Representative appearances

• Running the Product-Market Fit Treadmill with Brian Balfour | Mostly Growth

  Kyle Poyer and I launched a new show called Mostly Growth. It's where we dig into how companies actually drive growth, pricing go to market and everything in between. The episode you're about to hear first premiered on Mostly Growth yesterday, and it's already another episode out now that won't be shared here at all. So make sure you're following mostly growth wherever you get your podcasts. Now don't worry, Run the Numbers isn't going anywhere. You'll still get the same great CFO and investor interviews every week. All right, here's a taste of mostly growth. Now there's like, fake it until you make it. And then there's whatever this is. There's a term for it, Kyle. Do you know what Flintstones is? His car doesn't go. It's his legs underneath. Like going like. That. A 2026 plan that I am literally talking to my board about in two days. Number one, on the plan is how do we accelerate product adoption? We lived in this world for a while where that threshold was kind of increasing on a pretty consistent basis for most categories. I kind of basically dropped a giant bomb in the middle of this. C.J. was very excited that you agreed to join the podcast, because he's a bit of a fanboy. Time and time again, you know, if you were in the mobile battle, if you bet just on the iPhone, you had a chance at winning. If you bet on one of the other platforms, you clearly lose, or even if you spread your bets out. Those were also the clear losers. From most of the media. It's mostly growth. Go to market strategist Kyle Poyer and CFO C.J. Gustafson share. Smart takes on growing revenue and running a company. Pricing unit economics, AI leadership, candid insights for CEOs, CFOs and crows, or anyone else aspiring to climb to the C-suite. This is mostly growth. Hey, thanks for listening. We'll be right back after a word from our sponsors. As a CFO, you're expected to bring clarity to everything, including how your company manages ownership. The tools you should make that easier, not harder. But when they come with hidden fees data, you can't always trust and support. You can't rely on. You end up with more frustration and clarity. That's where Polly comes in. Polly is a cap table management platform built for finance leaders who need accurate, decision ready data from a system that works for your business, not against it. With Polly, you get intuitive workflows, audit ready reporting and support you can count on, all backed by predictable pricing. You can plan around so you can stay compliant, make smarter decisions about your company's future. Manage everything from fundraise modeling to AC 718 reporting to custom liquidity programs all on one trusted platform. Hundreds of high growth companies like Linear Fathom and Runway Trust Polly to help turn ownership into a strategic advantage. Joining them is now easier than ever. You can switch to Polly in as little as five days for a limited time. They're offering 25% off your first year to claim your exclusive offer. There's an all encompassing metrics that is p u l l e y.com/mostly metrics. Please, guys, I really need this poll ecommerce metrics. You just launch a new AI product. The new pricing page looks great. I'm talking crisp but behind it. Last minute glued code, messy spreadsheets, and running ad hoc queries to figure out what to build. Customers get invoices they can't understand. Engineers are chasing billing bugs, finance can't close the books. Well, with metronome, you hand it all off to the real time billing infrastructure that just works. Reliable, flexible and built to grow with you. They turn raw usage events into accurate invoices, give customers bills they actually understand, and keep every team in sync in real time. Whether you're launching usage based pricing, managing enterprise contracts are rolling out new AI services. Metronome does the heavy lifting so you can focus on your product, not your billing. That's why some of the fastest growing companies in the world, like OpenAI and anthropic, run their billing on metronome. Visit metronome.com to learn more. That's metronome.com. Welcome back to another episode of Mostly Growth. Joining us today is Brian Balfour. Brian started reforge in I believe 2016. Which is well, almost ten years ago. I remember it as a learning resource for growth teams, but I was on the site recently and it's now a full platform for AI native product teams. We'll get into that with Brian. And before reforge, Brian was VP of growth at HubSpot, where he took the company from a single product company to a multi-product platform. This comes from stocking Brian's LinkedIn site. His first foray into being a founder was starting a college specific social network in 2003, and his first PM role was at Zoom Info in 2006, where he added social features. So you can either thank him for that or blame him for that. Welcome, Brian. Did I get any of that right? That is right. Mostly right. Just like mostly growths. I'm interested what actually falls outside of the growth part into the. Mostly the other things are CJ's rants about people he runs into and various parts of Florida. I think knock knock jokes in black payback period calculations. Okay. Sounds good Brian. So you know, know you as founder and founder and CEO of a company for ten years. So you're building a business, but you're also like prolific writer and inventor of new frameworks. How do you do both of those things at the same time? Not very well. So the writing is incredibly sporadic. If people are subscribers on my blog, they will know that there is no continuous stream or routine of things. I am fortunate that for, large portion of free for just life, that the two went hand in hand. Together. I do think it's one of the things that has paid the most dividends for me in my career. Yet I look back and I regret not investing even more into it, even kind of going way back. Way back when, the writing is how I met my first investors and my first venture back company writing is how I met my co-founders and my second company. My blog is how all the HubSpot founders and executive team came to knew me, and kind of what led to that job. The writing is kind of how we found our first. I don't know how many customers. For reforge. So, it's just like I look back at every pivotal moment, in my career, and I can kind of like first touch, last touch. UTM source has been, the blog and the writing. Yet I still find it hard to make time