Ali Ghodsi is the cofounder and CEO of Databricks which accelerates innovation by unifying analytics across data science, data engineering, and the business. The company has raised $500M at a $2.75B valuation from NEA, Andreessen Horowitz, Battery Ventures, Data Collective, Microsoft, Coatue Management, Green Bay Ventures, and SineWave Ventures to name a few.
In this episode you will learn:
- The four ingredients of an effective CEO
- The keys to a winning investor pitch
- How hiring is different in a hyper-growth startup
- Creating culture principles that give your startup a competitive advantage
- The one most important thing Ali says he wishes he spent more energy on doing earlier
About Ali Ghodsi:
Ali Ghodsi, CEO and co-founder of wildly successful data analytics company Databricks and Adjunct Assistant Professor at UC Berkeley.
Ali was appointed CEO of Databricks in 2016 after holding the VP of engineering and product management title since co-founding the company in 2013.
Databricks has been on a tear of rapid growth/success, highlighted by a recent $250M Series E funding round which brought total funding to $500M and the company’s valuation to $2.75B – bringing an IPO as a likely scenario in the near future.
It was a university computer science class that emboldened Ali and a group of six UC Berkeley students to start what is now known as Databricks back in 2013.
Lessons learned from the group’s UC Berkeley courses have served as Databricks’ North Star for years and have led to some industry-altering innovations.
The idea began with the creation of Apache Spark open source project – a unified analytics engine for big data and machine learning that’s deployed at massive scale by internet powerhouses such as Netflix, Yahoo and eBay. The award-winning Apache technology is also the driving force/power behind Databricks’ Unified Analytics Platform.
Ali and the other co-founder are considered the true founders of AI, and are celebs of California for their groundbreaking work in AI (case-in-point, people come into the office each week to ask for autographs from all around the world).
They guys got their first breaks at Google and Facebook, and while building internal systems for them, realized these modern-day enterprises have enormous amounts of data that need to be crunched to be able to truly take advantage of AI. By starting Databricks, their goal was to bring AI to the masses, and this need is being met and realized.
Over 2,000 organizations globally, such as Nielsen, Hotels.com, Overstock, Bechtel, Shell and HP, are now leveraging Databricks to unify data science and data engineering teams across the end-to-end data and machine learning lifecycle.
This all started from the vision that Ali had back in the late 2000’s of creating a company that would be the clear winner in the big data platform race, knowing that data would soon become more valuable than oil.
Connect with Ali Ghodsi:
* * *
FULL TRANSCRIPTION OF THE INTERVIEW:
Alejandro: Alrighty. Hello everyone and welcome to the DealMakers show. Today we’re going to be learning a lot about data science and data engineering. Then also about how to scale businesses and how to come as a foreigner into the U.S. and also make it happen. Without further ado, I’d like to welcome our guest today to the show. Ali Ghodsi, welcome to the show.
Ali Ghodsi: Thank you so much.
Alejandro: So originally born in Iran during the Revolution. Obviously, not the best time to be around there, but what happened after this? Where did you go?
Ali Ghodsi: Yeah, actually in 1984, my family had to flee. We had about 24 hours to get out of the country. So they scrambled, and I think the relationship with the United States was not great at that time. There was a hostage crisis, so my parents got to Sweden. That was the destination they picked, far off in the north. Very cold.
Alejandro: Why Sweden, Ali?
Ali Ghodsi: I think when you have 24 hours, whoever will take you as a refugee, you’ll go there.
Alejandro: Got it.
Ali Ghodsi: So we showed up with leather jackets and many layers of clothes because it’s really cold there.
Alejandro: Wow. And you were five years old at this point?
Ali Ghodsi: Yeah, right.
Alejandro: Tell us what happened. You arrived in Sweden, and how was life there growing up?
Ali Ghodsi: Since my family didn’t have a lot of funds, we ended up moving to suburbs, which these are not the nicer parts of Sweden. I got a computer, Commodore 64. It was kind of broken. You couldn’t actually use it. The tape recorder was broken, so you couldn’t play any games on it. The only thing you could really do with it was program. I had a bunch of manuals, and you could learn programming, and you could start programming with it, so that’s what I did. Since none of the games were working, I started writing smaller programs. Then I eventually started writing games for it. Around the age of eight, that was pretty much what I was doing all day long was coding.
Alejandro: Why do you think you had this love for coding and perhaps for resolving problems?
Ali Ghodsi: I think it’s like there was not much else to do. I didn’t really have friends, so just sitting at home and figuring out what this machine could do. I remember my cousin came. He was much older, and he said, “These things can actually – you can tell them. You can build your own programs. You can build your own games yourself.” He pulled out the manual and showed me that you could actually start writing some basic. I was just blown away. So I started writing code immediately. That was it. I was stuck just coding from that age on up until just a few years ago when I transitioned to CEO. I’ve been programming probably since the age of eight continuously. It’s probably the one thing I’ve been doing every day almost.
Alejandro: Very cool, and we’ll talk about that transition because I think that transition is so interesting when you go from engineering to the business side. But let’s not get ahead of ourselves here. You went to the university there in Sweden in computer engineering. Then also did your MBA. Why did you do logistics and strategic marketing? What was the thought process behind that decision?
