The Power of Perseverance: Starburst's Journey from Open Source Vision to $3.35 Billion Valuation
This is the Three Data Point Thursday, making your business smarter with data & AI.
Let’s dive in!
Actionable Insights
Here’s the 3-minute summary of what businesses can learn from Starburst and its CEO, Justin Borgman:
Technology and science are key differentiators for data companies. The success of both companies Borgman founded is deeply routed in his continuous search to turn cutting-edge technology and science into businesses.
Perseverance is key to building companies based on deep new technology. Turning cutting-edge technology into businesses often relies on markets adapting and technology advancing further. That means long-term visions and perseverance to keep following them until markets & tech catch up to your vision.
Democratization of data is a thing. Democratization of data and getting data to everyone isn’t a sexy word at all, but it is what Justin has been doing all his working life. And it seems to work.
Open source is a player, whether you want it or not. Whether you’re starting a business or transforming yours using data. Open source is always there, an option to use, and a competitor to beat. And, as Starburst has shown, it is possible to maneuver through this jungle very well.
Starburst: A Beacon of Progress in Data Technology
At age 20, according to legend, Odysseus entered the war at Troy. He spent ten years fighting till the end of The Trojan War. Looking to return to his wife and kingdom, he embarks on his ship, only to be caught in a never-ending journey lasting another ten years.
Ryan Holiday tells this story in his fascinating book on how obstacles help us move forward to make one point: There’s persistence, and then there’s perseverance. Odysseus was on an entirely different level; what he needed was more than persistence; it was perseverance.
“If persistence is attempting to solve some difficult problem with dogged determination,.... [then] perseverance is something larger.” - Ryan Holiday in The Obstacle Is The Way
Perseverance single-handedly describes both the company Starburst and its CEO, Justin Borgman. There is something larger at work here; neither Justin nor the company he has built are only persisting against challenges - that most founders and companies do.
Instead, what they show is something on a different level: perseverance.
And that’s why it’ll make Starburst one of the biggest data companies of this century.
Just another DbtLabs?
What company do you think of when you hear “bootstrapped and data unicorn?”
The company DbtLabs pops right into my head. Founded by Tristan Handy as the data consultancy Fishtown Analytics. It went on to create a successful open source side project called dbt. Once that took off, they decided to rebrand to DbtLabs and go all in on dbt, turning DbtLabs into a unicorn data company valued at $4.2B.
It sounds like a fantastic story, but in my eyes, the success of DbtLabs overshadows a bunch of companies equally or more likely to succeed.
On the list of most valued data unicorns in 2023, another bootstrapper is hiding in the shadows: It’s the company Starburst:
A company valued at $3.35B
With a recent series D in 2022 for $250M
And three 3X YoY growth years in terms of revenue. That’s a 2019 to 2021 ARR growth of 27 X.
Like DbtLabs, Starburst is also based on an open-source project called Trino.
And just like DbtLabs, Starburst also started as a bootstrapped company.
Both companies share a similar start, impressive growth metrics, and unicorn valuations. Both are centered around open source technology. One can’t get away with thinking these two are indeed similar.
And yet, the more I research Starburst and its CEO Justin, the more it looks like the differences between DbtLabs and Starburst drove Starburst’s success to this point, not their similarities.
The differences between DbtLabs and Starburst drove Starburst’s success to this point, not their similarities.
What’s different about Starburst?
While it looks like Starburst and DbtLabs share a similar start, both bootstrapped and based on open-source tech, their differences are the engines of success for Starburst.
The bootstrapped Starburst: DbtLabs was a bootstrapped company because, well, it was a consultancy before, doing what consultants do. At the same time, Starburst had a much larger vision from the very beginning. Justin and his team bootstrapped the company by offering service contracts for the open-source tool Trino, which was still called Presto. However, they did that for a clear reason: to wait until the technology had unfolded enough for them to pursue their bigger vision.
While DbtLabs stumbled into a unicorn business model, Starburst was set up to become a unicorn from the beginning and was willing to wait.
Open source tech at Starburst: The best part about building businesses around open source is that it’s open! The hardest is… that it’s open! Interestingly, Starburst had to handle a severe conflict over the open source project prestoDB with none other than Facebook. They decided to separate themselves from the Facebook-backed project, a step requiring balls and the dedication to make it work. It turns out, as far as I know, they are the only open-source-based company managing a conflict with a big player gracefully without taking a hit.
