cost of attention and like all those kind of things. But it is just such a nice high level way to think about all these things. So today we organized your lessons of past few years as AI has been changing things into your top 12 tips. And we're going to do four tips around value creation, four tips around value delivery and four tips around value capture. And I thought it'd be fun to just just make it like a we're going to do an audio listicle. And it should be fun to kind of follow along and you can jump around and we'll probably clip some of these for YouTube and stuff. But let's just start with value creation. So again, core to creating a great business is like you got to actually make a great product. And then it is fascinating that AI, there are just so many ways that AI can help you build a better product. Whether it's integrating directly in speeding up engineering resources, all those kind of things. AI is coming to play in that value creation. But let's talk to your top four tips. So number one, providing a magical first session. So how does AI help you in that onboarding, which is so critical to success these days? Yeah, well, first I'll start by saying there's a reason value creation is the first step in the loop. And that's because without creating value for your users and subscribers, nothing else really matters. And so you've got to start there without that, nothing else is going to work. And the first tip that I have specifically for how AI is changing value creation is this idea of creating what I sort of jokingly refer to as an autonagical first time user experience. So what do I mean by that? What I mean is, if you think about consumer tech for the last 15 years, really since the first iPhone came out in 2007-2008, there hasn't been that many sea changes in terms of new consumer technologies that deliver experiences that consumers have never seen before. If you think back to, I graduated from college in 2008. And the first time that I used an iPhone, it was like an eye-popping experience because I'd never seen anything like the swipe screen and the app interface. It was just really, really compelling. Fast forward to late 2022. And that's when a lot of people were using Chattie B.T. for the first time when it hit the public mainstream. And it was one of these experiences that was so mind-blowing that you had, I remember being at a holiday party, the company I was working for at the time, and all anybody could talk about was Chattie B.T. And the CEO gave a speech that was heavily influenced by a Chattie B.T. prompt that he had put in the night before. And so it was in the zeitgeist, and I think AI has created this window of opportunity for consumer-aft developers to create these magical experiences that can hook users and grab their attention and ideally convert them into trials or even directly paid subscriptions within the first session, and sometimes even within the first 30 to 60 seconds. And the reason that's so important is you guys reported data on this and you're a state of subscription afterport, right? You guys first started doing that report in 2023, and at that point, 2023/24, it was like 70 to 75% of trial starts were happening on day zero. That's now over 80% and in some categories, it's almost 90%. So it shows that you really have to convert a user in that first session or you're not going to convert them at all. And so one example that people have pointed to historically of a company that provides a great first time user experiences do a lingo, right? Like in the first minute, they're not just asking you questions about language you want to learn and being pedantic about language learning strategies. They're just getting you into a language learning experience that's immersive and you're earning points and streaks and badges and it's compelling and it's gamified. And so that's how they reinvented the language learning space. But now there are new AI native companies and we work with one called Tolin that's sort of an AI alien companion that's sort of your best AI friend. And they've just done such a good job of in the first minute or two, you know, you have this oracle who asks you a bunch of questions about yourself. Based on your responses to those questions, you're then matched with a Tolin that has its own unique personality and character dimensions based on the language. off of your personality. And it explains the rationale of why you were matched with this toll, and it has its own sort of character to it. And then you're very quickly put into a voice-to-voice chat with this AI creature. And it's memory and its tone and its recall is just so compelling. It feels like you're talking to an old friend. So that's the type of magical experience I'm talking about that you can't really experience anywhere else. And the reason that's important is that's going to have a very high degree of influence on whether or not a user comes back and keeps using your product as well as whether it is decided to pay money for it. Yeah, and the toll in experience is just such a great example. I mean, it's like they're using AI at multiple levels because you are actually talking to an AI. But what I found so magical about that experience is that it was the most fun onboarding I've ever had. Because it asked you questions and then it would tailor the response. And this is what's so great about AI that you can do this now in near real time. That when a user says they care about this, then the response can be, oh, yeah, that's really cool. Or, you know, and can like make a joke about what they're talking about or something like that. I forget the exact example. But my daughter, I had my daughter go through the onboarding of toll in while we were driving in the car. And I remember cracking up. Like it was funny because they were able to take her responses and then kind of like in onboarding, you're really trying to convince a user there's some value. You know, they've come to the app for some reason and you're trying to get them over that bridge of like, okay, this is going to be something I want to use something that's worth paying for. And when it can be that fun and that meaningful, it just really creates a different experience. So in addition to like the way toll in does it, I think there's so many ways now that and we're seeing it in the industry, apps are able to do that level of like customization and make onboarding way more dynamic and way more fun and not just fun, but useful makes it feel like the products are already doing something for if you can deliver value in the onboarding by like taking one of their questions and like, you know, helping solve a problem or whatever it's just so powerful. And the way AI enables that is incredible. And it is such a new thing, you know, three or four years ago, do a link was doing a lot of customization in that onboarding, but it was, you know, hundreds of engineering hours and a lot of like set rules and maybe a little machine learning sprinkled in and you need a really big team to do that. And now small teams are able to do some really cool stuff. They say, AI and that onboarding is a great place to be experimenting with it because of how important it is to nail that first session. >>Yeah, I completely agree. And one last note I'll make on this one is in a world where AI has made it easy for anybody to vibe code an app in a weekend. The importance of personality and standing out from all the copycats has never been greater. And so for an app like Tolen, you can just, you can feel the craftsmanship and the, the soul of the company coming through in those first 60 seconds in a way that you just don't see in many consumer or prosumer software products. And I do think that's a huge part of why they've been so successful because they have a very opinionated product that they built. >>All right, let's move on to number two. Delivering hyper personalized experiences. We can already hint at this on onboarding, but it's great because it can go through the entire product. So what are some examples you've seen of doing that? >>Yeah, so when I say hyper