From 100K Installs to Product-Market Fit: How We Found Our North Star
Last week, the Jozu team gathered to celebrate our recent funding round and dive deep into our product roadmap. It was one of those pivotal moments where you step back and realize how far you’ve come—and more importantly, where you need to go next.
The Search for Product-Market Fit in a Sea of Possibilities
Up to this point, our focus has been on building KitOps and creating the platform to manage ModelKits—the open source ML artifacts our team pioneered. But here’s the thing: having great technology and knowing exactly what to do with it are two very different challenges.
When you start with open source, especially in the rapidly evolving ML space, the possibilities feel endless. Every day brings new problems that need solving, and it’s tempting to try to address them all. The danger? You end up with a product that halfway solves a bunch of problems, but none well enough that anyone would actually pay for it.
This is the reality of searching for product-market fit. In a perfect world, there’s a single, obvious problem with a clear solution for a well-defined audience. In reality, PMF usually looks like stumbling around in a dark room trying to find a light switch—except you don’t know which wall it’s on.
But here’s what makes it even trickier: PMF is relative. Thousands of developer tools achieve what many would consider market fit for specific problems and specific groups of developers. The catch? The market is often too small or unwilling to pay, making the business unviable despite the technical success.
With KitOps ModelKits, we faced this exact crossroads. We could lean into the developer space where tools like Docker and Hugging Face are playing, or target the operations/infrastructure/platform space where the market is less defined but certainly being eyed by major vendors.
100K Installs: The Power of Focused Iteration
When we first started, we were 100% focused on KitOps. We knew the core idea was solid, but we had no idea who would use it or what specific features would matter most. We got just enough right to drive a modest number of downloads, then used each iteration to focus ourselves a bit more.
The result? In just over a year, we’ve broken 100,000 installs.
Jozu tells a different story. Our product is less than a quarter old, and until recently, we’ve been addressing two key areas: speed to production and security. We knew both were enterprise pain points, but we were still figuring out which one deserved our full attention.
The Tradeshow Test: When Reality Meets Strategy
In April, we took our speed messaging to KubeCon EU. If I’m being honest, the results were less than exciting. It’s hard to pinpoint exactly why some messages don’t land—maybe it was EU regulations, maybe we were too deep in the open source space, or maybe it was simply a message-audience misalignment.
But here’s why I love tradeshows: they offer quick feedback loops, exposure to different personas and industries, and the ability to rapidly test different messages. Which is exactly what we did next.
A few weeks later at Red Hat Summit, we had our second chance. This time, we leaned into our security messaging—and the response was night and day. By the end of the week, we had just enough data points to be dangerous: a clear target persona, specific industries we could easily reach, and a handful of really strong leads.
Finding the Courage to Commit
At some point, every startup faces a critical decision: you have to commit to going deep and establish yourself as the solution for the problem. Without this focus, you’ll struggle to command premium pricing and find it nearly impossible to scale your go-to-market activities.
For us, this means going all in on the security side of DevOps for ML.
It’s scary to narrow your focus when you know your technology could solve multiple problems. But that’s exactly what product-market fit requires—the courage to say no to good opportunities so you can say yes to the great one.
The 100,000 installs of KitOps proved we could build something people want. Now, with Jozu, we’re proving we can build something enterprises will pay for. And sometimes, that’s the difference between a successful open source project and a sustainable business.
Get Started
Ready to explore ModelKits for your own AI/ML projects? Check out our getting started guide or try Jozu Hub for a ModelKit-first registry experience. Join our community on Discord to connect with other developers building the future of MLOps.
The journey from open source project to product-market fit isn’t linear, but every stumble teaches you something valuable about your customers, your market, and yourself. The key is listening closely enough to hear what the market is telling you—even when it’s not what you expected to hear.