Fundraising when you’re building in the dark, lessons from JFrog’s co-founder Ben Haim


so this is another quick post, before I’m going back into the rabbit hole.

So I listened to this Twist session with JFrog’s co-founder Ben Haim over the weekend and I thought there were some really valuable lessons in there about (1) building products before the market exists and (2) about the difficult fundraising journey that comes with this.

So what is JFrog? From a VC perspective it’s this late-stage dev tools startup, which has raised a $165m Series D led by IVP last October at a unicorn post-money valuation (exact figures aren’t public).

But what exactly is it that they do? This is when it gets interesting, because, like so many times when it comes to frontier tech and dev tools in particular, the average person is lost after the founder has uttered the first few words of the first sentence.

In the case of JFrog, which I would no longer categorize as frontier tech, there’s now an established narrative and terminology to describe their product and vision. But as the session with Ben and Jason shows, this has not always been the case.

In fact, the term “DevOps”, which is at the core of JFrog’s enterprise software product, didn’t even exist at the time the company was founded. As Ben puts it:

“When we created JFrog, there was no budget line for what we had. We couldn’t even explain what we replaced. It was replacing home grown solutions, or bad habit, or what you will need in the future to release faster”

So how can you pitch a “DevOps tool”, before there are multinational tech giants pushing out code seven times a day? Or in other words, how can you stay alive and build, when the market you’re addressing doesn’t even exist yet?

In fact, even today it’s hard to find an easy description of what JFrog does:

  • First, there’s always the “dumbed down” version, which could go something like this: “JFrog is like Github, but not for source code but for updates.” This is the low substance description, where you work by analogy and hope the audience knows about Github and source control is.
  • Then there’s the “nerd-out” version, which typically goes a bit like this: “JFrog is a binary or artifact repository manager, offering an end-to-end, high-availability software release platform for storing, securing, monitoring and distributing binaries for all technologies, including Docker, Go, Helm, Maven, npm, Nuget and PyPi.” This is the TMI version.

If I had to describe what JFrog is doing in more or less simple terms, then it would probably go something like this: JFrog is creating a way in which software, once it has been compiled into binaries, can be stored, monitored and updated. It’s much like a Github for production level software.

Well, so whatever this weird thing is that JFrog does, it becomes clear that its super difficult to describe. That’s even TODAY!

So imagine for a second that it’s 2008 and you’re starting out with this weird product, that only a hand full of open-source contributors get and you have to pitch to a room full of khakis, aka MBA VCs. You’re literally in the darkest woods of startup land.

The firm’s lack of a slick elevator pitch went so far that one investor told Ben:

“You either need an old elevator or a really tall building, because you can’t pitch that”

So how do you find the lightbulb and keep going in the dark. The JFrog story offers some pretty interesting guidance

1. Listen to signals other than financial signals.

So JFrog came out of an open-source JavaScript project, the “Artifactory”. The Artifactory is still JFrogs main product and it’s basically this binary/artifacts repository. Since they use the open core model, which I have to write about in more detail some other time, its core codebase is still open-source today. So what’s the point here? Well, while JFrog didn’t get a lot of investor interest in the early days, it could see massive traction in terms of code pulls and commits. This told them that there’s some kindle.

2. Prepare for a cold winter

As I mentioned, when JFrogs started out, the term “DevOps” didn’t even exist yet. The were no dedicated DevOps teams, as we have them today. Most software giants were still updating their software quarterly (remember the Microsoft update CD ROMs?), not pushing out production level code seven times a day like the Facebook or Snaps of this world are doing it these days.

When you’re building that far out, better prepare for a cold winter. As Ben describes it, it took him 4 years to raise the $3.5m Series A and 3 weeks to raise the $165m Series D last year. Notably, even when they were raising the Series A and they already had paying customers like Netflix, LinkedIn and Cisco, it was hard for them to convince VCs. The sentiment in the room was always that the market cap of dev tools was just not big enough, maybe $100m at best.

This meant that, apart form a small $1m grant in 2009, the guys had to bootstrap for almost 4 years. This is tough, not just mentally but also from a financial standpoint. In the case of JFrog, the three founders already had a small exit with AlphaCSP before and could scrape by. Also, they were based in Israel before the Series A, so they only started burning cash on overpriced real estate in 2013 when they moved to the Bay to open an office (or as they call it a frog “swamp”).

3. There’s light at the end of the tunnel

The bright side of starting to build before there’s a market, is that if you make it through the dark, the economic rewards can be much larger than if you build in a highly crowded space.

In the case of JFrog, the preference stack didn’t kick in before they had some decent traction. In the episode, Ben mentions that they raised the $3.5m Series A at a valuation of around $17.5m (indirectly, by saying that they gave away around 20% at the A). The $3.5m Series A would be considered a Seed today and the $17.5m valuation would be considered a top-quartile YC batch cap. So they did get diluted for sure, but probably much less than they would have, had they managed to raise earlier.

Also, they didn’t raise a massive round before it was already a no-brainer in terms of traction and the round could be raised purely on financial terms. It’s interesting how Ben describes the Series D in the session. He says that compared to the Series A, it was all about the “typical Wall Street financial metrics”. In contrast, at the Series A he would still let the tech guys do most of the talking. Selling Dev. The even more extreme example of this bootstrap for as long as you can is of course Grammarly, which raised a $100m Series A as their first outside capital, because the founders had managed to bootstrap forever, thanks to their past exit nest egg.

This is different from when you’re raising for an overcrowded vertical, like micro mobility aka scooter wars, where you have five startups come out of the gate with huge seed rounds. Before the founders have updated their LinkedIn profiles to ‘Founder of n+1 Scooter Company”, finished circling the TAM on their Keynote pitch decks, they are diluted more than a Theranos blood sample. Yes, this is the black hole of hyper competition at its finest. The narrative on your pitch deck may be ready on day one, but so are the wheels of capitalism.

Long story short, building in the dark is hard, but it’s also hard core lucrative if you make it to the other side of the dark woods.

So drive safe out there, don’t get diluted before you have to.



Note: I do voice recordings of these posts while walking on Sand Hill Road (not really SHR, but some busy street). Check it out if you don’t like to read.