personalized, what I mean is if you think about how consumer software has evolved over the last two decades, you know, I mentioned the launch of the iPhone earlier in 2007, 2008. When the iPhone first came out, it was like you download an app and everybody got the exact same experience. And then maybe 10 years later, 2015, 2017, there was some degree of personalization, but you maybe got put in a bucket. And so to use health and fitness as a category, maybe you got bucketed into somebody who, you know, was just trying to get off the couch versus somebody who has already sorted an average gym goer or runner versus somebody who's training for their next marathon. Now with apps like Runa in the running space or ladder in the strength training space, you're seeing more and more of what I call hyper personalization, which is you've got your onboarding quiz. And there are other reasons to do an onboarding quiz, which is you get all of this data that you can use to optimize paid marketing spend, especially if it's in a web app flow. But there's this core product benefit, which is now you deeply understand this user. And with AI, you're not just said, "Making users into a few buckets." You can actually create almost an in-of-one experience for every individual user based off of their unique responses to these questions. That's just the first time user experience. And then over time, to stick with health and fitness category, as a user logs more and more runs on Runa, or they log more and more workouts on ladder, the degree of personalization around how frequently this person should be working out, how long they should be running on Runa or how much weight they should be lifting on ladder. It's adapting iteratively based off of the feedback it's getting from each of your workouts. And so this concept of hyper personalization is another way that in an increasingly competitive app ecosystem, the best apps are distancing themselves from competitors. Because when a user feels like you're speaking directly to them and your product is constantly evolving and adapting around their specific needs, they really feel heard and they're more likely to retain and to pay for your subscription. And it just creates that extra bit of value. I mean, that's the whole thing with Runa. It's such a great example of what an incredible company. And in part because they actually started pre-AI, before any of this was possible. And when I first talked to the Runa team, which I talked to them super early in their journey, I was kind of thinking as I did with ladder, like, this is such a competitive category. There's so many winners already. Strava's already there. Like, what are you going to do? And what they did was they created a really valuable, compelling experience. And even from those early days, they had those levels of personalization. And then as AI made it more practical, they've just gone deeper and deeper on that. And it just created such a valuable product to people. And as you said, with ladder, being able to progressively overload and know that next right weight, the next right sets, and things like that, it's just such a powerful mechanism of value delivery to get that level of personalization and not just have those buckets of users where everybody's getting the same kind of plan, the same experience. So yeah, it's just such an incredibly powerful thing. Yeah, I completely agree. And one other thing I'll say, I had Rick, your CMO, on my podcast a couple weeks ago, which was a great conversation. And one of the themes we talked about was the nicheification of consumer apps that's being driven by AI. Right. For two reasons, one, AI is flooding the app stores with more apps, which is creating more competition. So you sort of have to niche down as a startup to be competitive. But two, AI is dramatically reducing the cost of software development, which means it's viable to go after smaller niches that previously just weren't worth targeting. And so I think run and ladder are two great examples of apps in the fitness category, which is one of the most crowded competitive categories there is, who at least initially focused on underserved niches to get a flip hold. And run as case, it was, okay, Strava is already the dominant app out there for people who just want to do some measure of running and track their performance. But for people who really want to train for their next race, for serious runners who are training for their next marathon or their next 10k, we're going to put together a customized dynamic plan that adapts to your running performance over time. And that's something that Strava wasn't doing. So run it goes and they do that and they create a great first time user experience and automatically onboarding flow. And they hyper personalize to their users. The stuff we've been talking about, they win this segment and eventually Strava comes along and buys them because they unlock this whole new segment for Strava. Similarly, ladder has this whole matrix of populations of people who want to get more fit, whether that's at home, or in the gym, whether it's men, women, and they've figured out how to identify the market size of each of these different pockets and then target the right coaches and personalities to attract the users in those pockets and then build a product that hyper personalizes the experience around each of those personas. And so more and more, I do think there's just this huge opportunity. It's sort of the classic riches and niches, right? It's niche down and find a very tightly defined early adopter base that you can get traction with. And then eventually you can expand beyond that. But it's always better to start small and really focus on one user persona. Yeah. And I think that's a huge opportunity too that we're going to see really great companies built over the next couple of years and probably already have, I can't think of any off the top of my head. But like with AI, you can do so much more personalization. You can't go so much deeper. You can create product experiences you haven't created before. So what's that new experience that wasn't possible before? Actually, I just thought of the perfect example. Cal AI blew up because they used machine vision and AI to capture your calories in a much more straightforward way. And is it perfect? No. But is it a great experience? Yeah. And so I think we're going to see a lot of great companies built thinking through those kinds of things. Like what level of personalization, how deep can we go through AI that just was never possible before? So I'd encourage folks. And then if you're an existing product that's doing really well, It's like you need to be think, you need to disrupt yourself and be thinking about like, what's the Cal AI to my fitness pal and you go build that instead of letting somebody else come into the market and then of course my fitness pal into buying Cal AI. And so I think companies big and small should be thinking about all these new unlocks that happen in value creation. Thanks to AI. Completely agree. All right. 