Ali Ghodsi: Well, it’s actually kind of silly. I had a 70% absence in high school because I was programming all night. I was actually on a U.S. time zone working with people in the U.S. So I ended up not getting into a very good school because of the 70% absence, so I had to move far away, even farther north to the northern parts of Sweden. I showed up there, and they gave me a roommate. My roommate said he was doing a business degree. I couldn’t really understand what it was. It was like he was doing logistics and stuff like that. He said, “What are you doing.” I said, “Well, I program. I do this.” He didn’t understand what I’m doing. Then I kept asking him, “What does that mean?” At some point, he just said, “Look. I’ll be your boss when we’re done. Okay? I’m like the manager, and I’ll be your boss.” I thought to myself, “That’s **. So, I’m going to see if I can do a dual degree. I’m going to do business as well. I’m not going to have that guy be my boss.”
Ali Ghodsi: That was really honestly what was behind it. I had no idea what it was, so I decided I’m going to do business. I’m not going to let some other guy with a business degree pass over me.
Alejandro: Okay. Good choice. Then after this, you went and did you Ph.D. I think that this was the transition for you to start to really get your head up and look towards the U.S. Can you tell us about this?
Ali Ghodsi: Yes. I started doing my Ph.D. on completely decentralized systems, which are precursors to what today is bitcoin. I started collaborating with the really, really smart guy in the U.S. or had some papers and I got to interact with him. His name was Ion Stoica who was at MIT. This was year 2002 or 2003. My research overlapped with him, and we kind of semi-kept in touch with these groups. I finished my Ph.D. in Sweden, and I took a professorship job there. Actually, I had a short stint here in the U.S. in Palo Alto for a summer, and I really didn’t like it here. I said, “Look. I’m definitely not moving to the U.S. ever. It’s all suburbia and big cars on highways.” But 2009, I got the opportunity to actually collaborate with that same guy from year 2000. He was at UC Berkeley. He was now a professor here, and he said, “You know, if you want to collaborate on these projects, maybe you can come here.” We talked over the phone, and I said, “Look. Maybe I should just go visit them for one year. I definitely don’t like the U.S. I’m not moving there, but I’ll do one year at UC Berkeley, and then I’ll go back to Sweden.” That’s how the transition started. I spent a year here working with Ion, and I was totally blown away.
Alejandro: What kind of stuff were you guys doing together there?
Ali Ghodsi: It was hard to know at that time, but it was actually an extremely important moment in history. Basically, what happened – and it was not super clear at the time, but what had happened is that they stopped figuring out how to make computers faster. So, mid-2000 is actually a big change in computing history. They don’t know how to make computers any faster. When I grew up, the machines would double in speed every 18 months. It was like 16 megahertz, 33 megahertz, and so on and so forth. Faster and faster, doubling – you get a new computer every year. But this changed around mid-2000, which then meant every computer was about 3 gigahertz. So what really happened is that they had to start moving a lot of the computations and a lot of the work into the data centers. Data centers started to become the new computers. We were in the middle of this. We were getting funding from Silicon Valley tech companies that were making this transition, and we did a lot of research on how do we actually build this new software for this new computer? And the new computer is the data center. We built a lot of software around this. It was exciting times. Many of the projects we started now are used by almost every enterprise on the planet. So the timing was sort of perfect. At the end of that year, I said, “I’ll stay another year.” I called back to family and friends. My girlfriend, at the time, said, “Sure. It seems you’re having a blast. Stay another year.” Then I stayed another year, and then another year, and then another year. So, that’s how it happened. They were really sort of magical years. Our timing was sort of perfect.
Alejandro: Got it. Tell us about one of the projects. I believe it was a very popular open source project.
Ali Ghodsi: Yeah. This was Apache Spark. Really, what it started with is, we were actually working on another project at that time. There were these guys that were doing machine learning next to us. They wanted to participate in this context called the Netflix competition. What Netflix had done is, they had given out movies or something like that. Here’s what people think – how they rated all these movies. Come up with a machine learning algorithm that can predict what movies we should recommend to our viewers. They were going to use the slide, and whoever gets the best accuracy, whoever can recommend the movies that people click on and watch those movies will, I think, win half a million dollars. The guy in the lab was saying, “I’m trying to do this, but the existing technology is really hard to use, and it’s taking forever. I’m using this thing called Hadoop, which was a big thing back then. “Help me.” Some of the technicians in the lab started working with him and started on top of the project we had built, which eventually became known as Apache Spark, to help him do these predictions. Interestingly enough, we submitted the results, and we were tied for first place. Exactly the same accuracy; exactly the same everything. But they only would give half a million dollars to one of them. They picked whoever submitted the results first. Actually, this team ended up not getting anything for it, but that was the beginning of Apache Spark. What the project did was, it enabled you to take lots and lots of data like movie recommendations and do machine learning on it so that you can actually start predicting things. For instance, predicting what movies people haven’t seen, but they would like to see.
Alejandro: Got it. I’ve heard as well that you guys are considered the true founders of AI. Real celebrities for the work that you guys have done. So would you say that this was the early beginnings to establish the pedigree that you guys have gotten from your work?