Built on technology: While dbt looks like a fabulous technology, in terms of technological progress, it’s more a coffee mug than a rocket ship. That’s not a bad thing per se, but it is a difference that matters. Starburst CEO Justin Borgman has shown a history of building companies around profound technology breakthroughs, just as he did with Starburst and Trino. Trino is a substantial scientific advancement, as was the project his first company, Hadept, was built on. It literally came straight out of university. DbtLabs is lacking any such inclination and likely won’t develop one. That stuff is hard.
In my eyes, DbtLabs lacks all of those engines. But if I had to put them under one wrap sheet, it would be that of perseverance.
It’s precisely those engines we will look at in this article.
“I think the single most important attribute is strictly perseverance. [...] It never gets easier, just different kinds of hard.” - Borgman in Interview
The origins
“I was very lucky to have an amazing mother who, not only worked exceptionally hard, but is just a very determined, tough, driven human being. She taught me to never give up. And I really believe that start-ups are an endurance sport,”- Borgman in interview with Alejandro Cremades
Justin Borgman learned the value of perseverance from day one and the value of grand visions deeply rooted in technology on day 2, not realizing how intertwined those two are as leadership skills.
In 2009, while he was at Yale, a recently published paper caught his attention: “HadoopDB: an architectural hybrid of MapReduce and DBMS technologies for analytical workloads,” authored by a Yale-based research group.
The technology blew away Justin, so he approached the lead researcher, Abadi, about commercializing this research. Abadi declined and declined again. But Justin stayed persistent and, in 2010, finally convinced Abadi. The team started to work on plans for forming a company around this technology.
In late 2010, they secured licenses from Yale and raised a first seed round of money to form the company Hadapt.
Keeping up the vision
In 2014, Hadapt had moderate success, Borgman his first four years as CEO under his belt, and the company was sold to Teradata, where Borgman worked as VP and GM. He had learned to continue pushing and developing a deep technology further to turn it into a business.
However, as luck would have it, Teradata only enforced Borgman’s visionary inclinations. His job became to figure out the future of data warehousing analytics. It is there that he stumbled upon the next opportunity.
“along the way discovered some guys at Facebook who had created a new technology for analysing data at tremendous scale with hundreds of petabytes of data and thousands of users hitting the system at the same time” (Justin Borgman in an interview)
More technology and one big problem
Realizing the opportunity but without a clear direction to act on, Borgman paired up with a bunch of the creators of this new technology in 2017 and started to do something, anything, to start making progress on this opportunity.
They bootstrapped their newly formed company and sold support to existing users of the open-source project to get it off the ground.
“The first couple of years we grew that way organically,”
By now, Borgman had spent roughly ten years on one problem: Getting people access to data. If you think about it, that sounds like a tedious problem, almost as boring as labeling data. And yet, as you can read in one of my past articles, labeling data is what Scale AI does and what made it a $7B company. And it was the singleminded focus of its founder Alexandr Wang who saw this one problem where even today, no one takes it seriously.
Borgman is just like that; he strips away the assumptions everyone has in his head and gets to the core: Accessing data is super hard for 95% of the people inside a company. Accessing 90% of the data is super hard for everyone inside a company. Data access is a huge problem and has always been.
At first, with Hadoop, Borgman tried to solve the problem of access to large amounts of data; now, with Trino & starburst, he’s solving the problem of accessing data in many different places, but the problem remains the same.
As with any deep problem, it takes time for the technology to develop and for the science to progress. So in 2020, when Starburst took up speed, after ten years of trying to solve this one problem, it was finally the time for Starburst as a company.
Luckily for Borgman, he learned the value of perseverance from his mother early on, and it’s probably the one thing that enabled him to stay focused on one problem over ten years and three companies.
The two big changes that make Trino successful today
Two significant changes happened over the past decade, one slowly and one more disruptive. One is the development of the so-called MPP capability. MPP stands for Massively Parallel Processing (of computing operations) and usually refers to data store capabilities.
It means you can split up any computation and do parts of it in parallel, thereby speeding up the whole computation. Think of it: if you have the most complex computation in the world, the fastest way to complete it is to use the best computer - and there is only one of those.