for it and and do it continuously. So, it's definitely it's definitely a challenge. And what I find great about your writing is that it's not just, you know, observations of what's happening. It's like bigger frameworks around it. Like the four fits framework, which I think is a great evolution beyond just thinking about product market fit but understanding, like the different degrees of finding product market fit and what that actually means over time, or, you know, growth loops as frameworks. But is there is there a framework that stands out to you as like the most evergreen, the one that everyone keeps kind of coming back to you for the most popular one that I've written, by all accounts, whether it's mentions, views, whatever you want is definitely the four fits. That one's kind of, you know, funny enough, in the last year has kind of made a resurgence. I don't know exactly the cause. It. I did do an update to it, but, I saw the resurgence even even before that. And so it's definitely the most evergreen one that people, refer back to. And I think at the end of the day, all of these things are just a way for me to make sense of very amorphous topics, not just make sense of them, but communicate, them to others. And so almost all of the writing has started from a standpoint of I'm either making sense for myself or I'm trying to make sense and communicate it to others in my job or in my company. And then it kind of turns in into a writing. I don't really necessarily set out to be like, you know, framework maker or, or however you want to describe it. If that's if that's a job title, that because I think a lot of things that we do in especially startups are messy and, they're these like concepts that at first glance sounds simple on the surface, but when you're actually in the thick of it, it's very complicated and much messier than kind of how it lives on the page. And so I, in my own endeavors, have tried to pull out a little bit more of the nuance and turn them into tools that either myself or others can use on a repeated basis. Now, the flip side of this is that you get the chorus of folks on ECS or, you know, our product management or, you know, whatever subreddit you are to go. And they're like, who are these people? Like Brian Balfour just making up this garbage frameworks. It's not how like it works, like all that kind of stuff. And, you definitely get the flip side of it. Like, a lot of people consume these things and view these things incorrectly. They view them as answers and I view them more as tools to use to get to a better answer. And so with any kind of tool, you have to combine it with some set of like raw ingredients that you have available to you. And if you combine those two things together, then then you can get to, a better answer. But it's not like, the answer appears out of nothing. You actually have to put work in every you have to combine it, with other things. And so it it's a way to kind of stand on, the shoulders of those who have been through these experiences before and have helped codify it into something simpler. That doesn't mean that it's codified into an answer for you. And I see I just often find a lot of a lot of disconnect. And as a result, folks tend to write these tools off, almost to simply, yes, I, well, the downside of writing a lot online is that if your writing is successful, everyone has an opinion about, that's fair. For, for better or worse. The trolls come out no matter what. The worst thing is actually no one talking about you, because that means no one's being fair. Yes. But one of the things that, you know, I think struck me about the forfeits framework and why it's having a resurgence. Is that there almost seems to be this, like misconception that product market fit was a, you know, binary thing like you had or you didn't have it, but also something that once you have it, you keep having it for at least for, you know, a long period of time. And now it seems like I think you call that a PMF treadmill. You have to stay on this treadmill that kind of keeps getting churned up. To maintain product market fit. And there can be some technology shifts that take folks from having PMF to maybe seeing a PMF collapse. And I thought, that's just a super interesting concept for right now, given, you know, the rise of AI products, but also just, kind of rapid change in market environments. But I would be curious to hear what observations led you to kind of this concept of PMF collapse and, and the PMF treadmill. Yeah, there's actually two components of it. There's or kind of mistakes around it. There's one is that thinking product market fit is binary. You mentioned that that it's constant. The second part of it, which is what the four fits really talks about, is it's like the only thing, you know, like once you achieve it, that's all you kind of really need. And both of those things are incorrect. I think on the first piece, the concept of the treadmill actually came from Casey Winners and Friedman about, who did the product strategy course. They came up with this concept of the treadmill because, we have these visuals, these graphs, which is that PMF is more like a threshold, right? So there's some threshold that you have to exceed in order to have product market fit. And, you know, a lot of 0 to 1 work is about finding where that threshold is. Are you able to build the right product, choose the right market to get above that threshold. There's actually two variables there. You could actually change the market, or you could change the product, to move that threshold around and stuff. But then, once you exceed that threshold, if we graphed it out, is not like a straight line across. It's actually a line that, progression constantly continues to increase over time. The reason it continues to increase over time is just because of relative competitive pressures and kind of progress in products, right? So your customers expectations are always increasing over time. You have zero control over that. There are always increasing over time here because you have direct competitors that are building things to increase them over time. That technology in general gets better over time and more importantly, that they start actually using products in other places. That increases their expectation. So you kind of see this dynamic mostly in B2B and B2C software, which is as the quality is, more and more consumers use B2C software and get used to those dynamics, they expect the B2B software to be simpler to use, to have