Stop Vibe Coding. Start Getting Customers

Episode Description

Building cool AI apps is exciting—but what if no one actually buys them?

In this episode, we explore why "vibe coding" alone isn't enough to build a successful business. Learn how to shift your focus from endlessly creating products to solving real customer problems, validating ideas, and generating revenue. Discover practical strategies for finding your first customers, refining your offer, and building something people are willing to pay for.

Whether you're an AI builder, indie hacker, entrepreneur, or developer, this episode will help you stop chasing trends and start building a business that grows through real customer demand.

🎧 Tune in and learn how to turn your code into customers.

Transcript

SPEAKER_00 0:00

Today, we're going to talk about how you can build anything with AI, but people are going to need to figure out distribution. How do you actually get customers to the thing that you're vibe coding? I believe that the wealthiest people will be marketers over the next 10 years. But every YouTube tutorial is telling you how to vibe code. So today, by the end of this episode, you're going to have seven really straightforward things that you can do this week: frameworks, ideas, alpha that's working right now, uh, so that you can actually build a software company, a SaaS, an agent company that actually gets customers. So I'm going to explain how to do it. By the end of this, you're going to get your creative juices flowing. I needed to make this episode. I really needed to make this episode. I'm actually recording at night. And, you know, I tweeted this that you know, people need to figure out distribution, you know, and it hurts me that people, you know, don't have the ideas around it. So like and comment if you want me to do more episodes on this. Let's get right into it. So I believe that there's this great flip happening. You know, I moved to Silicon Valley in 2014. And uh when I moved, it was all about engineers as number one. I actually studied computer science in university because I was like, engineers are number one on the hierarchy, right? And then it became product uh and then marketing as at the bottom. Like literally the laughing stock of Silicon Valley were marketing people. Now, today I believe that you know, because of AI, distribution, people who understand distribution, how to get customers is, you know, you're at the top of the list. Then it's product people, and then it's developers. Um this is uh you know a big deal because I actually think that not a lot of people understand drip distribution. They don't understand brand, they don't understand advertising, they don't understand content. Um, but I think that you know, if you can, you know, if you do understand it, you are at an unfair advantage. Um, you know, there's 200,000 new vibe coding projects being created every single day on Lovable. Um, how many people are actually seeing those products? Probably not a lot. Um, Peter Levils has three million dollars plus of revenue. He has zero employees. The reason why he's been able to uh you know get so much revenue as a one-person business is he's got 750,000 plus followers, I think. And also he's got great SEO. I mean, if you look at some of the stuff that he works on, you know, Nomad List, for example, a lot of the tools, uh like these could be you know copied quite quickly. Um, it's it's essentially a directory. I've done episodes and tours on how to vibe code directories before, you know, but he's he's obviously got good data and data's part of that, but it's also largely he's got an audience, he's got trust, and he's built good SEO. We'll talk more about that in this episode. So the big trap uh that I think a lot of founders and builders do is they vibe code something, you know, once they've built the product, they've got this great idea, um, then they're just gonna try marketing things, but it's just silence. So then they're gonna build more features and they're gonna launch again and more silence. So they basically are just thinking about how do I build a really good product? Um, and if you build it, they will come, but they don't come. What smart builders do is they start with the distribution. So they'll grow an audience uh to maybe a thousand people, they ask that audience what they need, uh, and then they go and build it in a weekend, 24 hours, 72 hours, a few days. Um, and then that audience is shocked that wow, you built the thing that I wanted. And then you launch it to a warm audience, you iterate with real users, and you start making money. So it's distribution first, product second, always. You know, the example of Peter Levels, I think I, you know, I think he's tweeted 125,000 times in the last like eight years. You know, he's he's put in the consistency and the work to actually go and build that audience, and he's been iterating with users and customers, you know, with them. So let's get into the actual tactics. Uh Strategy one is going to be about using an MCP server as your sales team. So, you know, if you're listening, if you're a listener of Startup Ideas Podcast, you probably know what an MCP is. Um, but you can think of it as basically like an uh an app. Um, you know, the example I'll use here is you know, plugins in OpenAI, use the MCP protocol. So you can just be like at Google Drive, find the right drive, then summarize a doc and update a sheet or edit a slide stack. So there's this massive opportunity to build MCP servers as your sales team. How does it work? Well, a user asks the AI, it could be you know Claude, it could be Chat GPT, as a question. The AI discovers your MCP server and AI returns uh you know your product and the user gets value from it. So there's zero CAC. The AI assistant basically becomes your sales teams. And I have friends who are, you know, I'm actually