PLG isn't Passive: Multi-Product Startups + AI Distribution Wars

You know, we've all seen those like those developers, solopreneurs who have in their Twitter bio like the cringe thing of like five apps with their MRRs all listed out. I don't think it's cringe. I like it. I know you like it. That's why I call it cringe. I know you love that stuff.

Hi, everybody. We're back. We're sorry we went out for some time. It was mostly because of me. I went for vacations in the Nordics with my kid and wife before my wife has to start work. And Hank also went to SF, I think, and did some traveling as

well. Just a little bit. You were gone so long. My hair is kind of getting awkwardly long. I need to do something about it. Me too.

But today we have interesting topics. Today, actually, we're going to talk about strategy. We're going to talk about strategy for like a seed/Series B startup, so smaller size startups, and then strategy for bigger startups. And then in the middle, we'll add some spice with the Bolt event. So it's not only strategy, but I present the first topic. We actually wrote a blog post about these on our go to market website. But most of the PLG companies that you know eventually become enterprise. Like, I think that's true for almost all of them. They all start with the idea of self -service, but then they realize that the money is on the enterprise. So eventually, even though they have growth to get people to come and to sign up and stuff like that, most of the revenue still comes enterprise. That was true, for example, for when I was at Auth0, it was true for a Clashyard. That's even true for Stripe. One of the companies that I've seen that has actually the most self -service still is Vercel, but they're around 50 -50, which is very interesting because they are huge and they still have a lot of self -service. Having said that, one of the companies that I advise is called Clerk. Clerk helps with authentication, and they didn't want to go enterprise. They've done a fantastic job. They've done 5x in revenue last year. They are planning to 3x again this year, and they've done all of that with self -service. And their idea now is instead of going enterprise, going multi -product. They just shipped Billing, which is a new product to make it really easy to show pricing plans and subscriptions, to get people to subscribe and then see a subscription and manage it. It uses Stripe inside, but they basically built all of the infrastructure around Stripe so that people can basically use it for having people self -servicely subscribe to their startups. And what they argue is that because they already have all of the authentication data from the user, it's not easy to save the information, the credit card, plan they have, and then just add like they already have a component for showing profiles of users and showing what subscription they have and how to do it. It's already there. So you don't even need to add code. Their idea is with billing, we're going to generate more revenue with the existing self -service customers instead of having to go enterprise. And their plan is to add more and more and more multi -products. What's your take on this? Like, do you think they'll succeed? Have you seen other examples? Like, how do you think about that strategy?

I like a multi -product strategy that has a couple of these ingredients. So one, it's a product line that could be big enough that it actually could become the main business. And what's his name? Is it Jason? Who's the SaaStr guy? Jason Lemkin. Jason Lemkin. He had this great bit about how he used HubSpot and Shampoo as examples of this. He said HubSpot did this great thing where they launched their CRM as their second product and CRM as of 2024 is now their bigger product, bigger than their original product marketing automation. And the shampoo comes into it because he had this great little metaphor of like, if you're a shampoo company and you want to launch a second product, the natural thing to do is to add conditioner, but fewer people will always use conditioner than you shampoo. So your second product will take just as much effort and distribution and everything, but it will never be a bigger, better product. So you should instead launch products that have the potential to eclipse the first product and HubSpot did that successfully. This feels similar to me with Clerk where they have off, how they have this whole billing thing built on Stripe. That's adjacent. It's related. They have advantages in it, but it could become, maybe it'll be bigger. Maybe it could be a huge thing. Maybe that's what they'll be known for one day. I like those types of launches. There's

actually existing products that only sell what Clerk is doing for billing. Like there's companies that actually sell, let me help you build your payment infrastructure and the plans and everything. So to me, it's interesting that for them, a new product that is integrated is literally the product for two or three other competitors.

Yeah. And this happens a lot when you have a company growing because then you get this distribution engine. And like you said, too, they're focusing on self -serve. They're Auth0, whatever. And that's interesting. Like you said, in that blog post we wrote about, okay, people often have to do enterprise. What that blog post is really about is a lot of founders fall into this erroneous thinking that PLG and focusing on self -serve is some sort of magical bullet where they don't have to hire sales or marketing and the product distributes itself. What it is is this, is figuring out, oh, how does your product have different synergies with other parts of the company? And how do you get different departments to all adopt your product at the same time? And it's not just passively allowing it to happen. Clerk is very good about pushing, about marketing. And we had another example when we were talking ahead of the show about our notes was CodeRabbit. They're also very focused on self -serve and they are not slouches when it comes to the distribution. If you follow any sort of dev tool content creator, you've seen an ad from them. They are doing the influencer marketing thing really well, and they're pushing, they're pushing on ads everywhere. They're not just hoping that PLG, that the product sells itself. They're actually forcing it out. So that was kind of another thought and comparison I had.