similar UI, to use similar modern technology. So it's not even direct competitive pressures. It's just like these things increase over time. And so as a result, you are constantly on this treadmill, to not only kind of keep up with how that PMF threshold is accelerating, but actually try to widen the gap between you and that threshold. That's kind of really what, really kind of what creates things. Now, the interesting thing is, in this dynamic or in this world, the AI world is that we lived in this world for a while, where that threshold was kind of increasing on a pretty consistent basis for most categories. And as a result, it was once you hit that BMI threshold, kind of layering on other things to kind of keep up with it, was a more quote unquote predictable exercise. AI is kind of basically dropped a giant bomb in the middle of this. And, and as a result, we've seen that threshold rather than increase consistently in some categories, basically hit this inflection point, like basically spikes started to go straight up, of course, the classic example that we pointed out and then, and then has been repeated a bunch is, is, Chegg. So, you know, ChatGPT, came out, it fundamentally disrupted, how users were achieving in solving the same problem that they were solving in a much faster, more personalized and in other ways. And so and because ChatGPT grew so fast. Right. It's not like, it had a long time to expectations kind of, adapt it incrementally over time. They actually changed in a very short period of time creating this spike feeling. So as a result, you know, Chegg just ended up in this negative flywheel where they didn't really have time to respond to adapt to this big change, which is pretty interesting because, they experienced one of these moments in the company previously, which was a shift from physical textbooks to digital textbooks. But the time that, you know, that took was actually much longer. So they had a much longer time to adapt. Unfortunately, it's ended up kind of where we sort of predicted and expected a year ago. I think they laid off since the beginning, like over 80% of their staff, like founders, kind of step back in to try to reboot the company. May have completely pivoted to an different segment of the market there. Now, in the professional training market, a similar, similar, even less rapid change in the professional, training space. But, I think what's interesting to me about this with the treadmill, I think of the CRM category as being historically really hard to build, something that competes against Salesforce and others, because there's this set of expectations where in order to, be, you know, viable CRM for a business, the number of table stakes, features that people just expect is like, there's hundreds of these things that people, you know, are almost checkbox features, but people expect in a CRM. But hey, what's what's kind of exciting about this concept of PMF collapse? And the change of markets is that, you know, now there's, you know, maybe this, this idea of what a CRM is and what you need from a CRM maybe changes with AI, where you have a database and you have systems of action and workflows. And so, you know, we're maybe going from this period of the treadmill, you know, being being on for, you know, 20 plus years with CRM and having to build all of these things to be a viable CRM, to now being in a position where there's maybe new markets emerging, where we could think about what a CRM is differently. So there's I guess there's a challenging part and a and a, you know, an optimistic part of, of the PMF collapse and, and, this phenomenon that they keep happening. But. Yeah. Yes. You know, there's two sides of these coins are always right. Like, wherever you see somebody crashing, it's typically the result of somebody kind of exploding in a good way. On that CRM note is we wrote this big blog post about your, a way to assess your risk of PMF collapse. And it kind of comes down to like a dozen different factors, right? Switching costs being one of them. Right. So the higher the switching costs, the lower the risk you are. But that's just like one of a dozen factors. Right. And you compare that tag like the switching costs were so low. And another factor was their audience was early adopters. Right. Like these the, the kind of younger audience that's a, that's another factor. The more early adopter your target audience is, the more risk you are of, of these types of dynamics existing. But I think in the CRM market, it actually points out something really interesting, which is that in the CRM you kind of build for two different audiences. This is kind of true for a lot of B2B software right now. AI, CRM quotes, that are going after markets where they are trying to do a rip and replace of a Salesforce or a CRM, and I think the dynamics that you actually see there are people are still coming to the table with a bunch of the same feature requests, the same product requests as they had in Salesforce or HubSpot CRM. Yeah. So there's all these stuff. Yeah. As as long as these AI has the these AI native CRM actually look like almost like AI tinted CRM, there's a whole other class of products. And disclaimer I'm an investor in a company called De AI. It's started by a couple guys I worked with. I, HubSpot and HubSpot CRM, which targeting much more of a market of people who aren't trying to do a rip and replace, but they're much kind of younger companies truly like operating in an in an AI native way and have a totally different set of feature requests. And these two things are different because actually, in that second category, I just describe, it's not a tool change, it's a complete workflow and way of operating change. And as a result, like you target those folks that they're going to pull your business in a very different direction than that first group, that I described the. But the they all have trade offs. They all have pros and cons. Right? Like first group I described probably is a larger market at this moment in time. Right. But the second group, it's a smaller market. But you're kind of taking a bet that this is the way that all companies will operate in the future. And so it's more like you're waiting for that. All boats are rising, rising tides. Right. Like you're waiting to kind of catch that wave, as, as it comes in and requires, I think, a lot more patience. And so I think you we could break down every single category, not just the CRM category. Do you see this dynamic playing out in