Founder To Founder: Why Can't I Just Vibe Code My Own CRM?

Greetings, friends on the internet. I'm Dharmesh, co-founder and CTO of HubSpot, which provides the number one agentic customer platform. And also, I think, the only agentic customer platform, which is basically a next-generation CRM for the age of AI and agents. And also, I love startups. I love startups and the founders who found them. And partly out of that love and some curiosity and some desire to learn, I did something weird yesterday. So I posted this LinkedIn post. I'm going to spare you the details and give you the TL;DR. I'm basically jumping back into early-stage founder mode and driving some founder-led growth. And so I'm going to try and Sell HubSpot, not the vision of HubSpot. HubSpot, like the actual product that we provide to customers. And as part of that, I have a web page, gohubspot.ai. You're welcome to visit. I encourage you to do so and sign up for HubSpot. But as part of this kind of effort to do some founder-led growth, I am committed to doing a series of founder-to-founder videos. And the idea behind these videos is if I were chatting with a founder that was considering HubSpot, I was making my pitch, what are the common questions they might have? So I'm going to try and address those questions in this series of videos. And for this particular video, I'm going to try and answer the question, why can't I just vibe code my own CRM? All right. So I feel like I'm reasonably well qualified to answer this question. Not because I happen to own vibecoding.com, which I do, a little domain flex there, but because I live in the matrix and I spend hours and hours a day literally vibecoding and agentic coding. My kind of current favorite is Claude Code. Opus 4.6 is awesome, but as are some of the other tools. But I'm not going to say, by the way, like one does not simply vibe code a CRM. That's not the answer I'm going to give you. I'm going to try to dig in and say, truly, can you vibe code your own CRM? And my answer to this, given the tools and technologies and the state of the art right now, the answer is yes. Like, for yourself, for your own needs, you could create your own CRM for just what you needed. But the better question, I think, is should you vibe code your own CRM? And to that, I think the answer is no. And I have a simple framework for assessing what makes for a good use of vibecoding and when is it not as good. So on the x-axis, we have narrow use cases, like, oh, it does this one thing, to kind of wider use cases, it does lots of things. So if it's a narrow use case, the narrower it is, like you're just building this One thing to do this one thing just for yourself. That's the extreme case. Vibe coding, awesome. As you move down the spectrum to kind of wider use case, it's going to do more things for more people. It starts to get worrisome and there might be cause for pause. On the y-axis, you have things that are like low maintenance. It's like, ah, it's out there. I don't need to fix a lot of bugs. I don't need to add a lot of features. I don't need to take late-night calls if the thing goes down. To high maintenance. It's like, I need to fix bugs all the time. People are asking for new features all the time. And I need to make sure it's up. That's high maintenance. Low maintenance, awesome. Vibe Coding is a great tool for that. But the more you move up towards high maintenance, Vibe Coding is sort of less and less viable, like building your own effectively that you have to support on your own. So you're solving for a narrow use case. It's just you solving for one discrete kind of use case that you have, vibe coding, awesome. As you move towards high maintenance, even though it's a narrow use case, maybe pause. If it's low maintenance, but it's doing a bunch of things, maybe pause. If it's high maintenance with lots of features, lots of use, this is where I think— You need to really kind of question your potential life choices here, and I would put CRM in that kind of high maintenance, wide use bucket. Biased, but not wrong. All right, so here are the kind of some of the common thoughts and questions founders I think often have around this kind of idea of vibe coding their own CRM. One is like, oh, CRMs are expensive. I can save. some real money by vibe coding my own. So why wouldn't I do that? So a couple of responses to this. One is CRMs are not necessarily expensive. So HubSpot has a CRM that starts at free, not a free trial, actually free. And even the paid version, the HubSpot Starter Edition, is $15 a month per user. No long-term contract. You can cancel anytime, right? So that's if it's just you and your founder, Uh, co-founder, that's just $30 a month. I wouldn't call that, uh, real money. Uh, the other argument I would make is that you're not really saving money because your time, uh, is not worth zero, and the opportunity cost for your time is definitely not zero. So, uh, the math doesn't math. You're not really saving yourself real money. That's my kind of counterargument to that. Another thought you might have is if I vibe code it, I can get exactly what I want And nothing that I don't. Now this is actually a pretty good argument, right? It's because you know HubSpot has 280,000 customers. We've been at this for a while, 20 almost 20 years. So there are likely a fair number of features in HubSpot that you might not need today, not right now at your current stage. That's true. So for instance, you might not need AI-powered automated sales outreach and follow-up. You might not need. a customer support system. You might not need integrations to other business applications that you're using. But I'm going to argue that even though you might not need all of that today, as you scale and grow, you're going to need more and more of it, especially the integrations. As you grow, you're going to need your CRM to kind of integrate with other things that you're using within your business. All right. Third argument, you might say, hey, we're an AI-forward, AI-native, AI-first company, and we really know how to do vibe coding really, really well. Awesome for you that you know how to do that. But that doesn't necessarily mean you need to be directing that talent at vibe coding your own CRM. Now, speaking of AI-forward companies, you may have heard of a company called Lovable. Lovable provides one of the most popular and one of the most loved vibe coding tools on the planet. Anton, one of my favorite founders of Lovable, had a tweet recently where they talk about, or he talks about, the kind of most common use cases for Lovable and for vibe coding. One is rapid prototyping, so you don't have to kind of wait for engineering or whatever. It's awesome for that. Building internal tools. This goes to the kind of narrow use case thing that I was talking about earlier. And in terms of software, the last one is like, you know, replacing kind of vertical SaaS with your kind of custom-built one. If the vertical SaaS is not quite what you need it to be and you're spending oodles of money on it, possibly a good use case for Vibe Coding. Now, as you might imagine, Anton, super clueful. Anton's team, super clueful. If anyone knows about Vibe Coding and the power of it, anyone has the talent and skill to vibe code their own CRM, Anton and Lovable would be one of those. And as it turns out, they did not vibe code their own CRM. Lovable is a HubSpot customer. And so I think Lovable has done what I think all great companies do, is that they spend all available calories on creating customer value. And this is my primary argument, not that you couldn't vibe code it. Not that you don't have the talent; it's that all available calories that you have as a founder, as a company, as a team should be spent on creating customer value. And I'm going to argue, and I'm right, is that your customers don't care whether you vibe coded your own CRM or whether you're using HubSpot or some other CRM. And so you should be spending all available calories on the things that create customer value. Speaking of customers, if you're not a customer of HubSpot yet. I encourage you to go to gohubspot.ai where you will find this video and similar arguments and links to be able to sign up for either the free edition or the HubSpot Starter Edition. Thank you for your time. I hope you found this useful. The next topic we'll be tackling will be, well, why shouldn't I just use one of these AI-native CRMs that are out there right now and be one of the cool kids? I'll tackle that one next. Thanks for your time. See you. See you later.