incubating uh some things, you know, around here. And actually, if you look at ideabrowser.com, you know, we we've been doing some MCP stuff. But I also have a lot of friends who just, you know, in in, for example, in the fintech space, you know, 150 plus installations in 30 days, $0 ad spend for something that he vibe coded uh quite quickly. So I believe that building an MCP server in 2026 is in a lot of ways building, like kind of building for mobile in 2010. There's just a lot of opportunities to show up in LLMs in these connections, plugins, or MCP servers. So how can you start uh this week? And then by the way, I realize that this, you know, an MPC and an MCP server, it's not for everyone, right? If you're building a boring business, maybe that's not for you. Or you know, even some software businesses might not be for you. But for some people, for SaaS, you know, if you want to get started this week, um, what can you do? Identify what question your product answers. That's the first thing you're gonna want to do. Number two, you're gonna want to build an MCP server that returns that data. You can vibe the code that realistically in 24 hours or less. You're gonna want to publish that to MCP registries, uh, Smithery, MCPT, Open Tools. And then every AI assistant that connects to your server is now selling for you 24-7. So that's a bet that you're gonna make. Um, and that's strategy number one. Let's get to strategy number two. Strategy number two is programmatic SEO. So it's how can you create 10,000 pages, SEO pages, in 48 hours? So you're gonna want to look at you know a keyword, a keyword pattern, best X for Y. So best uh CRMs for dentists, for example. Then you're gonna want to do uh look at uh data sets. So you're gonna want to, I actually did a whole episode on Firecrawl recently, a tutorial on how to use it. Um so you're gonna wanna, but basically, what does that do if you haven't seen that? It basically goes and scrapes and gets you clean structured data uh so that you can actually provide value in your pro in your pages. Uh then you're gonna want to do use a page template, Next.js, you can use Cloud Code or something to actually go and build this. Uh then you're gonna wanna actually create AI content, and then you're gonna go and create 10,000 pages. Now, I know there's a big gap. Uh I can see the comment right in the comment section right now. It's okay. Oh, you're just gonna press one button and then just automatically create 10,000 pages. Uh well, no, of course not, because you you know, you you do need to make sure that the content is good and doesn't feel like AI. So there's gonna be a lot of optimizations that you're gonna have to do to actually get it so it doesn't feel like AI. But once you get it to there, I mean you can start with just a few pages. And once you get it to there, then you can start scaling to more pages. Um, you know, the math is is pretty simple. If you end up creating, you know, 10,000 pages and each page gets 30 visits a month, which you know, it's not a lot of visits, right? Um, that could go up to you know 300,000 monthly visitors. If you convert at 2%, that's 6,000 conversions a month. Maybe that's $10 each, and then all of a sudden you're making $60,000 a month from pages you built once. Now, those 30 visits a month, that's not going to happen overnight. Um SEO does take some time to compound. And but like think about it directionally, right? Directionally, the idea is build a lot of pages, get really good at creating AI content, get structured data that's ex that's high quality, um, and uh and and and rinse repeat. So, how can I actually start this this week? If this is something that's interesting to you, how can I go and start this this week? Well, you're gonna want to pick a keyword pattern, so like the product type for niche or service in city, um, like for example, the CRM for dentists or uh you know uh roofing in Miami. You're gonna want to build your data set. So you're gonna want to scrape with Firecrawl, or you can use existing databases. You can create a template in your framework, like you know, NestJ, Next.js, Webflow, WordPress doesn't, you know, just stick with the thing that you're gonna use. And then you're gonna want to use AI to generate unique content per page, so not just variable swaps, actual paragraphs, high quality stuff. Um, and you know, you might also might want to have a human in the loop to actually edit and look at some of this content. Publish a hundred pages as an MVP, monitor indexation, and then that's when you can go ahead and scale. So that's number two. I think that's applicable to really anyone. It doesn't make a difference if you're building a service, a SaaS, an app, you've got a boring business, you've got an agency, programmatic SEO is still under tap. Strategy number three, uh, three on seven, is doing a free tool as a top of the funnel uh for your top of the funnel. So the tool is your marketing. Um, you know, the example I would use is uh href for for a while has had this backlink check uh checker. So, you know, you you put in your domain, for example, greg eisenberg.com. It's gonna see what are the different backlinks, like who's uh connected to me? Uh what links are connected to me, um and you know, it's high quality and stuff like that. And you know, but the the reality is it gives you a taste of their full product. So if I want more, I have to go and spend hundreds of dollars or thousands of dollars a month with hrefs. So they kind of hook you in through that. So, you know, step one is building a free tool. It could be a grader, it can be an analyzer, it can be a calculator. Uh step two is the user gets instant value. So, you know, your site scores 43 on a hundred. Um, generally, what you want to do is obviously grab their data in some capacity, you know, text message, email in order to get access that. Uh