Ali Ghodsi: Yeah. At the time, of course, nobody knew what this thing is or any of the other stuff that we had. So we actually struggled in 2009, ’10, ’11, ’12 to get people to adopt this technology. We were going around Silicon Valley giving them this technology. We were giving it to these existing companies and saying, “Please take this technology and run with it. Go commercialize it. Make money on it. You can take credit for it. You can say you created it. We don’t want anything. We just want the impact. We’re Berkeley hippies.” Their response was, “Nay, this is not that interesting. This is academic. Some student wrote it. The student will quit and go on and do something else in their life, and then we’re left with the source code. So, we’re not interested.” It was an uphill battle. None of these companies ended up really adopting it. So, tough years.
Ali Ghodsi: We were just trying very hard. We would send in students as interns, kind of like Trojan horses in the summer. As Trojan horses, they would try to get them to adopt our software, but it just wasn’t working.
Alejandro: Obviously, now, everyone is talking about Artificial Intelligence. Everyone. Every single startup that is trying to pitch their idea, there is some form of Artificial Intelligence. So, it’s interesting how far along we’ve gotten now with this.
Ali Ghodsi: Yeah. You know, AI is an old term. The research on AI started in the 1950s and ’60s. It was not a hot term in those days, say 2009 or ’10. Maybe it was called Machine Learning. Yeah, at that time there was not that much interest in this stuff. And it wasn’t until 2013 that Ben Horowitz, who actually grew up in Berkeley. So, from Andreessen Horowitz, the famous VC showed up. He said, “I’ve heard about you guys, and I think this thing you have is pretty amazing. I think you can create a 100-billion-dollar company with this. But you guys have to do it yourselves. You have to start a company. No one else will do it for you. You have to build this company yourself.”
Alejandro: How did he find you guys? How did he find you earlier?
Ali Ghodsi: He knew some of the professors at Berkeley, so he knew another professor whose name was Scott Shenker. He had invested in a company there that was called Nicira, which had just got bought for 1.26 billion dollars by VMware. Scott was involved also in creating Databricks. Somehow, they started talking. He heard from Scott that “We’re working on this thing called Spark, and it does machine learning and much better than Hadoop.” He had heard through the grapevine, and he came to us. We were actually not that interested in taking his money initially. It didn’t look like we were going to take his money.
Alejandro: Then he started talking about how this could become a 100 billion and formalizing more into a corporate-type of structure. Walk us through the process until you guys finally said, “Okay, let’s do this thing. Let’s take the money, and let’s build something big so that we can really take AI to the masses.”
Ali Ghodsi: We were engineers, so we said, “Who is this guy? He’s going to come in and spend all this money. He’s not building this stuff with us. We just need $200,000. If someone gives us $200,000, we can take small salaries, and we’ll work on it for a year. We’ll start a company. Then later we can raise more money.” But just in case, we went around the founders. There were six of us and said, “Is there some price if he comes in at that price, where we would just go with it, and despite odds, just let him invest?” Remember, we had nothing. The first guy said, “Yeah. If he values our whole company at say 20 million, then maybe we’ll maybe let him invest.” Then we went to the next co-founder, and he said, “No, no, no. It has to be more valuable – maybe 25 million. The numbers kept going up. I think it when up until someone said, “It should be a 35-million-dollar company. Then maybe we will let him invest.” Ben came in one day, and he said, “Hey, guys. I’m not going to haggle with you guys. I’m valuing your company at 50 million. I’m going to give you this much money, and take it or leave it.” We were like, “We’re taking it!” So we jumped in and took it immediately. That’s how we – just from the get-go outbid all of us, and we took the money. I remember the day when the money actually came into the bank account. All the founders went to that laptop, and we saw the bank account. It said 14 and like six zeros after. We were like, “Wow! That’s so much money.” I was making $59,000 at that time at Berkeley.
Alejandro: Yeah. Wow!
Ali Ghodsi: We were like, “If we put this in a bank, we can get interest on this. How much is the interest?” Someone pulled out the calculator and computed how much interest. “Yeah, we should just put it in a bank account.” That was how it started.
Alejandro: How many of you guys really jumped into this?
Ali Ghodsi: Interestingly, I was the only one that was one-foot in. I was like probably the least interested of the six. I wanted to continue – we had such an amazing streak with the research. We had so much impact with this research in academia, so I kind of wanted to continue doing that. So I said, “Hey, I don’t know if I want to do the startup,” but then it turned out that the other co-founders were taking every member in this research lab, and taking it to the startup. I realized that it’s going to be completely empty at UC Berkeley, so I might as well just join them. So I joined part-time.
Alejandro: Got it. Walk us through what happened then. You take the money. Everyone is joining, and you start recruiting people. Then the company is born. What happens next?