However, you could also use computers that are only 1/10th as powerful as the best in the world - and the economies of scale will teach you that there are not 10x but 1,000s of those.
That’s the basic idea behind MPP systems: parallelizing into small common chunks is good. Over the last decade, MPP capabilities have been scientifically explored, then built into most common systems, and are now available to almost everyone and every system.
The second significant shift that happened more gradually is network speed. You must transfer data over the network at least once to work with lots of data. If you want to use MPP systems, you must transfer them often! Thanks to much faster networks, today we’re at a point where working with enormous amounts of data in a distributed fashion is very accessible to everyone - technically.
Jumping on more trends
Justin is aware of the data technology landscape and constantly leverages the whole ecosystem to realign Starburst.
I remember him contacting me in 2021 about my first article on the data mesh. It took me till then to realize Starburst was already pouring tons of marketing dollars into catching the data mesh trend. This is a trend that admittedly fits the tooling very well. Funnily enough, neither presto nor trio is the default choice for data access inside data mesh architecture (for no good reason).
But Starburst doesn’t stop there; Starburst also doubles down on the data democratization trend, putting content everywhere.
Starburst Galaxy
If you’re wondering what the company Starburst does, that’s a good question, and it’s one that a lot of companies looking to build businesses on top of open source struggle with. Starburst has found a nice balance, strengthening its position as a data access enabler.
While they still maintain trino, the open source framework, and probably cut off their support contracts for open source users long ago, their first product was Starburst Enterprise. Fittingly named, Starburst enterprise aims at enterprises that can deploy and manage a high-quality trino version themselves. They get an enhanced Trino version they can deploy into their infrastructure.
The company's new product, Starburst Galaxy, goes one step further and aims to provide more data access to a broader class than big fat enterprises able to manage things themselves. It is a nicely wrapped, much more user-friendly SaaS version of Starburst Enterprise (my description, not theirs).
Finally, hooking even deeper into the data mesh trend, Starburst developed a “data products” feature into their enterprise version. Borgman describes it as HelloFresh for data; you don’t need to source all the ingredients; it is curated data ready for consumption, shortening the time to insight.
Borgman and open source
“First of all, I’ll just say for me, as I was thinking about starting my second company, open source was an important criteria of the type of business that I wanted to build, because I think there are some really inherent advantages both for the company and customers. The first is, you get the benefit of contributions from a wide audience. I think that really enriches the technology and allows it to grow and evolve at a faster rate than perhaps a single vendor pushing it forward. And what I mean by that is, for example, in the early days the geospatial functions were created by the ride-sharing companies. We didn’t build those. I mean, maybe we would’ve gotten to it eventually. I don’t know, but they built that. So as a result, pretty much every single ride-sharing company in the world now uses this technology. The other benefit is it gives you very broad distribution. It is open source and therefore it is free”
“That was one of the lessons, painful lessons for me, actually, I learned in my first business, Hadapt. Although it ran on top of Hadoop, we were selling proprietary software and when Cloudera introduced Impala and that was free and open-source, included with the distribution. So, you know, that was really hard for us because we weren’t getting the same number of looks or evaluations.” (Interview at Madrona)
Borgman’s relation to open source is undoubtedly exciting and different from that of other COSS founders. Maybe it is because he himself was neither the creator of HadoopDB nor presto, but instead simply amazed by it.
In my own words, I would say that’s a good thing. Borgman learned the hard way that open source is there. It is a force, a significant force in business. So, if you want to do business, you need to learn how to work with open source as a business strategy. For most open source enthusiasts (and you might count me into that category on a personal level), that takes away the romance and beauty of some of the open source ideas.
But it still is reality.
Notes on resources
Ryan Holiday's book “The Obstacle is the Way” is one of the books I keep on rereading; I love the story of Grant running towards flying grenades.
Open source as a business tool is a real thing, and I’ve put a few articles out over time on a few separate topics on unpackingbos.com.
Here are some special goodies for my readers:
👉 The Data-Heavy Product Idea Checklist - Got a new product idea for a data-heavy product? Then, use this 26-point checklist to see whether it’s good!
The Practical Alternative Data Brief - The exec. Summary on (our perspective on) alternative data - filled with exercises to play around with.
The Alt Data Inspiration List - Dozens of ideas on alternative data sources to get you inspired!