step three is the user shares their score. So that's gonna help this thing go viral. Uh, social proof, back backlinks. Um, and then step four is more users are gonna discover it. So you've got this organic viral loop, and step five is upsell to a paid product. So you want to fix these issues? Well, we've got the product for you. Now, this uh strategy has been around for a while, right? Um you know, but why is it different? Well, now it's different because you can use Claude Code uh to go and actually build these free tools, right? It's now like not crazy to build a free tool a week, whereas in the past, these things would have taken you months. So you can vibe code a free tool in a day, ship it by lunch, and it markets itself forever. So uh I really like this one. This is something that I'm doing in all my businesses as much as possible. Um, and if you want to get started this week, I would just say uh, you know, I would ask uh, you know, your LLM of choice to be like, here's what I'm working on. And give me 10 ideas for free tools that can act as my top of the funnel. And that's you know, you're gonna have your 10 tools, you can prioritize the best possible the ones that you think are the best. You can even ask your audience, hey, what would be more interesting, this or that? Um, get them involved and then go and build it. And then just, you know, we we think of uh when you do social media, you you think of like an editorial calendar, a content calendar, but no one really thinks about like a free tool calendar. And I really do believe the tool is the marketing, this is working. Don't know why more people aren't doing it. Um, but strategy number three, one of my faves. Strategy number four, answer engine optimization. So you want to be the source that AI cites. So the old way of you know getting seen obviously was SEO. You just wanted to get, you know, on Google page one, you do 3,000-word blog posts, you're doing backlink building, keyword stuffing. Oh my god, um, click through your site. You know, I'm just having flashbacks of all of doing all this stuff. Um it's there's still opportunity in SEO. I mean, we talked about it today, but it's declining um and zero-click searches are growing, right? So um if AEO is in 2026, is where SEO was in 2010, first movers will own these niches for years. So uh the idea is answer engine optimization. How do we get cited by ChatGPT and perplexity? You know, structured direct answers, you can do an FAQ format, schema markup, comparison tables that AI can parse so that you know that way uh yeah, you're getting seen a little more. Um and then AI, AI is gonna send users to your product. So um, yeah, I think it was Peter Level's actually uh I believe that he said like his AI referrals jumped from 4% to 20% in one month. I think that it's only going to increase for you know across the board on e-commerce, SaaS, apps, you know, you name it. So um it's worth kind of uh uh exploring. It's worth exploring. So okay, great. You know, you probably already knew this. Obviously, go you know, you want to be seen in Perplexy or ChatGPT, but how do you how can you start this this week? Well, Google the top step one is to Google the top 20 questions your customer asks. And you're gonna step two, write the definitive structured answers for each. So I'm not talking about 3,000 word fluff, I'm talking about clear, direct, citation worthy. It's all about getting citation worthy content. Then you're gonna add step three, add schema markup and FAQ blocks. Step four, you're gonna want to publish on a domain with authority or build authority through the other strategies. Um, obviously, it's bet you know why a Peter Levels could could get his SEO and AEO up quickly, quick more quicker than the average person listening to this is because a lot of his domains have built authority over time. So uh, you know, there's no better time. Start building that authority today. And then lastly, monitor your perplexity in Chat GPT citations. You can use tools like Auditerly, there's profound, there's manual testing. Um, but the important thing here is that you're just getting started. So get started this week um on AEO. Strategy number five is how can you make your outputs of your product shareable? So we all know you know, the end of year December Spotify wrapped. Um, they get 100 million shares every December of that thing. It's absolutely crazy. Um everyone wants to share it, your mother wants to share it, your sister wants to share it, your kids want to share it, everyone is sharing it, laughing about it. It says it says about something about your identity, who you are. So, what is that for you, right? So, what is the free distribution output shareable thing that you can be creating? You know, the way I the way I the way I have it in my head at least is you know, ask what does my user want to brag about, and then make that thing beautiful and shareable. Uh for GitHub, they've got the contribution graph. Um, and then you see like devs bragging about sharing those green squares on Twitter. Uh Stripe Atlas, which allows you to create companies, you know, the artifact there is the incorporation milestone. Founders tweet, uh, I just I just incorporated, or it's been five years since uh I incorporated. Uh here's my you know my anniversary post lessons learned. Duolingo, perhaps you know, one of the most famous uh examples, daily streak count. And then uh, you know, users are bragging about 365-day streaks. I don't know if Snapchat was before them. I think they were. Um But like I think, yeah, I think that was an original Snapchat thing too, right? Um I like to look at like how social products, like what are social products doing? Um, because you can actually take a lot of that thinking, that psychology, and bring that into whatever it is you're doing. So your product, what is the artifact that you're gonna