both? Everywhere. Yeah. So if you think about the way that a Pal is traditionally constructed, you have your cost to serve your customer. So revenue less Cogs is your gross profit. And we've always thought about it like, that that's how you show SEO ability in your model. That's how you show leverage and everything that's in R&D. You can spend a ton more there. You can bring more, can make your, profit margin look worse because that's all for future growth, etc.. But I think with the collapse of PMF and this is my takeaway from reading your piece, just as someone who thinks about allocating money within a company more and more, the R&D component is actually the cost to serve your customer because they expect more and more over time. It's not like you give it to them and the product stays the same. It's kind of like you're graded on this sliding scale, and you have to keep investing in your product and creating new stuff in order to keep them around. It's very much, I think, at the top of the panel, more than I thought towards the bottom. It's actually partially why HubSpot I, I can't talk about the most recent years, but at least while I was there, we always invested way more in R&D than the comparative benchmarks for companies. Our size, where I can't remember as a percentage how much more about. But we very much believed we didn't we didn't have the need to name these things. Right. Like at the time, like, I hadn't I hadn't come up with those things. But from a principal and other approach perspective, it did it did end up working out that way in terms of how we thought about you keep keeping up over customer expectations and how we, and how we invested in those things. It's interesting to think that, there's been a bunch of firms and companies that have honestly made a decent amount of money by coming in and cutting our R&D and in efficiency and in a bunch of other places because of like, high switching cost and because of a bunch of other things. I think we're now seeing opportunities to take take market from those places. I find it fascinating that even more so now, customers are buying your product not just for what it does today, but what they believe it will be able to do for 100%. We see this a bunch in reforge. Yeah, and we actually guide on it too. It's like we're very transparent in our sales process about here's where we're going with our tools. Like we want you to align to that future that does a couple things, at least from from our perspective, I think one is it prevents possible future churn, which is, you know, bad too, is that you make sure that we're getting the right customers in the door that are going to pull us in the right direction. Three it actually also helps us counter position a bit against our competitors because we have ten x the product velocity. Any of the competitors in our categories right now, and we show it and we back it up and we prove it. And like we just cut them, we're like, do you want to go with the company that shipped two things. This year or literally ship something major 2 to 3 times a month? So, and we show them the list of things and the side by side comparisons and stuff, and, and I think that's right. We've even our own adoption of tools internally at reforge is we've very rarely centralized around, a single tool because things are changing so much and we don't know exactly where some of these are going. But in the cases that we do centralize, we want to know, okay, where are they going that align with where how we want to operate as a company. And we are very much kind of picking, picking for the future, not just where they are today. 100%. Yeah. I think one of the interesting challenges for companies that hasn't really been a challenge before, but I think I saw, a post from you about this that it's going to become a challenge for many companies going forward. Is communicating about the roadmap and all of these new features. If you're launching, you know, a big release multiple times a month or maybe even weekly, by the time you've gone through a year, you have dozens and dozens of launches out there. And, the, reason why someone came to your product in the first place is, you know, very different from what you have available now. And so how do you keep educating that customer and drive adoption of ongoing features, like how have you thought about that, especially as a company pivoting from, you know, more of the traditional learning and community and professional development space to increasingly, an AI product company? In this past year, we've launched five major new software products. We actually ended up sunsetting one of them. Right. And so but right now we have a portfolio of four software products. One is Reforge Insights, which aggregates analyzes all of your customer feedback, plugs it into your workflows. The second is Reforge Research, which helps you capture new customer insights with are we have like an AI interviewer product. The third is reforge build. That's our prototyping product. It specifically helps you prototype on top of an existing product, rather than doing kind of a 0 to 1 app building that a bunch of these other ones do. And then the fourth is Reforge Launch, which is, helps you kind of launch, manage, analyze all these AI features, these non-deterministic, experiences. We did this with a team of about somewhere between 20 and 25 folks. You would have asked me if we could have done that a few years ago. I would have told you you were absolutely bonkers and crazy. And so. But look, strategies today kind of need to be bonkers and crazy to compete. That's kind of what I've never stopped myself to do is like, if your strategy doesn't feel kind of bonkers, then you're probably not pushing hard enough. We launch all of those, plus a ton of updates to each of those products, and we have definitely are experiencing like, our product velocity is outpacing our, product adoption. And I'm seeing this in a bunch of other companies as well. I think that there's actually three big pillars to adopting AI, especially in product teams. There's product discovery that's figuring out what the hell you should build a launch. There's product delivery, which is actually like generating the production level code and kind of getting it out there, and there's product adoption and just do the nature and the strengths of AI as well as companies priorities. Everybody has been very quick to adopt AI in the product delivery pillar, and that's why you see the rise of the cursors in the cloud codes