I think that makes it very different. Like in the past, we always said that for all dev tools, when you were doing PLG, it's always to eventually do product -led sales or to go to enterprise. And I think that when we said that, we were mistaken because we were talking about how most companies do it, which is the idea of we just have self -service and then people will come and it will just work. But in Clerk's case and others now, but I think they were the first ones who started to do this, their thinking is different. It's more around actually having a strategy on it. It's like, okay, we drive expansion from existing customers with new products. We'll get revenue through that. We'll also sell eventually this separated so that people can buy us for this new product. And I think the timing is right. Like in the past, you could only sell to drive the revenue to enterprise because there weren't that many programmers or companies that you could sell. But now with AI, there's starting to be a lot of these lifestyle startups that people just build an app with vibe coding and a bit of one or two coders, and that's it. And they ship an app or there's now a lot of solopreneurs who are using like lovable, for example, or something similar. And I think there's a huge opportunity now to target and sell services to a lot of these people who have one or it's just a solopreneur or it's a team of three. And I think there's going to be a lot more PLG self -service businesses in the era of AI, just because teams will stay small and continue small and still drive a lot of revenue. So what clerk nailed, I think, is timing. Like I work with them. So, for example, their signups starting to grow like crazy in the past few months. And they realize that a lot of them are actually coming from Lovable and v0 and Bolt because it's these people and Lovable, most of the people don't even code, but they are still trying to buy authentication and clerk is trying to decide on, do we want to serve these people who are not developers, but still need authentication or is it a mess and we shouldn't do it? And they are thinking about it. And that's the right thing to think about because then how you think about your product strategy is different than who the user type is, but also what the next multi -products are. I know they're actually already building the third product that they haven't shipped. I can't share yet what it is, but they are already working on what's next. So it's not just winning, they're more coming.

One thing to think about with Lovable, Bolt, v0 is more than ever, and just cursor and co -pilot more than ever, a single developer now is shipping multiple apps and may be monetizing multiple apps on the same products. And one thing I've had a discussion with with some teams is it's less about how many developers there are. It's more about how many apps are being created that could use your product. And I think that's something that clerk is a great example of being well poised for that. How many apps is clerk going to be used in? It's the same with the databases that are being recommended in all these products, because if you win a developer, you might win 10 apps over the next year because they're getting so productive. I guess that's kind of a new framing on PLG is it doesn't have to be about getting into a large team where you can grow across departments. It can also be about how do you get to the most productive developers that are going to produce the most apps and ship the most things. And, you know, we've all seen those like those developers, solopreneurs who have in their Twitter bio like the cringe thing of like five apps with their MRRs all listed out. I don't think it's cringe. I like it. I know you like it. That's why I call it cringe. I know you love that stuff. And I mean, it is cool. Like it's it's great to see that success. But yeah, like one developer who ships five profitable apps, it can be the same as five customers.

And they all get a lot of users, for example, for clerk or a lot of people that are paying or something similar. What clerk also did really well was in their latest round of funding. They actually got big investment from Stripe, who is now their partner for the billing. So they're investors, which means it's likely easier to for them to get some type of discount or something similar. That's number one. And number two is billing. They charge per transaction similar to Stripe. So they just charge an extra percentage on top of Stripe, which means that it's easier for people to understand how it works. To me, it's going to be interesting to see how they evolve it. And see, I love your framing on it. Billing could be the product that eventually clerk sells and that most people want to use. But I actually think that the main value for them is just how easy it is that once you have one of the two implemented to add the other one. And that's what I think will add a lot of value in the new era of AI, where it was just already have something you don't need to even touch the code. It will automatically show the others and will make it a lot faster. That's, I think, going to be a big differential for a lot of these new SaaS. And I think multi -product is going to be a new strategy that is used more and more common.

So, OK, so what are our takeaways and principles here? I think, one, it's OK to go multi -product. A lot of times VCs discourage that because it splits your focus. But if you do it in a way where, hey, actually, even though these two things look really different, they're part of the same developer experience. Like a developer working on auth also has to work on billing. They're both heavily security involved. If you can trust one, then you can definitely trust the other. The way those tie together, that's one less integration to do. That's all interesting synergy. It's also OK to not try to go for the enterprise if you're pushing hard on distribution, if you're finding ways to get out there and you're not being passive. PLG is poison when you think it's passive. I like the takeaway

that you should build the new product that could potentially in the future be your main product. And then the last one, I think, is it's very important to start thinking now of a new user type, even if you're a Dev Focus company, who are the vibe coders or the very productive developers, which I actually think are very different. But it's understanding the people who use the Lovables and Replit agents of the world are the vibe coders or solopreneurs. And then the very productive developers might be using Cursor and v0 and Bolt. And I think you should consider them different type of users. But think about what's your PLG strategy to target both. And with that, since we're talking about Bolt a bit, maybe we can switch to the Bolt topic.