Ali Ghodsi: We’re in downtown Berkeley, and we started coding, and we started bringing in these super-smart people. Unfortunately, the first engineer that we hired is one of the smartest students we ever met. His name is Aaron. This guy is one of the most brilliant guys we ever met. He’s the first guy we hired. We’ve never interviewed people before, so we asked him some questions, and he just sort of handled them. Then from now on, the next five or six people we interview who are actually also amazing, we immediately rejected them because we said, “Ah, not that good compared to the first guy.” We had a super-high bar initially. Actually, we had to adjust it, and we started hiring these engineers, and we started coding away, but we were in an uphill battle. These were tough times because people didn’t believe in Spark. Now that we had a company around it, actually the Silicon Valley tech companies that were selling these other competing technologies, they started spreading quite a bit of FUD, fear, uncertainty, and doubts about Spark. They would say, “Ah, you know, this thing. It’s not fantastic. Maybe it’s good for one or two things or one thing, but it cannot do all these other things.” So there was a lot of skepticism toward our software. In particular, one thing they kept saying was, “Spark is great if you have massive amounts of memory. But what if you have so much data it doesn’t fit in the memory? Then Spark will not work, so don’t use it.” It was very frustrating to us because actually, our technology would work both in memory and on disc, so we could actually handle all these big data sets that were on discs, but we kept hearing it from everybody. They would go to these conferences, and they would say, “Yeah, we heard from this other vendor that your technology doesn’t really work if it doesn’t fit in memory. We have a lot of data. So, sorry. We can’t use your technology. This was for several years. We kept fighting this and fighting this and fighting this. It was just going nowhere. These conferences have thousands of people go into them. They’re spending a lot of marketing, budgets – it seemed impossible to get our message out and tell them the truth that, “Of course this works. This works on everything.”
Alejandro: Got it. Did you guys experience any challenges as well with being six co-founders. There are a lot of egos there to manage.
Ali Ghodsi: Yeah. I’ll get to that. I mean, it’s a real strength today. It’s one of our superpowers.
Ali Ghodsi: We used to always joke that the PayPal mafia was six people too. So they were six. We’re six. But it’s also a lot of opinions, especially if you come from academia, and you have six people, six researchers, there are a lot of debates about everything. Yeah, it slowed us down, and it’s created a lot of, should we go left; should we go right? But really, we were all just focused, “How can we make this Spark thing take off? The first two or three years, it wasn’t taking off; no one was adopting it. So we had all these uphill battles. Then something amazing happened in 2015.
Alejandro: So, what happened?
Ali Ghodsi: In 2015, there was this marketing thing that we decided to do on the side. We were fed up. Everybody kept saying if your data doesn’t fit in memory, their technology doesn’t work. So we decided to participate in a geeky contest. Not the Netflix contest. This time, it was what’s called the Sorting Challenge. This is a contest where you take a lot of random data, and you have to sort it. You have to sort all the numbers. In particular, we wanted to sort one petabyte of data. That’s a lot. Right? We didn’t have one petabyte of memory. We had much less memory than one petabyte, and still, we wanted to sort this petabyte of data, so we worked on it day and night. With one of my co-founders, Reynold, we were able to get it down and beat the world record and do the fastest anyone has ever done sorting a petabyte of data. We submitted the results. This got out, and then the media started talking about this. “The Berkeley guys just won the Sorting Contest, and they sorted one petabyte of data faster than anyone’s done before, and they didn’t even have a petabyte of memory. This made it very, very clear that, of course this works, even if your data doesn’t fit in memory. The pendulum swung almost overnight with this event. Suddenly, this becomes the most popular software. Gartner puts it at the top of the Hype Cycle. Every one of these companies that had been rejecting it put on their web page, “We love Spark.” In fact, they started saying, “We are the Spark company. Not Databricks.” Before you know it, overnight, everything has changed. It coincided with us winning the Sorting Contest.
Alejandro: Got it. Then I understand as well that right before this around 2013, that’s when you become the CEO. I understand as well that there were perhaps some conversations at a board level. So walk us through this change as well.
Ali Ghodsi: So this is 2015. It’s awesome. The technology is now being adopted by the world, but it’s actually open source technology. The bad news is our revenue at that time is only 1 million dollars. So here we are. At that point, we’ve raised quite a bit of money. The company valuation is probably somewhere – I think that year after 2015, we got a 500-million-dollar valuation. The company is valued highly, but we only have 1 million dollars of revenue. The board is getting really anxious with us. They said, “You have less revenue than the local restaurants. And great, you built this amazing technology that you’re giving away for free, and everybody else is making way more money on your technology. You’re just not business guys. You can’t figure this out.” At that time, my CEO co-founder, Ion Stoica, was a professor at Berkeley, so he had to go back to Berkeley because they only give you absence of leave for a period of time.
Ali Ghodsi: They started also, now, doing a CEO search. Most of us threw in the towel, thinking, “We had a great experience. How many times in life do you get to write software that the whole world adopts? We impacted the world and changed it with the software. We’re just not business guys. We don’t know how to make money. So let’s just go back to academia and come up with the next cool thing.” That was my thinking, as well. They talked to a bunch of CEOs. But then at the end of it, Ben Horowitz is a big fan of founder CEOs. Towards the end, he was like, “Hey, maybe this Ali guy should be the CEO.” We scratched our heads, and even I was like, “Is that a smart idea switching one Berkeley professor for another professor as a CEO?” But that’s what he did. In January 2016, I took over as Ion was going back to Berkeley.
Alejandro: Why you out of the six?
Ali Ghodsi: Probably, I was the oldest after Ion.