10 number three under value creation. Getting extrinsic triggers to build habits quickly. What do you mean by that? Yeah. So what I mean by that is we've already talked a bit on this podcast about how competitive the api-gustism is getting and a corollary that is the average consumer or prosumer just going to bombard it with information and ads and lifecycle marketing emails and push notifications all the time. Right. There's sort of this dystopian future I can imagine where the average person is just getting thousands or tens of thousands of happen preshens every day. And so what does that mean? What it means is, you know, if your product competing for consumer or prosumer attention, as you're starting to build a habit with a new user, it helps to have an extrinsic trigger on the user's mobile device or on their laptop or whatever device they're using. It just reminds them to use the product. Near I All Rights this book called Hooked where he talks about the loop that helps build habit formation among consumers. And it starts with an extrinsic trigger, which could be an email, a push notification, a widget. But eventually as the user goes through enough loops and builds a habit, it becomes an intrinsic trigger. And then they don't need the reminder anymore because in their head they just know, okay, if I want to do my workout, I go to ladder or if I want to listen to music, I go to Spotify. So the tip here is, it's not going to work for every product. But if possible, find a way to build an extrinsic trigger into your target user's workflow that reminds them when they should be using it. So a couple examples here, they're both more prosumer versus consumer, but two products I use every single day now are granola, which is my AI note taking software. And then whisper flow, which is voice to text dictation. And both of those apps have these very subtle ways of just reminding you that it's time to use their product, right? In granola's case, it's linked to your calendar. And so every time I have a client call or some other meeting coming up, I get a little notification on my desktop and on my mobile device letting me know that the meeting is about to start, which is helpful in its own right? So it doesn't come across as annoying because it's beneficial to me to know what my next meeting is and who it's with and when it's coming. But then on top of that, it reminds me to turn on granola in cases where I want to be taking AI powered notes in the background. And then similarly with whisper flow, so whisper flow for those who don't know is voice to text. And so you can press a button in my case, it's the function key or the FN key on my MacBook. And if I press and hold that and then I just speak into my MacBook Pro, then it will convert my voice to into text. And it turns out that's a much faster, more efficient way to write long-form content, whether it's me writing a LinkedIn post or a blog post or I'm in Cloud Code trying to do research or analysis. But they have this really little widget on the bottom of your screen that's very subtle and non-descript. It's not getting in the way, it's not a distraction. But it just reminds me, oh yeah, whisper flow is there. And so as I was building at this point, I'm using voice almost as much as I'm using typing in order to put my ideas down digitally. But when I was first getting started, it was a very difficult habit to break. Because talk about 10,000 hours, I mean the average knowledge worker is spending huge amounts of time every single day typing. And so it's a habit that doesn't die easily. But that little widget just helped me get into the rhythm of like, okay, whenever possible I should be using voice because it's going to save me time. And then this is where I give the whisper team so much credit because they've taken this little widget, which is this little black sliver at the bottom of your laptop. And now kind of like the dynamic island. So whisper flow has turned at this little black reminder of its existence into its own island that will pop up and give you tips and best practices to sort of extend the amount of value you're getting out of the product. And so it's just this really elegantly designed way that they are training their users, how to use their product, but then also as they climb the learning curve, how to get more and more value out of it. Yeah. And whisper flow also, they have a built-in keyboard. I don't know if you use that, but I've been using that more and more. And it changes the way the regular keyboard looks and works because you can tap really quickly to get into whisper flow, which works so much better than the Apple's dictation. And so now I find myself using it a ton. And it's right there in my keyboard. And so similar to like, you know, they do have a live activity that will pop up and things like that that you can do. And that gets us really AI, but it's maybe AI apps using these kind of features to do extrinsic reminders, but a great tip to be looking for these kind of platform features that you can use effectively to deliver those extrinsic reminders to use the product. And yeah, whisper flow is such a fantastic product and such a fantastic example. Yeah, definitely. All right. Number four, staying ahead of the curve with rapid innovation. Yeah. So this one's sort of like, blindingly obvious. But I think it's sort of like when I wrote the first version of this for Lenny's newsletter, that this was a couple years ago, and I wrote the top 10 tips for growing consumer subscription apps, tip number one was build a great product and it's sort of like, well, duh. Like, nobody needs to be told to do that. But if you don't put that number one, then you're missing a very important point because without that, nothing else is going to matter. I think similarly now with AI, this has just become so important. Casey Winners and others have written about how product market fit is a moving target. It's not like one line and then once you cross that line, you permanently have product market fit because the market is evolving and your competitors are pushing the bar forward. And so what that means is with AI, the pace of innovation has become so fast that if you're not shipping new products and features every month and increasingly every week or even every day, then you're falling behind, which means you're going to lose product market fit relative to your competition and very quickly you'll become obsolete. I think the textbook case of this recently that everyone's talking about is clot and specifically clot code. I mean, the pace at which they are shipping new features and optimizing the performance of their existing features is absolutely mind-blowing in a way that wouldn't have been possible pre-AI. But it now is possible and it's also just, it's absolutely necessary for survival. And then one other example I'll throw out is study fetch. So this is another company we've been working with for about a year now in the edtech space. And they're primarily serving college students, but they just have lots and lots of different AI-powered features. So you can use study fetch to record notes in the classroom similar to what a professional might use granola for. You can use it to drop in a set of handwritten or digital notes and convert it into an outline or have its synthesized important takeaways to help you study for an upcoming exam. You can use it to, let's say you're a medical school student, right? You can like upload a diagram and have it extract the most important terminology for you to learn as part of say a biology class, right? So all these different use cases and the advantage of shipping so many features so rapidly, one is you're creating a more dynamic core product experience that can lead to longer-term user retention and a greater probability of subscriber conversion. But two, if you ship in public, which more and more of these startups are doing, right? So you're constantly letting people know on TikTok, on Instagram, on Discord, what your new product features are and why they matter and you're having people react to them in real time. It also has a user acquisition benefit. And so this can be a good segue into Valley Delivery because it's sort of a two-fer in that you're improving Valley creation while also creating Valley Delivery opportunities at the same time. Before we get to Valley Delivery though, I did want to stop and say this speed of innovation can be a trap. So I wanted to get your take on how to avoid it becoming a trap because with it being so quick and easy to throw a feature on, it's easy to turn an otherwise great software product into this like catch all or way too many buttons, too confusing. I mean, this has been kind of a long-standing challenge with iPhone app specifically, mobile app specifically. It's a screen real estate, it's so small. The use cases, you're often like in line at a grocery store or whatever. And so yes, you can move fast and build a lot of features, but it can be a trap if you're not thinking through carefully how you build those features, where you put them in the experience, how that relates to the rest of the