create? What is the viral artifact that you're gonna create? Um, what do your users want to brag about? So if we want to get started this week, what are we gonna what do we need? Step one, we're gonna have to identify the output or milestone your user would screenshot and share. Step two, let's make it beautiful, let's make it shareable, let's make it branded. Your logo actually should be subtle but present. I've seen some examples of artifacts that the logo is super big, and it's just no one wants to share someone else's logo. It should be it should feel uh it's about you. Step three is you're gonna want to add a share button that pre-fills the post with the artifact. Step four, every share is free impressions to your exact target audience, so that's a huge win. Um the user is doing your marketing because they're bragging about themselves. Uh people are gonna say, well, these are you know mostly consumer uh examples. Will this work in B2B? Um or you know, yeah, B2B. Yes, you know, at the end of the day, B2B is are people too. There's things that these people know people want to share. Maybe they're sharing it within Slack, maybe they're sharing it within Microsoft Teams, but just think about what is something that they are willing to share uh with your with uh the group that you ultimately want more of. So that is strategy number five. Strategy number six. So you know me, you know, I think audiences are really important. Um sometimes, you know, it's just hard to build an audience. Listen, a lot of times it's hard to build an audience, more more often than not. Uh it takes it takes sometimes years, it requires daily content. Uh there's no guarantee it works too, right? Sometimes you just it doesn't work. Um, and then it it's tough, right? Because you have zero subscribers on day one. So uh what if you just acquired a niche newsletter instead of building from zero? So this is something a lot of people don't think about, but you can go and like buy a 10,000 subscriber newsletter for five to twenty thousand dollars, you inherit that trust from day one, you plug in your product immediately, and then you can repeat that across niches. Um if you have a product that you believe in and you understand the LTV, this just might make sense for you. Uh so and a lot of these uh smaller newsletters, like they're not really monetizing that well. So they might be like stoked to get a ten thousand dollar offer from you. So if we want to just get this started this week, uh, I would go to deuce.com, uh, which is like a you know listing of selling newsletters, newsletter investor, or just search your niche newsletter on Twitter or Substack and just try to find these newsletters. That's step one. Step two is you probably want to look at the 5,000 to 50,000 subscriber in your target niche. You obviously don't want to buy something so big in the beginning. Uh, but I mean, listen, if you have the money and you really do understand the LTV, maybe you do. Um it's all about confidence, uh, how much confidence that you have. Uh step three is DM the owner. Have you ever thought about selling? Uh you'd be surprised how many people would sit uh would take the call. Um next step is you know, a lot of these, a lot of these folks, as I mentioned, are not really making that much money. Maybe they're making zero to five hundred dollars a month, and they would be open to a fair offer, uh, you know, five to ten to fifteen to twenty thousand dollars. And then all of a sudden you want to direct channel to your exact audience. And more importantly, it's not a social media account that could be taken away from you at at any point. This is a yeah, a direct line uh that you can go and put links to, and if you want to put your product in there and say, come buy my product, you know, it's gonna get seen, unlike places uh like social media that could suppress your reach over time. So I think this is something that not a lot of people think of. So I definitely want to throw it, throw it in there. Um the last strategy is uh thinking about uh AI building an AI content repurposing engine. So the way I think about it is this you can think about uh creating one you know hero or pillar piece of content. That could be a podcast episode, a YouTube video, a long blog post, an essay. Um and that one piece of content ends up becoming uh a lot more. You know, it could be 50, 75, you know, 20, depends how much you want to get out of it. So you can actually use AI to take that pillar content, turn it into five to ten you know tweets, three to five LinkedIn posts, two to three short form videos, one newsletter, and the short form videos would be using like a remotion uh skill uh on something like a claw code, um, you know, a newsletter edition, a blog post, five to ten quote graphics that you can use creative LLMs to go and create for you, email sequences. And a lot of this stuff can actually, you know, for the people into OpenClaw, they can go and automate with OpenClaw. You can also use um just claw dispatch and and I think uh you know, claud co work, um, perplexity computer, any of these to go and automate some some of this stuff for you. The big the big idea here is you get one thing that'll equal a bunch of different contents, and that gives you way more touch points across uh multiple platforms. Um so if you want to get that started this week, um you know, I I I would suggest just starting recording like one 30-minute piece of content. It could be a podcast, it could be a video, or just a voice memo. It's way harder to actually go and write a uh 3,000-word essay on something. And I just find if you just voice memo something or you do a podcast, uh, it could be like almost it could also just be an internal podcast or just a pot like literally a video of you that you don't end up putting out, and that like the actual value is in the you know, the repurpose.