and, all of those places. And but the challenge with this is that our job is not to produce more code. Our job is to produce, more product value that our customers adopt. Like, those two things have to be true at once. So if you apply only to one of those pillars, you actually just move the bottlenecks, you know, to one of those other pillars. And as a result, companies need to be thinking a lot more about how how do I adopt AI on all three pillars at a roughly equal rate, if they actually want to impact that outcome of more product that my customers adopt now reforge internally ourselves, we've done a very good job on those first two pillars this past year. That's been our big focus was we were basically landing all of these products. We were kind of getting them out there and our next year, lot heavier investment in that third pillar. I have not solved the challenge yet. It is my challenge, my 2026 plan that I am literally talking to my board about in two days. Number one, on the plan is how do we, accelerate product adoption, go to market velocity? How do how do we get the velocity of that pillar up to be equal with the velocity of our other pillars? Because it's definitely unweighted. Now, the challenge with this that I haven't actually solved yet is out of those three pillars, the one that is slowed down the most just by speed of humans is product adoption. You can accelerate things towards the speed of computers with AI, but the ultimate bottleneck is us and, what we're doing as humans. And so I actually think the bigger lever is on product adoption will not be how to automate certain things and creation of certain things. That's not the pain that I'm feeling. The pain that I think is going to be much tougher to solve is as buyers, as adopters of things, right? We only have a limited bit of time. There's we sort the built habits on new products. We have to change the psychology of of people, how they think about you, how they think about the alternatives, like all of those things. That's actually where I think the real leverage is going to come from. In, in how to do this. I don't know exactly how to do it yet. I'm, I'm in the figuring it out, but I am open to ideas. If for any of you. Thankfully, I think once you do figure it out, you're going to make a really great framework. Which is awesome. This is what I want to feel like I gotta keep secret. But we'll we'll see, we'll see. I usually I'm usually very open and and sharing. Yeah. Well, we'll, we'll check back in next year. And, I know, so C.J. Was very excited that you agreed to join the podcast because he's a bit of a fanboy. And I'll let, C.J. take it from here. This is great, because Kyle teed me up in a in a very kind way, this time with Emily Cramer on last week, and he's like, hey, C.J., meet Emily. Can you insult her as a CFO now? Oh, wow. All right, so this is great. I'm paraphrasing. You're your person. Well, Brian, you've been living rent free in my head since June when you release the next great distribution shift. It was my favorite piece this whole year of all of 2025. My only complaint is it didn't have a clear answer at the end of it because this really, really impacts my business. Yeah. What answer are you looking for or I he wants he wants one platform that he could just spend all of his time. And that has been helping him grow as newsletter. And you hinted at it. You hinted at it. But yeah, this truly was a great piece on on the platform cycles. And it felt like such an moment for me to go through and actually apply the open, grow, close and monetize to each of the different networks that I've experienced over the last ten years. Yeah. I think a lot of people were looking for an answer. Look, I made a very clear prediction and this is before ChatGPT. My prediction was it was going to be ChatGPT. And I had a big explanation for why now that some time has passed and they released some more things, I would put even more chips on that. But me personally though, I did try to provide a bit of a framework in a follow up post of how I would evaluate it if I was in everybody's individual's shoes, because the answer is not going to be the same for everybody. The answer will be different depending on who your customers are, what your product is. You know, all of those sets of people because I think besides the big players in these experiences, the ChatGPT is the Gemini's clods of the world right there. There's a battle for that. But there's also kind of this sideshow going on where there's these other new platforms or experiences that are emerging, like cursor, that I do think will have very similar type of app platforms and will be distribution channels for certain types of businesses if you are targeting those customers. So in the same way that everybody's kind of focused on the battle of the big behemoths, I actually think there is going to be some really interesting companies made by, yeah, taking a bet on some of these sideshow, platforms as well. And Brian, just to back up for the benefit of listeners, can you maybe just go through what this brutal pattern is that that's happened over different platforms? Yeah. Yeah, there's I wrote three steps. There's actually four steps in the post. Right. So we've been through these cycles so many times. Most times there's a technology shift. There's also a distribution shift, meaning the new way to distribute your products to to your actual customers. And the distribution shift is sometimes as equally as disruptive as the technology shift itself. Now, the I with the technology shift came a few years ago, but the distribution shift hadn't happened. That's actually pretty normal. The distribution shift actually tends to happen about 2 to 3 years afterwards. Right. And so we've seen this play out time and time again. So when the distribution shift happens it goes through these four phases in knowing the phases is really important to know how you play the game. I'll call it phase zero is kind of the ingredients for these things to emerge, which is okay, the technology shift happens. It's pretty fricking clear that there is going to be some new huge, type of consumer and user destination. Like it's a it's a consensus idea. And as a result, you tend to have somewhere between 5 and 7 major players start to battle it out. Now in AI that's Google, that's meta, that's open AI, right? All kinds of. But you can look at all the previous ships too in social. It was Facebook, it was Myspace, Friendster. It also happened in mobile. You know everybody