experience. And for some apps, creating a second app instead of adding more features to the existing app actually probably does make sense. So how do you think about avoiding that speed of innovation as a trap? I like to think about the value to noise ratio. So what I mean by that is I think a lot of not just consumer products, but tech products in general fall into this trap of just maximizing value. And a fallacy that a lot of companies fall into is like well more products and more features equals more value and therefore we should just ship products and features as fast as we possibly can. But the reason that's the wrong way to think about it is the bottleneck is the capacity of a human brain to absorb the most valuable parts of your product. experience. And as your product gets more bloated and complex, sure, absolute value of the product, if you just sum the number of products and features goes up. But the amount of noise and complexity also goes up, which means your value to noise ratio may drastically decline. And so what that means is, as you're adding new products and features, you need to be constantly talking to users, analyzing user engagement rates, the correlation between feature usage and long term retention and subscriber retention rates to understand what your hero features are. And then pruning away the products and features that are no longer adding any value so that your value to noise ratio stays as high as possible. Yeah, that's such a great tip. All right, well, let's jump into value delivery. So your first tip there is to leverage product-led growth loops to amplify viral word of mouth and SEO. Yeah. And so this is another one that isn't strictly enabled by AI, these product-led growth or PLG loops have existed for a long time. But I think they've become even more important in the world of AI because competition has become so fierce and the largest paid ad networks like Meta and Google have gotten so crowded, which means they're getting more expensive. And so what I mean here is, when I say product-led growth loops, I mean things like personal viral loops, where people are inviting colleagues or friends or family members to a product to make their own experience better. So social networks are a great example of this, the more people who are on Facebook or Instagram, the more valuable it is for every individual user. There's also social viral loops, which is I want to tell people about this new product, whether offline or on a social network like TikTok or Instagram because I want to get social credit for being an early adopter and being in the know. There's financial viral loops, which is I'm actually incentivized through dollars and cents or free subscription credits or AI credits or some other mechanism to invite other people to the product and I get a tangible financial benefit out of it. And then there's also content-driven organic loops like SEO or increasingly AEO, which is answer engine optimization, that are generated by content spun off of a product. And so an example that I use for this one is Gamma. Gamma, they just raised their series B recently from Andrews and Orowitz. I think they're at north of a $2 billion valuation at this point, building AI-powered presentations and websites. And they've grown extremely cost efficiently with very little paid user acquisition because of these PLG loops. And part of that is they have this inherent advantage, which is their product naturally spins off all of these different presentations and websites that people are creating using Gamma. And so some percentage of those are going to end up getting shared, either internal to a company or team or at a conference or a presentation. And so there's all sorts of personal and social viral loop activity happening through the Gammas that are being created. And then there's also SEO and AEO acquisition benefits because as more and more of these Gammas get indexed by search engines and are now discoverable by other users or they start to get referenced in responses to LLM prompts on chat Gbt or cloud than other people can discover Gamma. And all of that is free user acquisition that the company doesn't have to spend money on. Are there specific ways that you've seen companies use AI specifically to help do this better? So either better sharing screens, more engaged loops, anything like tolin, where that kind of experience becomes better because you're using AI. Have you seen anything like that? Yeah, I'll give you an example. So one is going back to value creation. When you create an auto-magical product experience that people have never seen before, they're naturally more likely to share. And so tolin's a great example of, okay, I go through the onboarding flow, I get maps with my toll in. They have this great infographic that shows the toll in's personality profile and how it maps to yours. That's a viral artifact that people are likely to share. So that's one example. Second example is, and we'll talk about this as part of value capture, more and more AI products are using hybrid monetization models where you've got a subscription, but then you also have to purchase additional AI credits even on top of your subscription if you're a power user. Well, that leads to an opportunity around incentivize referral programs. We're now if I invite a friend or colleague, I get additional AI credits that I can use for my own product usage, but it's a natural hook for me to invite more people. And then a third way that AI is directly impacting these BLG loops is through content-powered growth. And so we talked about search engine optimization, which is the sort of legacy way of acquiring users through Google and other search engines when your content gets crawled and indexed by those search engines. Well, now there's an equivalent with LLM outputs that are increasingly taking share from search engines like Google. And so there are some specific tactics you can use to optimize your likelihood of appearing in a prompt response in chat with ET or Clod. One specific example of that, this is less true now than it was six to 12 months ago, but Reddit overindexes as a source of data that these LLMs use for responses. And so like one early tactic was go-do gorilla marketing on Reddit in order to get a following and build a community there. And then the more your product shows up in Reddit, the more likely you are to appear in these LLM results, which helps drive AEO. Well, that leads perfectly into the next tip, number two. Investing in community for UGC driven growth. Yeah, so again, this is something that companies have used long before the advent of AI or the mainstream adoption of AI. But it's become so important because people are still trying to figure out how to use these tools. And I think as as knowledge workers and as people who work in tech, we often overestimate the degree to which mainstream consumer populations are fluent in how to use even chat GPT or Clod, let alone all of these longer tail AI products. And so the example I use here is notion. I had Rachel Hepworth who is formerly the CMO of notion on my podcast when I was first getting it started last year. And she talked about how crucial community was for notion when it was first getting started because notion is an unusually horizontal product experience. Unlike most consumer or prosumer tools, it didn't launch with a single point solution. It launched as essentially like a wiki replacement for something like it lassied. And so as a result, they really relied on influencers and early adopters who were creating content on YouTube or other online platforms to be able to educate other potential users of notion. And then when they became an early adopter of AI with notion AI in 2023, they went back to the same playbook. And they had these early AI adopters on YouTube, on TikTok, on Instagram, on Discord, just sort of talking about how they were using notion AI, publishing templates or databases that they were using tips and tricks for how to get the most value out of it. And that really helped to drive a lot of organic adoption of their new AI products. But