Alejandro: So you think that Ben Horowitz’s suggestion was just based on age?
Ali Ghodsi: Yeah. The other guys were in their 20s. They looked a little young and didn’t have enough grey hair. So, let’s take the older-looking dude.
Alejandro: Got it. Pretty cool. Then, after all of this and all of the adoption, you take the reins as the CEO of the business. Then what do you start implementing or what do you start doing to really understand or figure out how you make money?
Ali Ghodsi: Keep in mind that at that time, my psychology was, I have nothing to lose. And a little bit of disappointment because we hadn’t made much revenue. I just wanted to go back to the university and the research just like my co-founder, Ion. When I took over in January, “What do I have to lose? I can make some pretty big changes. Right? Three changes were done. One was we really did this pivot into enterprise sales. We started focusing on large enterprises. Before that, we as a team tried to make Databricks work in almost a self-serve way. People come in, swipe their credit cards, and they just use this technology, and we don’t really have massive salespeople that go there and try to sell them. Now, we pivoted and went all-in on the sales. At that time, we also had hired an enterprise sales leader, Ron, started at the same time. He started building up, and we started hiring these very expensive sales guys. The thinking was pretty simple. The thinking was, who in the world will pay you 10 million dollars if you improve something 1% for them? It’s the really big corporations. The big corporations have a lot of data, and if you say, “Improve by reducing 1% fraud using AI, they’ll pay you 20 million dollars. So let’s get those guys as customers. It turns out that those really big enterprises have bought software the same way for the last 30 years. They buy through relationships. So the enterprise sales teams are really, really important. That was change #1. We started investing a lot in these guys. I remember the first guys we hired. Each of these sales guys makes $350,000. That’s a lot of money. Right?
Ali Ghodsi: Especially for me, I was making $60,000 at Berkeley two years earlier. So did we really want to pay these guys? And they don’t have any PhDs, but we decided to take a bet on it, and it started paying off. That’s one. The second was I was like, “Okay. If you want me to go on the front lines and fight this war, and I’ve never done it before, then I want all my lieutenants to be super experienced. I started hiring an exec team that was over-indexed on experience. People had done it before because there were a lot of us. We were six co-founders with PhDs. We’re smart, but we had never done it before. So, let’s bring in the pros. I did 12 executive searches and hired, over time, these 12 executives. That was the second change we did. We really up-leveled the team to people who really know marketing, people who really know sales, people who know customer success, people who know finance. Before that, we were trying to innovate as researchers and figure out each of these departments ourselves. We were trying to almost reinvent the wheel, and maybe even come up as researchers with a better way of doing marketing than anyone has done before. But at this point, I just thought, let’s focus our innovation on the technology. Then on this other stuff, let’s just get the pros. That was the second big change that we did. The third was, focus on the enterprise features. The open source technology is great, but let’s start building proprietary software that is really focused on the enterprise features that the large enterprises need so that we have something really valuable we can sell. Those were the three changes that really helped the company turn it around.
Alejandro: At what point are you able to be somehow relieved because as the founder and CEO, there are always fires that you need to put out, but at what point do you realize, “I think we’re into something here, and I think we have a meaningful business that is actually producing revenue and addressing the concerns that the board has”?
Ali Ghodsi: Yeah, it took a while. The first year, I think we had said we’re going to do probably 10 million dollars this year of revenue. I was new at sales. I had never done this before. My Head of Sales said, “I don’t want to lower the number, but I have no idea if we can go up to 10 million or not. I don’t know what this is based on. I don’t know how someone came up with this number. If they said 10, let’s do 10. Let’s figure it out.” We had no idea. We didn’t do any modeling or anything like that. We just sprinted really fast that year, and at the end of that year, I think we came in at 13 or 14. That was great. So we survived one year – didn’t get fired. I don’t think any of us thought, “This is awesome. Now we have an amazing company that has a great business model. Okay, we survived one year. Let’s do it again. Let’s see if we can actually beat the number this year.” The second year, 2017, we promised the board, I think 24. That seemed like a really big number. We started sprinting really, really hard. We started getting our first million-dollar deal that year. That was a sign that maybe it was doable because, before that, people would just pay us $12,000 maximum. Sometimes $18,000. Then we had a few $30,000 deals. But $30,000 seemed to be the magical limit, which no one would pay us above. I almost thought it was some kind of rule of law. But now in 2017, we were getting a million-dollar deal here, and things started to slowly take off. I think at the end of that year, we did 30 million, and now, it’s 24. Our recurring revenue at the end of that year was 40 million dollars. Now we had a pretty sizable business and started to realize that maybe there is a product/market fit here, and maybe we can actually build something. I think in 2018, was the year where I felt like maybe we’re out of the woods, and maybe we’re onto something here. Maybe we can really scale and push this.
Alejandro: Got it. I believe that last year, you guys did about 100 million, and this year on track to double. How did you manage to do this?
Ali Ghodsi: 2018, we started investing heavily then because then we realized this had the shock to really be big, so we started investing big time. We raised quite a bit of capital and invested that capital in partnerships. We have a big partnership team. We have probably 70 people who just focus on just partnerships. That was a big investment – started investing in more marketing, making sure that we are present all over the world. We started investing more in sales in Europe, so we started hiring a big sales team in Europe. We probably have about 100 employees now in Europe. Many of these investments stats set us up for large scale in the future. And it started happening then.