remembers Apple Android now. But there was a bunch of people. Facebook even made an attempt at Amazon made an attempt at it. And then phase one is like, okay, one of them really figures out what is the the moat, what is the defense ability, the thing that they need to basically hit, escape velocity away from the others. And then once they figure out what that moat is, they tend to go through this open phase where, it's it they either invite, developers, contributors. It could be it could be content, it could be applications, all that type of stuff into the environment. And, and as a result, what they do is they extend all of their use cases much more quickly. All of those apps, all those other people bring in more consumers. They bring in more engagement, they help create more of the moat. And then once there is clear escape velocity on that platform, then tends to go from open to close, they start to either levy, transaction fee, or they shift organic traffic to paid traffic. There's a number of different ways to tax the system. And, and as a result, it becomes much more closed. So it goes through this cycle every single time it happened in social Facebook, Facebook, when I was I had a company that had a bet in the space and experience all phases of the cycle. We then shifted to mobile and unfortunately experienced all the phases again. Right. So I kind of got punched in the face twice before, I really learned my lesson. But this happens over and over and over again, and it is 100% happening again with AI. You see, OpenAI, I launched the chat GPT apps platform. We can talk a little bit more about this. We will 100% go through an open phase. Their mode is about memory and context. So you can see this in the value exchange. When you install these applications, if you like, go dig in a little bit deeper, which is AI models have commoditized. Right. So the platform that owns the most of your data in context will be able to produce the best output for you. And, if they're able to produce the best output for you, that's going to lead to you engaging more and more into that platform over time. So they kind of figured that out with memory and context. But you can see this with their early application partners. When you go to install it, there's a little screen that says, that that application is getting access to your stored memory in context in ChatGPT. So the app gets some of that context to produce a better output for their user, which creates more usage on the platform, which lets OpenAI capture even more and more context, and memory about you. Right. And so this is just going to this is going to keep widening the gap of the outputs that ChatGPT will be able to produce for you, compared to all the other platforms, as they know more and more and more and more about you, that's kind of the game that we're we're going through right now. And that game is going to play out in other, other places as well. OpenAI and specifically ChatGPT are going through this cycle right now, but becoming a dominant platform. If you were to think about some of the platforms that could be next, or if you were in go to market or products, and you might want to build for emerging ecosystems or platforms that haven't yet closed but maybe are riskier but are more open. What are some of those emerging platforms that you'd bet on? Once again, this all comes down to who your specific customers are. There's the the mass market platforms. Right. And that's been that that battle is I don't think we're going to see a new player there. I think you're taking a you're looking at one of the five players that are fighting it out right now in you're taking a bet on them. At the end of the day, there's probably will only be two at most left standing with any kind of significant traction that just the history kind of shows that that's what it kind of consolidates down to 1 or 2. And so you're more kind of taking a bet on that. And we're still in the early phases of that. Like you still like, yeah, you can play around with your own ChatGPT app and kind of experiment on your own and stuff, but they haven't opened up the platform. They clearly said they're going to open up the platform. I've talked to a bunch of the early partners. That game is still coming out, and I imagine one of the other platforms, whether it's Gemini or Cloud or something like that, will counter with their own version of this. The question will be, what does it matter? At the end of the day, I'm actually of the belief that it doesn't matter that I think, you know, like it's open AI's world or more just all living in it at the moment. But who knows, right? There's always surprises around the corner in the AI world. So that's kind of one set of bets you're kind of you're taking on. And a lot of times people would sit there and think, well, I'm just going to diversify across all of them and let the winner emerge. Yeah. Does it doesn't work like that. I'm sorry. Most people, especially early stage startups that benefit the most from these distribution shifts, don't have the resources, the time, to actually diversify their bets like they actually have to choose one, be all in on it. And if they're right, they make the gains. And that's actually been proven true time and time again. You know, if you were in the mobile battle, if you bet just on the iPhone, you had a chance at winning. If you bet on one of the other platforms, you clearly lose, or even if you spread your bets out. Those were also the clear losers. There are all these boy. Once again, like these other, I think, you know, smaller ecosystems that emerge. And it really depends on what type of business that you're building. Right? Like I do think new developer experience like cursor and stuff, are pretty or will be pretty, sustainable. They will draw a lot of habit. They will draw a lot of usage of that specific crowd. And so if you are roughly targeting that customer set, there is some kind of play there with like the whole agent platform that they've been talking about in signaling about. But I think there's going to be new ones. And all of these, just like Salesforce is a channel for a bunch of new businesses. Maybe one of these new AI CRM that we were talking about earlier. You take a bet on one of those, at some point. But, so I think there, there, that tends to be roughly two bets. You have to take a bet on a major platform and maybe a minor platform, depending on what type of product and business you're in. The second one that comes out with the platform