Yeah, I've got my swag bag from an event I went to, Unleash the Vibes. Yeah, I was in San Francisco, not for config for some different reasons. But while I was there, I went to Bolt's party thing, Unleash the Vibes. They had the Chainsmokers there before, like they had this whole concert with, dinner and drinks and stuff. I dipped out of there after two songs. I'm not there to lose my hearing and jump around with a bunch of nerds. So

boring. I know, I am boring. But the lead up to it was awesome. They had a little park outside where you had, they had a couple food trucks to earn your dinner. You had to do a couple activities around the thing, just fun stuff. So I got a very nice burrito, talked to a bunch of cool people. There were a bunch of really amazing people there. And it was interesting because they were, yeah, there were a lot of like developers clearly, but there were a lot of designers. It was config. And we're in this interesting age where vibe coding allows all this blending of, well, who's creating the app? Is it designers who made a Figma file? And now they are like, I can take this to life without a developer on my own. And they're trying that out.

I think it doesn't matter where it starts, but this role that Vercel has, and I think other companies have, that's called design engineering. I think it's going to conquer the front end world. Of course, Figma will say, designers can now code and they don't need developers, but then the v0s will say like, no, developers can actually now design, but it doesn't matter. What matters is that in the end, you get the best designers and the best developers to become more of a hybrid role of design engineering, which will make them more productive.

Yeah, it absolutely does. Having worked with many design engineers and basically the bar for what a design engineer is, is getting lowered. What do you think they did really

well? Did they do a good job of sharing it social media? Did they do a good event live there? What are your thoughts on that?

It was definitely an awareness play more than anything else. So I think getting a bunch of people who are at config, very aware of bolt and that, Hey, this thing's available and powerful now, which was especially important because of what Figma released. They, you know, Figma has its own Figma to code stuff now and Figma to app type of stuff. Now they've really evolved dev mode or gone far beyond that. So I think that was an important play to get that audience and to take them outside of the config building for just a little minute where they could be like, Hey, we're here. And, and, and we do this stuff that was important. And then they, it was also largely to hype up their upcoming hackathon. They still haven't started their $1 million hackathon. Uh, that starts later this month, I think next week. And that's going to be interesting to watch because they've just put a million dollars of prizes up. Who knows what's going to happen? I'm very curious about that hackathon. I know you love a hackathon, so I'm sure

I love it. I actually love it that it's a one me, like I used to not believe in hackathons for selling because people who went to hackathons were students that were not like your enterprise developer that would buy. But now I actually think that anybody can go to a hackathon and anybody can build something as we were talking on, on the clerk topic. So the game is changing. Like how you think about marketing for developers, how you think about P and G on developer marketing is changing so much now. It's going to be interesting to see what happens in a year from now. Last recent leak of the OpenAI strategy. I don't know if it was because of the trial or how it happened, but there was a big leak of the OpenAI strategy, which we're showing now on the tweet. If you're seeing the video here, otherwise you'll see it on the notes. But the TLDR was that their thinking is that this year is the year of agents first for coding. And then it's the year for what they call like a super assistant. So the objective for OpenAI for this year is to ship a super assistant who can help us answer emails and answer chats and book as a trip and sending emails and calendar management and stuff like that. So basically replacing an EA with AI. To me, however, the most fascinating thing about that is that they actually talked about competition on that leaked OpenAI strategy. And when you think about who could fight versus OpenAI, my first thought before reading that was they're scared potentially of Google because they have Gemini 2 .5 pro that is actually so fucking good. And it seems that they are doing better models now than OpenAI. Like, could they win or that it will be anthropic that has pure and they are doing great. But actually OpenAI is the most scared about meta. And to me, it was crazy because I know if you saw, that's always the

right answer.

Everyone should be terrified of

Zuckerberg.