it also helped to improve new user retention because they got more new users over the activation hump of understanding how to use the product and get value out of it. So they didn't just turn in the first 24 to 48 hours. Yeah, I've talked about this before, but it's almost a good filter when building products as well is, you know, would this make a good TikTok video? And so if you can build a feature into the app, and you got to be careful because this can be a trap as well as like a really fun TikTok video doesn't necessarily mean that it's a great product feature that delivers value and stays core to the app and everything like that. But it's a fun filter to think through. I talked to the 11 Labs team on the podcast a few weeks ago. And they were talking about how they write the tweet before they build the product. And, you know, thinking through like what are people going to care about? What's the hook? What's the like three-second demonstration? And so many of the apps that we do see going viral and doing really well with UGC, it's because they have those hooks. And it can just be such a powerful thing. Yeah, it's almost like the new version of the Amazon memo. I've never worked at Amazon, but my understanding is if you were a PM at Amazon, you needed to write the launch memo of how the product launch was going to appear, you know, in the New York Times or in some sort of online publication. But people's attention spans have gotten so much shorter. So now instead of, instead of, you know, a news article, it's a tweet or a TikTok video. Yeah. And it is really helpful to start with the end in mind because the reality is no matter how good your product is, if people aren't willing to give it a shot, it's not going to matter. It's like a tree falls in the forest and nobody's there to hear it. Did it really happen? And so you need to start with the end in mind, which is how are you going to get your first 100 or 1,000 users? And part of that is, you know, writing the tweet up front or figuring out what the TikTok video needs to look like up front. And to bring it back to AI and the theme of this podcast is that these new experiences that are enabled by AI, like we talked about with Cal AI, like that was their hook, was like, you know, they show a picture of like, yeah, I'm tracking my match grows, but look how easy it is. I'm just taking a picture and then it gives me all the macros for this meal attracts my calories. These new AI experiences are those like really fresh things that are more likely to catch attention and go viral and things like that. So yeah, so many ways to do this these days. All right, tip number three, using AI tools to build and test hundreds of creatives per month. Yeah, and so this is probably one of the more standard tips that I imagine a lot of companies are already using because it's pretty straightforward is very important, which is if you are relying on paid advertising as one of your channels to grow and most subscription apps are, then there are tools you can use now like 11 labs when it comes to voice or like VO3 when it comes to video that will allow you to just exponent. increased the number of creatives you can test in any given month. An example that I use here is RUNNA. RUNNA went from tens of creative concepts per month to 400 plus creative concepts per month. That their marketing team is spinning off each month. And when you think about the implications that has for the company's growth trajectory, they're massive. But not just because more creatives leads to more A/B testing leads to lower CACs because you're figuring out which creative is working faster and you're able to optimize and double down on your best performing ag creatives through networks like meta and TikTok. But it also leads to more rapid learning cycle. You're learning more about your users and what resonates with them, which can inform product roadmap decisions in addition to just paid user acquisition. So this is another example of where the subscription value loop has these three discrete steps, value creation value value capture. But they're really all connected as a system. And oftentimes the best teams are finding ways to make investments in one step that also positively impacts other steps in the loop. How do you think about the downsides of this flood of creative though? And I'll name one specific. The FTC had a ruling I think a year ago. And it applies more broadly than AI. But it's essentially that you cannot fake user testimonials. So technically if you're creating an AI video or a slideshow or whatever and it's like a person, even if it's a made up person, and that person is saying, you know, I used run to crush my 5K. And if it's not a real person and it's not a real testimony in the US, that's technically against FTC guidelines and you can be fine and all that. How are you thinking about navigating the kind of like truthfulness of it, the preserving the brand? I mean, yes, you can just spit out hundreds and hundreds of creatives. But how do you do that botfully and carefully and in ways that won't get you find in the long run? Yeah, I'm so glad you brought this up. And again, we could do a whole episode on this. But the short version is there's a difference between creating AI-powered ads wholesale through a platform like VO3 that's handling everything into end, right? The video, the music, the person's voice, versus using a real human being, but using tools like 11 labs for voice or sooner for background music that allow you to just create a lot more permutations of the same ad unit at the same time. And so what I've advised my clients is be very, very careful about going the wholesale AI route and using a tool like VO3. One, because these tools are still evolving, and I don't know that they're quite at the point yet where those ads are fully convincing, two, because of the regulatory risk you brought up. And three, because there are a lot of consumers who have a really negative reaction if they detect that you're using an AI at, which is increasingly hard to do, but it's a risk, right? Because if that blows back in your face, it could negatively impact your brand. But I think the risks are much lower and the benefits are every bit as high. If you can create real organic content coming from real human beings, but that just layers on 11 labs technology so that you can take one spokesperson and translate their message into 30 different languages using AI, or take one 30-second ad unit on YouTube or TikTok, and then use Suno to layer in multiple different background music tracks and see which one performs best. To me, that's the right place to be playing in right now. And one of the better uses to, for diversification of creatives, is also just using it as a brainstorming partner, is that mining your user reviews, mining feedback you get, mining anything you can get your hands on, competitors, competitor ads, and then using that to be able to more quickly iterate and try new things as well, not just, like you said, wholesale creating assets using AI. All right, tip number four, highlighting AI features to reduce CAC and unlock new markets. Yeah, so I mentioned PhotoRoom briefly on the last tip, but this is one where PhotoRoom has really shined. I had Olivier Lemary on a podcast episode last year, and he talked about how historically PhotoRoom had really struggled to penetrate some of these longer-tail emerging markets, so countries like Mexico, Brazil, and Indonesia. And the reason for that was PhotoRoom, like many consumer apps, is heavily dependent on paid user acquisition to scale. But they just didn't have sufficiently high conversion rates in those markets, in order to get the unit economics to work through paid acquisition channels, even though the average cost for an impression on ad networks in these markets tends to be quite a bit lower than in markets like the US. But when they launched their new AI-powered background feature, where you could remove a background or replace a background using AI, not only did they see subscriber conversion rates and the product go up