will tend to to in order to attract developers or contributors will make the value exchange look more lucrative. Myspace, a bunch of others kind of did this in the in the early social days to try to incentivize and pull people over to the platform. I actually think those incentives are a sign that those folks are so far behind, that it doesn't it doesn't matter. Like, you could take the incentive, but your time is actually wasted. Yeah, to your point, I was, talking to the, growth team at Web Flow and, how they're optimizing for AI discovery right now. It's becoming a very sizable source of leads for them. And even just looking at Lem referred traffic, it's, you know, one of their top and fastest growing channels. But then within the LM discovery, more than 90% of the share is ChatGPT. And then just a couple of points are perplexity for Gemini. I don't know if that's just web flow or if others are experiencing this, but it's wild just how different, that level of demand and kind of high intent traffic is on ChatGPT relative to other platforms. And just to be clear, like OpenAI has bets across multiple distribution shifts. So the shift from search to EO right is chip number one. That's pretty clear and obvious at this point. And yeah, there's this whole debate about is I a different than SEO blog? It's missing the point. The point is, is that a huge part of your target audience has shifted their time spent from one area to a new area. And they're not only spending more time in there, they're behaving differently. You know, you have to be where your customers are. And as a result of that. So that's number one. Number two is the app platform. Right. And that's kind of still developing. But they also have these other bets, which is kind of crazy to think about which is they have Sora, which is kind of in the line of social. I that's I have no idea, like where, where that ends up going. But it's interesting that they are going out there and they've been actually very public that they have a device. Bet two. So maybe they go after the phone mark. So that's why I kind of say I feel like it's OpenAI's world. We're all kind of living in it. They don't need all of these to play out, but it is interesting to see that they when I talk about the distribution shift, I'm not just talking about the first category, I'm talking about all these distribution channels collectively. All right. So, Brian, I know that you're doing a lot of hiring over at reforge. You're probably looking at a lot of resumes. I think Kyle even mentioned you're hiring a go to market engineer. I'm not really sure that exists yet, but Kyle keeps telling us all it us, which is pretty neat. So, I'm terminally online, and I saw, I saw, the CEO of Replit retweet a resume that he looked at and this guy under interest, he actually put his powerlifting axes on there, I guess. Squats, squats, 365 pounds and benches 275 deadlift. Wow, 475 pounds. Brian, would you be impressed by this? Would you hire this man? Yes. Yeah. Well, no, I don't know if I would hire this from this guy, but I think, oh, would I be way more interested in this resume versus one of the I slop auto generated things that I get ten x? Yes. Like, we don't even post a bunch of our jobs at this point because we know that we're just going to get AI automated crap. And it's it's more time for us to sift through it then then trying to find the maybe the the needle in the haystack there. So I don't know, I don't know what role this is, is for, but at least shows some personalization and humor and, and other parts of itself. And this would catch my attention, for sure. You guys are. Putting your job postings in gyms. Yeah, yeah, yeah. If you're a high rocks competitor, send me your high rocks times and, and I'll and I'll know you listen to this device. Use an AI no taker at all by Chad's. I use fixer because you recommended it nicer. I do use one. I haven't heard of fixer. What? What's the ten second pitch? So fixers like an email, calendar and notetaking kind of app all in one. So it's kind of superhuman granola. And I don't know what the other one. And calendly, I guess, all in one I app, I do, I use, I use granola. Yeah. So as, as, as the rest of our team, I think we also use gong on the sales side, as, just as, as a thing of habit. But we'll see how long that lasts. Yeah. Feel like granola is the darling these days. Well, I keep getting shocked at how big the market is. And fireflies is another company in the space. They have, raised at $1 billion valuation. And recently they shared some of the back story of how they got there. And the founders actually pretended to be an AI note taker before they had an AI note taker. They join meetings on you, take notes by hand while sitting silently, and then send the AI generated in quotes, notes ten minutes later. Now there's like, fake it until you make it. And then there's whatever this is, which is wild to me. Yeah, do things that don't scale, then this makes sense. There's a term there's a term for it. Kyle, have you heard? You know what Flint Flintstones is? So I call it that. I love that analogy. His car doesn't go. It's his legs underneath, like. Oh, like that as it. They're doing. They were Flintstone ING. I like it there at Theranos. This was just called fraud. But, it is different and I cannot bear it. The fine line, fine line between. Does this make sense that I'm surprised by the market size? You. Did they announce revenue with this? I'm interested. I did not see it. But if if they're worth $1 billion in a category like this, I have to imagine they're, you know, getting close to 100 million. There are. Yeah. I mean, there's I've seen metrics on multiple of these. And like, all of them close or above 100 million. And so every knowledge worker is essentially the TM here. Right. Like the Tam here is every single knowledge worker. And maybe beyond that to maybe beyond the knowledge workers to, over, over the long term. So it's an incredibly wide market. But just like I think like AI coding, you know, the vibe coding and stuff, I think this is all going to segment and specialize over time. And so I'm interested to see how it how it all and how it, how all these players end up chunking up to create more defense ability and higher price points and stuff. Clearly the ly on everything app, they just pivoted to this. Yeah. Yeah. So when in doubt and I note taker, there are some actually interesting companies in vertical specific notetaking spaces like I know Take Care is for wealth