Alejandro: Got it.
Ali Ghodsi: We also started building a lot of enterprise features that the enterprises need. Those are difficult and costly and frankly quite boring features around security. We started investing in doing those, and these things start paying off. It takes a year or two, but they start paying.
Alejandro: How many employees do you guys have now?
Ali Ghodsi: We’re close to 900.
Alejandro: Wow. So I guess for you, just thinking about the incredible transformation that you’ve also experienced as a founder. You started on the engineering side. You were an engineer all your life, and then you have to go into a business guy. What was that transition for you like because I’m sure that there are a lot of people right now listening to us and probably, they’re about to make another similar transition, maybe going from an engineer to putting the business hat on? What was that transformation journey for you like?
Ali Ghodsi: I loved it because, for me, I got into computers as a kid because I liked the challenge. I think it was probably the most challenging thing I’ve done in my life is the transition, especially to CEO. I loved that challenge. You learn a lot. It’s a steep learning curve. I loved that. But it’s quite different. Basically, the skill sets needed by a CEO are extremely different. They’re almost the opposite of what you need as a researcher or as a research leader. I had to learn all of those, and talk to people, get coaches.
Alejandro: Got it. Ali, to be an effective CEO, what would you say are the key ingredients?
Ali Ghodsi: I think the CEO job can be summarized quite simply. One, you need a vision? You need a vision that’s compelling and can get a lot of people behind it, and it is sort of audacious and really forward-looking. A lot of people don’t actually come up with that for their companies. They run their startup, but they haven’t actually articulated what that vision is. So you need to articulate that, and then you need to repeatedly repeat it to the company. Repeat it to all hands every week. Repeat it in all kinds of context, to new hires. Just do this over and over again so that everybody’s aligned on what is the vision? What are we trying to do together? And it needs to be something inspiring. It needs to be something that’s not around how much money we’re going to make; not around how successful we’re going to be. It needs to be something that’s really aspirational. That’s one. Two, you have to figure out a strategy because great, anyone can say, “Look. We’re going to solve world peace.” Great. How? What’s the strategy? What are the things we’re going to do in the next two years? So come up with a strategy, and the strategy shouldn’t be 20 things that we’re going to do. It needs to be three, four, two, maybe max, five things. Then keep repeating those for the company, and make sure that you have a way in which you can operationalize that for the company. So that’s the second is the strategy, and also, repeating that strategy. Third, probably the most important, you have to hire an executive team that’s really strong that can execute on this. That’s probably the hardest part: finding those people in the team that report to you that can really execute. That’s probably the most difficult part, and this is where I see a lot of people failing. Then finally, four, you have to make sure that team is aligned, working together, and executing on that strategy and toward that vision.
Alejandro: Got it. My follow-up question on this, Ali is, also making sure that there’s money in the bank is critical. I think that you guys have done a tremendous job when it comes to raising money. How much capital have you guys raised to date?
Ali Ghodsi: We’ve raised half a billion dollars.
Alejandro: Wow. We were talking a lot of zeros before with Ben’s Investment like at 14, but 500 million is quite a bit. I’ve also heard that the valuation is 2.75 billion. So, really, really impressive, Ali. Can you walk us through what were the different financing milestones that you guys accomplished over time with the business?
Ali Ghodsi: Yeah. Our story is a little bit different in the sense that we had this technology. And as I said, eventually people started adopting it. Our issue was, every time we’d raise money in the beginning, we would tell them about how amazing the technologies are, and we’d tell them how many people in the world are using it. The first time we raised Series A, they said, “This is phenomenal. You guys are awesome, but next time you come back to raise money, make sure that you’ve made some money. So, the next time we came back, Series B, we still had only lots of adoption. We didn’t really have any revenue. This time, they told us, “Look. Last time. The next time, you probably cannot raise any more money unless you have some revenue. We repeated this for Series C, and I would say even Series D, we didn’t really have any significant revenue. Our revenue was 1 million. So that was the pressure on us was, “We see the momentum. We see that there’s a lot of excitement. Can you actually make money?” That’s why 2015 was a rough year because I think the board was starting to get the sense that, “You guys are brilliant, but I don’t know if you can make any money.” After that, Series D, E, and so on, it pretty much just boiled down to, “Can you beat the numbers? What’s the momentum? What are the metrics?” A lot of it starts becoming centered around how can you actually monetize this? That’s really the key thing. So, in my opinion, if you can sell them, it’s similar to what I said what you do as a great CEO. Same thing when you’re pitching them. Tell them that you have a great vision and that there’s a huge term. Then tell them that you have some secret sauce that others don’t have because yeah, there are great terms out there, great markets, but anyone can do that. So you have to have some secret sauce. Then you have to show them that you have some special team. Then if the team’s really strong, and they bet on the team, you have to show them that you have momentum around the numbers.
Ali Ghodsi: It’s really all that, and if that presentation is convincing, usually things go well if the markets are liquid.