significantly, but they also found that by leading with those AI features in ad concepts on meta and TikTok, they got much better performance, especially when they put it within the first six seconds of the ad. It just really caught users' attention. It goes back to the automotive experience. If I'm watching an ad with the tenth version of a calorie tracker or a language learning tool, it's like, "Okay, I've seen this before, and I'm next." But if I see this AI-powered background removal feature as somebody who's interested in photography that I've never seen before, then I'm much more likely to click through and give the product a chance. And so not only did this significantly improve their subscriber cap numbers and their subscriber conversion rates, but it also fundamentally unlocked new international markets that they previously were unable to go into at all. We've already talked a lot about AI unlocking these features to go viral and that are really powerful in user-generated content and stuff like that. And so it makes a ton of sense that those same principles apply to your paid ad spend as well and getting that into the onboarding and getting it into the product in ways that really does drive that excitement about the product and everything. So let's move on to value capture. We've got four tips under value capture. The first one is using cheaper, faster LLM's when performance is good enough. I'm going to push back on this in a little bit. Why do you think that's okay? Yeah, and welcome a healthy debate on this one. And I will admit it's controversial because at this point, the best open AI and then theropic models are just so powerful that it can be easy to just say, "Well, why would we try anything else?" But Gamma's a great example of a company that when they were first getting started, they famously had this viral tweet in March 2023 when they launched AI-powered presentations that finally altered the trajectory of their business. And six months later, they were profitable. And there are a number of reasons why they got to profitability so quickly, but one of them was they were very savvy in the underlying LLM's they used to power the early versions of their AI-powered presentations product. Which is not to say they weren't ever using the most powerful models, but in some cases what they found was using some of these longer tail models, not only was the performance good enough in terms of the quality of the output that they were delivering to their users, but also the speed of the product was significantly faster. And the cost to the business of the compute power needed to serve these AI-powered features was significantly lower. Which as a startup that was fighting for its survival in early 2023 was really important. And so it's just a lesson that at this point, we're all waiting with baited breath for the next release of the hot new open AI-powered, the rapid model. And that's great. The frontier of what these models can do is so fascinating and compelling. But don't forget that there's this long tail of thousands and thousands of other models out there and sometimes those models tailored to your use case can be the optimal solution. I wanted to push back when I saw this one, you know, as we were prepping for the podcast, because I think where you draw the line of what is a good experience or not is where this can be a trap. And we've been talking about this a lot inside Revenue Cat. We just released this really cool chat bot where you can talk to your data at Revenue Cat and you can ask it things that you would otherwise spend five minutes looking up in the dashboard with all these filters and stuff. You can ask it things you couldn't even find in the dashboard because it can combine things in ways that the dashboard doesn't currently enable. And we are using the highest end models. And we're doing that because we want the product experience to be the absolute best it can possibly be. And we're doing that with the recognition that over time that should get cheaper. So I think you need to be careful not to over optimize too soon and figure out exactly when it makes sense. And I like the way you put it though sometimes you can create a better experience with a less powerful model if that model is tailored to that specific use case. For example, I've been tinkering with a little, if I could adapt, I've been working on. And I actually think Gemini and Gemini Flash specifically is probably the best experience for that app. Because Gemini is really great for some of some things that open AI and anthropic models aren't actually good at. And it just also happens to be cheaper. But I think if you're trying to create the best possible experience and you're selling it at a premium price and you're trying to get people excited about it, you just got to be really careful not kind of prematurely optimizing on cost versus just going for that like best possible experience, especially in this age where the competition is too fast. so fast where you. you're going to have a competitor come out and copy that feature, but maybe use the better model. Any follow-up thoughts to push back or echo what I said? >> I think it's such a great point, and I'm glad you're pushing on this. But I would argue that the two points don't have to contradict one of them. So what I'm arguing is, one, the product experience is a function not just of the output that the LLM provides, but also how fast it provides that output and how expensive it is to generate that output. So if you were on the bleeding edge of a prosumer category, where peak performance is the thing that matters and cost is no object, then absolutely go use the most powerful open AI or anthropic model. But to use your analogy before, if your Cal AI competing with my fitness panel and your target market is a 20-something young professional who is not yet at the point where they can afford the most premium tool on the market, and a faster, cheaper model is good enough, then it may very well be that you're not best served using open AI or anthropics latest model. The point you're making, which I think is equally valid, but I think you can reconcile it with the first point, is all of these models are seeing their costs decline pretty rapidly. And over time, I think there's going to be more and more commoditization of the models. Like deep-seek was big whatever it was a year or two ago, because they came out of nowhere in China, and their performance was every bit as good at one point as some of the top open AI or anthropic models. And that has changed. All that competition is creating downward pressure on the pricing of what these models charge for their API use. And so what that means is you should be thinking about where things trend over the next six, 12 to 12 months, not just the short term. And so I think both can be true. Consider who your target customer is and how much speed and affordability matters relative to pure performance as you decide which LLM to use. But then also keep in mind that the cost of these LLMs is going to come down quickly over the next six to 12 months and factor that into your calculations. All right, let's move on to tip number two. Offering basic versus AI-powered subscription tiers. Yeah, so this one's sort of obvious once you break it down. But I think that the key point is it used to be the marginal cost of serving a user subscriber for most consumer or pro-sumer subscription apps was close to zero. Unless you were a hardware product like Aura or Woob, or you were say a meal kit or physical goods service like blue apron or stitch fix, your underlying cost structure was very minimal because you're just a software business. Right? But now with these AI-powered products and features, that's no longer true because you do have the underlying compute costs to support the AI features. And we just talk about how the cost of LLMs is coming down pretty rapidly, but it's still a non-tribial cost structure for these companies to support. And if you're an early stage startup or especially if you're a bootstrapped solo printer who doesn't have a lot of