advisors, or in the medical space. And I think it's kind of a cool wedge to become more about system of record for these companies and, and firms. Yeah. I mean, there's essentially to two directions. It's either becomes a feature of some vertical kind of a software stack, you know, for a specific audience, or these things expand horizontally into whatever the replacement is for the notions. And Google Docs of the world like that probably ends up being the two baths. Right? Ben, do you want to bring us home with the intro of something? I tried this week? Sure, yeah. Every every week the fellows, they try some some different things and this week is no different. Does anyone, any of you three got something you tried this week? It's worth talking about? I tried something partially because it's trendy. But I it, spark my curiosity. Have you guys heard of a company called Gamma Abrams? Yes. So it's, trying to be an AI made of PowerPoint, essentially disrupting the product that, you know, it's been around for decades now. We all love and maybe hate. And, they claim that they've hit 100 million, er, profitably and with 50 people, they also just raised a bunch of money to, to go after this opportunity. But I was curious to give this a try. One of the things that I actually really liked about it. So when you go through the onboarding, most apps these days just drop you into a prompt bar UX where you kind of type in what you want, what you want to build for you, and then the agent goes and starts building something. But with gamma, when I signed up, actually gave me four options I could paste in text, essentially from notes and outline or existing content. So like paste in text, they'd make something from it. I could use the prompt bar to generate something. So like a one line prompt, or give it a few seconds and it'll kind of generate something. You could import an existing file or URL. Or you could remix a template. And so I just like that variety of onboarding. And then it was a really interesting onramp to essentially, you know, you could make PowerPoint slides out of it in a full deck, but you could also get started with carousels, single social images, like I think the the PowerPoint messaging is a powerful brand narrative. But the reality is it's actually, I think, more of a Canva that's trying to get to PowerPoint than fully PowerPoint. So far, yeah. I'm really interested to see how karma collides with Canva, at some point in the future. So this is like one of those products I started. I tried really early on, and it just wasn't at the bar for my specific use case, but they just kept cranking on it and found different use cases where the technology did meet the bar. And I see this across the AI landscape like they're, you know, early on in the AI avatars, they weren't really good enough to do a lot of use cases, but the one use case that it did do was good enough was like all this internal compliance training, there was kind of a minimal quality bar, and people were just looking for efficiency. And so that's where, like, I think it's contagio, kind of came out and has done really well, 100 million, you know, r and, where the avatar technology was like good enough. And so I think a lot of the question about AI is just where, where can you create a good enough experience to get above that threshold where we, you know, started this conversation and it might start in a very specific subset to the segment of the market. And as the technology gets better, you just kind of unlock more, more and more, adjacencies. Either you guys have something you tried this week. I try a lot of things. So in the AI growth course that Kyle, you, very graciously, contributed some thoughts to we have this whole module on retention engagement and how AI's impacting different types of onboarding experiences. And so I was trying a bunch of products through that. The two that I found were most interesting that created these like really interesting experiences for me were 40. Have you have any of you played with 40? No, no, I haven't seen that yet. You won't be able to do because it immediately pops you into WhatsApp as an experience. But so when you onboard, it pops you into WhatsApp, asks you a couple questions and then says, hey, can I give you a call and it calls you in? It's an AI voice agent starts asking you about your your career, what you're doing, what you're looking for. Like all this type of stuff. It's a really like dynamic conversation. It's done in five minutes or less, and at the end it's like, oh, I think I have. I have three other people in the network that would, that you'd really enjoy meeting, and it goes and pings those people off, right, to really personal emails, to all those people, based on all of that information it collected you for in this like five minute conversation, if they opt in and it sends like a really personal email introduction back and forth. But it was really interesting because you can see that rather than like an onboarding experience that just had form fields, just was like a few minute call, the amount of other interesting stuff that you would never be able to capture in the form fields. There was like tons of very similar experience, but the product is like this alien body. It's targeted towards a younger audience, but the alien chat body that it presents to you at the end is so personalized. It's not just about what it talks about, but how it talks to you, as well as other things. There is this emotional like resonance with me that was like almost like it was a WTF moment. I was like, oh my God, how how am I emotionally resonating with this thing? I just like create it in a few minutes. So I highly recommend go try those two products in onboard them like you were actually trying to use them. Don't just like flip through the mindlessly. It's they're really interesting experiences. Very cool examples. Well, thanks for joining us on the pod brand. Thanks for having me. Mostly growth is a mostly media production executive produced by Ben Hillman. That's me. Nothing said on this podcast is intended to be business or investment advice. It's the sole opinion of Kyle and Ben. Just some guys who love talking growth mostly. If you like this podcast, hit subscribe, give us five stars or hit like the Algorithm overlords. Thank you. Drink water. Call your mom. And have a great day.

• Relentless curiosity, radical accountability, and HubSpot’s winning growth formula | Chris Miller

  Mentioned as the person who introduced Chris Miller to Lenny.