Zuckerberg is awesome. I love Zuckerberg. Very founder mode always. But to me, what's fascinating about it is Meta just shipped Llama 4. Llama 4 was so fucking bad that 80 % of the Llama team is gone. I don't know if they resigned. I think they were fired. Like Llama 4 was so fucking bad that I think they fired 80 % of the team. And even if Llama 4 is bad, OpenAI is still scared of Meta. And it's not because of Mark Zuckerberg. It's actually because of distribution, because most of the world and actually now a bit more in the US uses WhatsApp to communicate. So if you could get AI in the app that you used to communicate like that one or in the app that you used to share what you're doing on your day to day like Instagram or in the reels, which if TikTok ends up being banned, it's going to be even huge. The main thing that OpenAI is scared about is the distribution platform. And I think that speaks to how important distribution is. What's your favorite distribution, Hank?

Hey, it's nice to know we're important, you know, smart guys. Exactly. Because that's, it's true. Like the product, no matter how strong the product is. And I don't know, I think frankly, OpenAI, I think it's good for them that they're thinking about this. Because I think their distribution has been weak. Their marketing has been weak. They've relied so much on the strength of their product and being ahead on that. And yeah, their distribution has been lazy. I mean, their naming is lazy. Even Sam Altman is tweeting about how like, I'm sorry about the names. Well, we'll fix it. And then they released 04. And I'm like, wait, I've been using 4 .0 for five months. I don't know. I don't know which of these is which. But I think we're going to see OpenAI get better about distribution and become more conscious of it and not so reliant on the strength of their models and start really thinking about, I mean, they have with Microsoft, their partnership with Microsoft was largely, you know, it was a, hey, we can get into Windows and stuff like that a little bit more. But that's not a great distribution play as, you know, the only Windows user on the call here. Yeah, they've got to think about it because Google, both Google and Meta have such an advantage on distribution. I mean, I see that dumb little sparkle emoji everywhere in my Google products. It's in Drive. It's on Docs. It's in Gmail. But I think Meta has

even higher distribution with the regular human being like Google has a strong distribution on people who work in the enterprise or in companies because of Google Docs, Google Sheets, Google Meet, et cetera. But Meta has WhatsApp and Instagram, like all of the fucking world uses them. And that's, I think, their killer edge. And to be honest, it wasn't in their strategy, Doc. But if I were OpenAI, I would probably focus on building a phone and building a browser. That's your distribution. If you can get the next Chrome, like buy Arc, which is getting killed or something like that, do a new phone or a new Android operating system. But do something. I'm trying to save.

Quit trying to save Arc, Gonto. You're hanging on to a zombie, man. It's Arc's dead. I don't know. I saw a tweet that somebody

suggesting Raycast by somebody by Arc. But anyway, it's fascinating that distribution matters so much. And I think a lot of companies don't think enough about distribution. And if you only focus on writing the best content, doing the best YouTuber content and stuff like that, you're not going to win. Which is why, like Clerk and CodeRabbit, for example, as we talked before, have done this huge influencer marketing YouTubers strategy. And that actually works in most cases better than if you have a really good internal YouTuber building YouTube videos, because you're nailing distribution before content.

Yeah, I wonder what it'll look like. We'll have all these competing models. Oh, another related footnote here is co -pilot. The AI editor is free in VS code now. So that's an interesting distribution play by Microsoft and GitHub, you know, fighting against Cursor and Juni and Windsurf. I

know. We'll see what happens. But I think it was very interesting on how both Clerk, which is a SaaS company and OpenAI that is, I don't know, like a fucking monster, both think in their strategy about their distribution and the type of users. And that's a big part of their strategy. And I think as a takeaway, I see so many startups that don't think about distribution, that don't think about what's their go to market strategy in that sense. And I think it actually is important to stop. Most of them just think that if they do PLG and it's passive, it just works. It will continue to work. But actually giving some thoughts into it is a worthy spend of time.

Yeah, I think people would think more about the importance of distribution. If whoever came up with the PLG term had called it PLM, product led marketing or product led distribution, even because so many people are too passive about this. And if you're passive, then the people who are active will eat you for lunch. So OpenAI is aware of it. They're going to make some interesting changes. I bet the next year, the next two years, we're going to see some really interesting up -leveling in the marketing for all of these companies on their AI stuff.

Everybody needs to think out new strategies on distribution, on new type of users, as we were talking about, like v0 users, Lovable users, solopreneurs, and stuff like that. It's just very interesting time if you're not a lazy marketer, because things are going to change so much, I think in the next year or two, that marketing, I think, is going to be one of the most fun jobs to have.

I agree. Very strongly agree. Anyway,

we did a very fucking long episode, but we're back after three weeks and we're going to chat. So thank you all for listening. As usual, if you have feedback, let us know. But otherwise, we're back to our weekly cadence. So you'll hear from us next week. Cheers. Cheers.