venture funding, you have to be very conscious of this. And so what you're seeing is more and more apps, Dualingos, a great example, are moving to a two tier or multiple tier structure, where you have a basic subscription plan that gives you non-AI features and maybe a taste of the AI power of the product. But you don't get full or unlimited usage. And then if you want the full-blown AI version of the product, then you have to pay for a pro plan or a max plan or whatever the tier is called. So you're seeing this with Dualingo with their Dualingo max product, which is their new immersive language learning tool that allows you to talk to Lili and do role plays, all using voice, not just text. You're also seeing it with products like perplexity or chat to PT or cloud themselves. And then there are all sorts of other examples of tools. We've talked about 11 labs, we've talked about 12 and like all these products have multiple subscription tiers, which used to be very unusual in consumer subscription, but it's becoming a lot more common because you have to protect against the underlying cost of supporting AI features. Yeah, and maybe this kind of reconciles our last conversation on prices that, you know, as a product guy, I want to just build the best product possible. And so this is that opportunity where if the costs are unreasonable to provide that best experience, maybe you don't just eat the cost, but you create this premium tier. And Kelly would maybe be a good example of like, you know, and again, I've actually said this multiple times. I mean, in press that they were only charging 30 bucks. And I actually think that's part of been part of their success was that cheaper entry price. And it was just a kind of easy hurdle to get, especially with the kind of market they were going after. There's that perfect opportunity to say, hey, you know, at 30 bucks a year, you're getting the 70% accuracy, 80% accuracy, whatever. At a hundred bucks a year, you're getting the 90% accuracy. So I think it is a good opportunity to still build that maximal potential product. But figuring out how to tier things and how to message the tiering, I think can be tricky. So it's definitely something to be careful of, but maybe an option if the costs are just totally out of bounds and you're not in a position to be able to cover those costs. Definitely. Yeah, definitely. The companies just have to be more conscious of their unit economics and their underlying costs. Yeah. All right, tip number three, limiting trial links and premium to minimize AI costs more, more cost-based ideas here. Yeah. Well, and you know, as a business guy who's worked in technology most of my career, it's refreshing because it's great to have a software product that essentially has no underlying costs, but that's not the reality for most businesses. And so this is just another flavor of what we just talked about, which is in the same way that AI-powered features probably belong in a higher price subscription tier. Trial links and premium product offerings need to be a little bit more conservative in a world where giving away AI features has significant cost ramifications for the business. And so Tolen's an example of a company that started with the 7-day free trial and has adjusted its free trial length and its premium offering over time. There are plenty of examples of other companies that are opting for shorter trials or adjusting their premium offering for the same reasons. You guys had in your latest data subscription apps report this data point that called out that despite longer trials, trials as long as 14 to 30 days, often performing better on conversion, the majority of companies are actually training in the opposite direction toward shorter-term trials. And I suspect the reason for this, if you were to cut it for AI versus non-AI apps, I suspect this is being driven largely by AI apps that are shortening their trials out of necessity. Because if you give away 30 days of free AI feature usage and you have whales racking up massive compute costs without paying you a dime and then they turn, then that's not a sustainable business. No, not at all. All right, and along the same lines is tip number four, charging for additional AI credits beyond usage caps. And this is a great way to maybe offer premium but make it super limited and then allow the ability to either upgrade through subscription and/or these credit systems, which we see more and more apps using. And one of the things I'll all say too before you jive in is that I do think consumers, there seems to be a bit of like a consumer reset in how they understand these cost structures and the value that's being created. When chat to BT is 20 bucks a month, I feel like, and I've talked about this on the podcast recently, there seems to be a little bit of a reset of like people willing to spend more and understand that a lot of value is being creative and almost understanding maybe because they see the numbers of hundreds of billions of dollars to these big AI companies that like this stuff is just expensive. And so it's like a unique place where we're at in the market where it's like the alignment of the technology and consumers is such that I think you can get away with this in a way that you couldn't in years past to align those costs with the actual usage and value delivered to the user. Yeah, I think that's absolutely right. It goes back to value creation in a way, right? Like the idea that a consumer subscription product can charge $20 per month or in some cases $100-$200 a month for max plans even three years ago would have been absurd. Like it would have been rejected on its face. But these AI tools are creating so much value that the people have the means to pay for them are very willing to shell out that cash because it just unlocks so much value. And then the other thing you called out is I think it used to be consumer subscription tools needed to be very, very simple in terms of, okay, you get your seven day free trial and then you've got a single subscription tier, you're either in or you're out and maybe there's a monthly plan in an annual plan, but that's it. That's all the complexity there is. But more and more, I think B2C is taking a playbook from B2B where for a long time there has been value or usage based pricing, not just basic subscription tiers. And so you're starting to see this crop up in products like 11 labs, like Tolen, even Claude and ChatsyBT or Gamma where you can buy a base subscription tier, you can buy a pro subscription tier, you can buy a math subscription tier. But then even if you're on a subscription, if you're a power user who goes well beyond the cap that your subscription provides you with, you just have to pay additional money for additional AI credits or going back to our PLG loops conversation, you can invite more people to the product to get more AI credits. And in a way, that's just another mechanism for providing value back to the company because you're bringing in free users. Yeah, yeah. It's a great way to align costs and value. And we've talked about this a time hybrid monetization and finding ways to charge more to users who get more value out of the product. And I do feel like AI has enabled this in a way that it wasn't possible before. And it's kind of training users too. I mean, if you're on the free ChatsyBT, you get you run out of messages and it says you can't use it again for an hour. I've actually been frustrated with Claude specifically When I. I just run out and they don't even prompt me to increase my tier, which I did recently when I've been vibe coding my little app I've been tinkering with. I ran out of usage and it said it wouldn't reset for six hours and I was like, well screw that, I'm gonna just upgrade. So I upgraded, but I had to quit the app and jump through all these hoops just to get the upgrade. So you gotta create a good experience around this as well, but anyways, the point being, I think that consumers are kind of recognizing these kind of limits in the AI age and so it is more natural to introduce these kind of things and