Building a Venture Scale AI Developer Tool Company: Moderne CEO Jonathan Schneider (Part 6)

Sramana Mitra: I have a question that comes to my mind. Enterprises have very diverse applications and frameworks. Are there particular frameworks that you specialize in, do better in, or have more training in? Or you can basically walk into any enterprise, do this training right there, and basically build from there?

Jonathan Schneider: You’re hinting at something very important, which is that it’s less about the frameworks and more about the tech and language stack, which was very important. To build that representation, we had to integrate deeply into the compiler, which is very language specific.

Sramana Mitra: Very language-specific.

Jonathan Schneider: Then we started in the JVM and started broadening out into infrastructure as code and then walked back into COBOL and then JavaScript, Python, and C#. There’s a go-to-market challenge there at the beginning as we think about how we build outbound.

Sramana Mitra: I was going to ask you about that.

Jonathan Schneider: We don’t really care about the industry vertical of a company. We don’t care about the size. I care about the tech stack. The qualifying criteria in the early stage of the company is tech stack segmentation. A lot of data sources for outbound are built around industry vertical or size, not tech stack segmentation.

Sramana Mitra: So, how did you do that qualification? Did you buy databases? I know there’s a company up in actually the Seattle area in Vancouver, Washington that has that kind of data. Is that what you’re using?

Jonathan Schneider: We didn’t find a good source of data we could buy. This actually wound up being a question we asked back to Gradle. So remember the history here. Gradle also had a tech stack segmentation problem. They could only serve JVM customers or at least initially. That was a problem they faced as well. So it was a question of how have you done it, right? You’ve gotten the thirty or forty million, so how have you done that?

They had developed a heuristic around LinkedIn job description. How many JVM developers are there proportional to the size of employees or the total number of employees in the company? They had identified certain thresholds for 5% and 3%.

We just used a very similar heuristic, developed that same technique and tested it. I could go to an event where we pre-calculate that. I’d know this person’s going to attend and I can see they’re at 5%. If I talked to that person, I would find out how accurate that number is. It’s pretty spot on.

Sramana Mitra: Very good. So, you need a way to qualify and you’d found a different way to qualify. What about the company structure? Are you all now in Miami? Is the company fully in Miami or is it spread out virtual? What is going on?

Jonathan Schneider: It is spread out in engineering, although we deliberately had a concentrated time zone at the beginning. Early on, we didn’t particularly care about country, but it had to be Americas region- Canada, North America, or South America. We just tried to align the time zone so that there isn’t too much spread.

As we grew the engineering team and had some independent functions, like language engineering, the best talent for that is in Europe. It’s okay to have a team in Europe because they’re kind of working together on the same problem whereas the SaaS platform team is more concentrated in the Americas.

We’ve deliberately avoided both customer acquisition and hiring an APAC just to not put too much time pressure stress on the company.

Sramana Mitra: How big is your team?

Jonathan Schneider: We’re still just under forty. So there’s an echo of the Netflix mentality here. When I left Netflix, they had I think 500-600 engineers. At that time, Facebook had somewhere between 10K-15K engineers. These are very different size organizations, but the Netflix mentality is – we hire senior, we pay more. We don’t hire folks that are just newly out of college, Facebook will. Our company is built more on the Netflix model. Our average age of the company was over forty for a long time. They had families, were much later in their career, needed less supervision, and had more experience.

Sramana Mitra: More experienced entrepreneurs and developers. What is the split between developers versus go-to-market team?

Jonathan Schneider: It’s close to fifty-fifty now. A little bit more in engineering. Things have kind of evolved in tiers. So I mentioned we did single tenant SaaS at the beginning. We found that it has a relatively high activation energy. That’s a phrase that Matt Carbonara from Citi Ventures uses.

How much belief and pain does a champion have to go through to even bring your product in?

A very visionary buyer is just going to charge through walls for you. Somebody that’s a little less interested, may be willing to do the procurement element of it, but just doesn’t want to do cloud security review, right? It requires too much activation energy.

We eventually developed an on-prem air gap addition that would be easy to replatform onto single tenant SaaS that allows us to sell an initial term, which avoids cloud security review altogether. We wanted to lower the activation energy.

At the end of or mid-2024, we also started inserting technical account managers into customers from the outside. As the scale of open source technology started showing up more and more in consultancies, we decided to place them into accounts to do custom migrations for them with our platform, which further lowers the activation energy. Now, I don’t have to reallocate engineering resources internally to make use of this. I can accelerate timetables. That’s something we have to deliberately think about.

This segment is part 6 in the series : Building a Venture Scale AI Developer Tool Company: Moderne CEO Jonathan Schneider
1 2 3 4 5 6