Alejandro: Got it. And, obviously, amazing investors. We’re talking about Andreessen Horowitz, having Ben there, one of the founders, NEA, Battery Ventures, Data Collective, and I also see that you guys have an interesting investor, Microsoft. So what is the reasoning behind bringing a strategic like Microsoft into the equation?
Ali Ghodsi: Yeah, we just wanted to deepen our partnership with them. Microsoft has actually OEM-ed our solution. There’s something called Microsoft Azure Databricks, which is a Microsoft product. Just like you buy Office, you can buy from them the Microsoft Azure Databricks. This was obviously developed by us, and we’re actually running it for them, but it’s a very deep partnership between the engineering teams. This was going really well. So we just wanted to further deepen that partnership. For that reason, they came in as investors in our last round.
Alejandro: Got it. As we know, Ali, the journey of being an entrepreneur, is just full of ups and downs. There are those dark moments where we really get to thrive and to grow. We experience that growth that comes perhaps with breakdowns that lead into breakthrough situations. What would you say, as you’re looking back in the journey of building the business, was perhaps one moment that was really challenging for you and that ended up being a massive breakthrough for you and perhaps for the business as well.
Ali Ghodsi: I think CEO transitions are extremely hard. They’re very difficult. They’re painful for everybody involved, and they take a long time. I think if you do it really well, the company only loses one year. So I feel like it took us almost a year to both prepare that and make it happen, and then for me to settle in and take over. That was probably, for me at least, personally, the most challenging thing. But I also think for many others in the company.
Alejandro: Obviously, a hypergrowth business that you’ve built from the ground up to become a 2.7 billion valuation, that level of growth also comes with an unbelievable level of growth for you. The company is growing at 100x level or whatever multiple that is you need to be growing in parallel at the same level. So how did you manage to do that?
Ali Ghodsi: This is a great question because most companies, most large companies, people’s careers can grow much faster than the company. Most really large companies that grow maybe 20, 30% revenue every year, you can actually if you’re smart, grow your career faster than that. And you can get promoted, and promoted, and so on. In a hypergrowth startup, it’s actually the opposite – the company’s outgrowing almost every employee. You hire someone, and they’re awesome at their job. Six months later, they’re already over their skis. This is a challenge at startups, and it’s also a mindset change that you have to tell your employees about. The problem is, the norm is that you likely will get layered. We’ll probably get a new boss for you, and it’s okay; it’s normal because we’re growing so fast. We’re going to be several thousand employees in a couple of years. So this is a challenge. The same challenge applies to the CEO, which is me. So I have to keep up, and the way to keep up is, build a network of others who have done it. Pick their brain. Read lots of books on the topic, and have an open mind. Make sure that you don’t have the attitude that I’m really smart. I did this and that yesterday, so it’s also going to work when I do it today. No. What you did yesterday might not apply anymore. That technique from yesterday might not work at this scale anymore. So you have to change your ways every time. For instance, for me, transitioning from managing really smart people to managing experienced execs is very different. It turns out with experienced execs you can’t tell them exactly what they should do. You don’t hire seasoned execs in their 40s and 50s to micromanage them. You have to learn and change your way that with them, maybe you should be a little bit more declarative and tell them what the goal is and let them figure out how to do it. These kinds of things you have to all the time change your style every year to adapt to what’s needed. As you expand your business to Europe and APEC, which we’ve done now. The cultures are different. How do you manage those teams? You have to adapt all the time.
Alejandro: Got it. Talking about getting ahead of your skis and scaling because scaling also comes with another set of challenges. What is probably, let’s say from the network that you built where you had meaningful discussions to get their advice, and then also from where you have experience, what is the most common challenge when you’re scaling?
Ali Ghodsi: Well, it’s really if you’re in a hypergrowth company, it’s always the same. How do you find phenomenal leaders that can give you leverage? So in some sense, my job as a CEO is super-simple, and anyone can do it if I have an amazing team underneath me that’s really stellar, and they work well together. Then I kind of don’t have to do anything. So the question is, how do you build that team? Once you’ve built that team, how do you make sure that team is also amazing and scaling because the team that is awesome for a 100-person company is different from a 200-person company? There’s Dunbar’s law around 150. And it’s different for a 500-person company, from a 1,000-person company, from a multiple, multi-country company that spans several continents.
Alejandro: Yeah. Like you were saying, hiring great people is critical, but I think also having them working well together is even more critical. How do you think about culture, Ali?
Ali Ghodsi: Culture is super important as you scale. We have new people coming in all the time. You can do things two ways. You can micromanage anyone and put down policies and lots of rules for everything. You can tell people exactly what are the rules and exactly how to behave in every situation, and you can have sort of an iron grip. But that obviously, doesn’t scale. So the other way to do it is to have some culture principles and get everybody aligned with the culture principles. Then they can figure out what the exact rules are themselves, but at least they fit within the big umbrella of the culture you want to have. For example, if you want to have a frugal culture like Amazon, you can see if frugality is the culture principle. Then everything else will sort itself out. They’ll have processes, rules, and budgets, but you don’t have to micromanage them. So as you scale a hypergrowth startup, it’s super important that you actually are governing through culture. It’s a governance tool, and it’s a competitive advantage against your competition is the culture principles that you put in place. But it’s also rules of engagement internally. How do we actually work together? Can we criticize each other or is it okay that we get in rooms and shout at each other, or are we teamwork-oriented? Which type of company do we want to be, and what kind of behavior is okay? It’s super important that helps you also scale because it makes you more aligned internally. You have to spend a lot of energy on culture as you scale. Right after 150, 200, 300 employees, that starts mattering much, much more, especially when you start having people remote in other places of the world, other time zones far away. You have to really work hard on the culture.