it has been in the past. And so, and especially when cost can't get out of whack it's very important to be thinking about it and then it's great that it really is an option now and not something that consumers will just instantly bulk at. - Yeah, well you're certainly not alone. I've done the same thing with Clawd I've upgraded my usage multiple times, not just for me, but for the rest of my team. And it's the ultimate sign of product market fit when they don't even have to prompt you to upgrade or provide an upgrade slow because they know they're creating values so fast that they would rather spend their team resources just creating more products and features and value than worrying about their upsell flow because they know people are gonna upgrade because the product is damn good, they don't have a choice. - I do think they should work on automatically being applied immediately, that's probably just a bug that they're fixing. I mean, it's crazy, we talk so much about product and we have this whole conversation about being a really great product and everything like that. But when you're delivering a ton of value, you can get away with stuff and people will jump through those hoops. So it's kind of an interesting lesson in this new age of how much value is being delivered and then what you can get away with in that. - Totally. - All right, well let's wrap up with the three questions I now ask every guest. The first one being what's the biggest win you've seen in the past year? For you, you work with a ton of clients. I'm sure you've seen a lot of incredible wins, but what would you say is the top win you've seen the past year? - Yeah, so one of the fun things about being a growth advisor is you just get to work with so many different companies and you sort of start to develop pattern recognition of what works. So there are a lot of different examples I could point to, but we've been doing a lot of work recently, specifically on helping companies move from hard paywalls to freedom. Which could be challenging. It's sort of like moving from playing checkers to playing chess because it requires a lot more sophistication. But one of the tactics we found is really helpful there is what we call a multi-step paywall. So basically moving from a single hard paywall that's just basically, hey, you've got to pay for this product now to, this product is free and it will always be free, but we want you to try the best version of it and we'll give you the best version of it for a seven day free trial. But after that, we'd love to have you continue paying for subscription to get maximum value. And so there was a case recently, I'm gonna name the client, but we saw a 75% increase in LTV per user through the implementation of this multi-step paywall along with some other pricing and packaging optimizations, which is pretty fundamentally altered the full potential size of this business, right? Because now they've moved from a hard paywall that's excluding lots of people to a premium model that is just growing much more quickly through organic acquisition. - Yeah, in the status registration app report, we showed that in our data, hard paywalls performed five times better. So I love that your biggest win is going against the grain. And we talked about that in the episode Jacob and I did about the state of subscriptions is that there's a lot of reasons to still use freemium and don't take that 5x increase in LTV as a sign you should absolutely do a hard paywall. So it's great to hear your biggest win is pushing back against the hard paywall and opening it up to more people to create a bigger opportunity. All right, what has been your biggest fail of the year? What fail have you seen with clients that, especially something you could learn from? - Well, I'll stick on the hard paywall versus freemium theme because it doesn't work for everyone. And so the other side of this coin is, we had another client we worked with where we attempted the shift to freemium and the initial results didn't work at all. Like we saw more than a 50% reduction in subscriber conversion. And so we very quickly pulled off of it after a couple of weeks but we learned a lot of really valuable information that I think will inform what I hope will eventually be a successful transition to freemium. When I use this checkers versus chest analogy, what I mean is for the vast majority of apps, hard paywall is the right answer. It does convert 5x better. And if you're a bootstrap startup or you're operating with limited outside capital, it's a much more reliable and low-risk way of growing your business. But if you want to build a billion dollar or a $10 billion or $100 billion company, there are a lot of examples of apps like Spotify, Dualingo, Strava that have done that through freemium. And it makes sense because you're just going to attract a much larger user base at the top of the funnel if you have a free version of your product. - Now, all right, last question. Growth would be easier if. I'm going to go with growth would be easier, specifically in my world, which is consumer subscription. If consumers weren't so fickle. And that's true now more than ever. Because when the app store first launched and you've talked about this on your podcast, it's like in a world where there are only 100 or 1000 apps in the app store, you can get discovered very easily and people will give your product a try. For the simple reason that there's nothing else out there. But we couldn't be further from that world now. You got hundreds of thousands of apps across the Apple and Google app stores. And so what that means is consumers are constantly being bombarded with potential substitutes. They may never download your app in the first place. If you don't have a reason for it to really stand out. And we've talked earlier in this episode around some tactics you can use to help it stand out. But even once you've converted a user and converted them into a subscription, the odds of them turning are much higher than if you were a B2B SaaS business. And so this goes back to why you need to constantly be innovating and launching new features and constantly iterating on your onboarding flow, not just to get to product market fit, but to maintain product market fit. And so that's the hardest part is that consumers are just very fickle. They have short attention spans that got tons of other options if your product is no longer serving their needs. But I also think that's part of why our job is fun as a growth advising firm. And it's part of the reason why we exist is to help companies navigate that. >> Yeah, I've seen a lot of tweets recently going both sides of the consumer market. Like it feels like in some ways, it's the most exciting time to ever be in consumer. But then a lot of founders talking about just how challenging consumer is for those reasons. But that's what's so fun about it is that when you build something that hits, when you build something that resonates, the opportunity is just so massive. And then even like you were saying earlier in these niches, even if the opportunity isn't massive, maybe the monetization opportunity is massive that you can be charging $100, $200 a year, $20, $30 a month because you're delivering so much value. And so yeah, it's just such a fun space to be playing in despite the fickleness of consumers. >> I agree. I've never had more fun in my career. And I think it's a great time to be in consumer tech. >> All right, well, as you wrap up anything, you wanted to share with the audience? >> I just say, if you're working on a consumer subscription app and we can be helpful, you can check out Elemental Growth at my website, PhilcheeCarter.com. And then I also have my own podcast called SuperSive. So give it a listen if you have a chance. >> Awesome, thanks so much for listening. So much fun, dog and a day. >> Thank you, David. This was a lot of fun. (upbeat music) >> Thanks so much for listening. If you have a minute, please leave a review in your favorite podcast player. You can also stop by chat.subplub.com to join our private community. (upbeat music)