Alejandro: Yeah. Absolutely because at the end of the day, every office in every different country, for example, is going to have its own culture type of thing.
Ali Ghodsi: Yeah. Absolutely. Some of the culture principles you have are maybe a little bit harder, or maybe easier to actually land in other cultures on the other parts of the planet. So you have to work through each of them and adapt them and work with the leaders there, and you have to make sure that your leaders are also continuing to also establish that culture on the next level. It’s not just the CEO’s job to talk about the culture all the time. All the leaders have to continue doing that, and at each level, they have to adapt it to what’s going on the ground so that it doesn’t become some super abstract principle that you can’t really map the reality. Then, of course, everyone knows this, but you have to hire, you have to promote people, and also sometimes, manage people out because they fit or do not fit your culture.
Alejandro: Yeah. Absolutely. One of the questions that I typically ask the guests that come on the show is, if you had the opportunity to go back in time – and I know that it is obviously impossible, but if you had that opportunity and to have a chat with your younger self, obviously now knowing everything that you know now about building, scaling, financing, everything around hypergrowth companies, what would be that one piece of business advice that you would give to your younger self before launching a business and why?
Ali Ghodsi: I would focus it on hiring leaders. I would say spend even more energy on hiring great leaders and finding leaders and growing leaders internally because that’s going to be one of your main bottlenecks. So spend a lot of energy on that, and do whatever is needed to get those people. Pay them whatever is needed, and bend over backward. Do whatever you can do in your power to grow those leaders and keep those leaders because as the company grows, that’s going to be your main bottleneck. It’s very hard to find very talented, great, amazing people who are in a place in their life timing-wise and situation that they want to join you and give their 100%. So steam and more energy and time are on that would have been something that I would have given myself advice. And try to do a lot of other things. As a CEO, you want to weigh in on all the decisions and get things right. You have an opinion on this or that, but I should have probably spent more energy just finding the right leaders, and the rest will kind of sort itself out.
Alejandro: I’m sure that there are a lot of people that are listening now that they’re probably wondering how do you find the right leaders? What would you tell them, Ali?
Ali Ghodsi: Actually, I think it’s very simple. Any hiring is a tradeoff between time and quality. So I can find the best leader easy, not even hard if you give me two, three, four, or five years. I’ll just interview all day long until I spend all my time, but that’s obviously not good enough. I can also hire extremely fast by just getting the first person that walks in the door. So the question is, how do you do both at the same time. I think the trick is you have to hire ahead of the curve. You have to start a little early. If you’re starting to hire, which a lot of CEOs do, which was my advice for my younger self as well, if you start hiring when you actually really need the person, and it’s almost you needed them to join you yesterday, it’s almost impossible to succeed them because now you’re under time pressure, and as I said, it’s a tradeoff between time and quality. So you don’t have much time now because you needed this person yesterday, and it’s going to be difficult to find that quality that you need. If anything, hire ahead of the curve. Hire when you don’t really need them. People say that about money. They say raise money when you don’t need it. Well, do the same thing with the leaders. Don’t wait until it’s absolutely necessary. Start a little bit earlier. Take your time. See lots and lots of different candidates. Go meet the best ones out there. If you can’t hire the best, just have coffee with the best, and pick their brain, and tap into the networks, and start building those networks and also remember, if you can’t get the particular person now, keep in touch. You can get them in half a year or a year. Many of my executives that I’ve hired, I kept in touch with them over a long period of time, and they joined me eventually later because as your company grows and you become more successful, you suddenly have access to talents that never would have joined you when you were six researchers out of UC Berkeley with no money or cash flow being generated. Later, when you actually have a lot of revenue coming in, and you can pay those bigger salaries, a lot of people might change their attitude about their company. The risk profile changes. Those are my advice for how to start early and interview a lot of people, and keep them in your network, and keep them warm, and hire ahead of the game.
Alejandro: I love it. That’s very profound, Ali. I love how you say that you need to hire when you don’t need it, just like raising money. I 100% agree with you. When you need money, people are not going to give it to you. And when you need to hire, it’s too late. That was great, Ali. For the folks that are listening, what is the best way for them to reach out and say hi?
Ali Ghodsi: Just LinkedIn with me. I’m Ali Ghodsi at Databricks at LinkedIn or type in: /alighodsi
Alejandro: All right, fantastic. Well, Ali, thank you so much for being on the DealMakers show today.
Ali Ghodsi: Thank you so much for having me.
* * *
If you like the show, make sure that you hit that subscribe button. If you can leave a review as well, that would be fantastic, and if you got any value either from this episode or from the show itself, share it with a friend. Perhaps they will also appreciate it. Also, remember, if you need any help whether it is with your fundraising efforts or with selling your business, you can reach me at email@example.com.