How ClickHouse powers Netflix, Uber and Spotify’s Analytics | Aaron Katz, CEO of ClickHouse

Thanks, Matt. Thanks for the invitation.

Well, let's start with the technology. So ClickHouse is an analytical database that was created by one of my co-founders, Alexey Milovidov. And it was initially designed to power web-scale analytics. So think about a database that would sit behind something like Google Analytics, but in this case, it was Yandex Metrica. And so, internet-scale event data. And it was open-sourced in 2016 under an Apache 2 license, which is still the license that governs the project.

And it took off in popularity, quickly became one of the most popular open-source databases in the world. And that's when I first discovered it, when I was at the previous company that I worked at, another open-source technology company. We started to see ClickHouse pop up in the wild, and we actually started to lose some pretty prominent workloads to the technology. So it was on my radar. And then in 2020, 2021, we embarked on a process to create a company around it.

Now a Delaware corporation with headquarters in Silicon Valley, it exists as a venture-backed startup. And we are taking this very powerful analytical database that's used by companies like Spotify and Netflix and Disney and Uber and eBay and Cloudflare, and we're building a managed service in the cloud that we're calling ClickHouse Cloud. ClickHouse, as you know, is columnar, versus something that's row-oriented like Postgres or a transactional database. And so it aids in very fast aggregations. People claim that it's one of the fastest and most resource-efficient databases that they've ever used, and that combination is pretty impactful.

Faster, cheaper. And the database market seems to be something that people are drawn to. And it's just the way that Alexey and his team have been thinking about the design of this technology over the past 10 years in terms of how they think about data compression, how they think about data ingestion. We're positioning the company and the technology as the real-time data warehouse in an effort to essentially create a category here that takes what technologies like Snowflake did so well.

And architecturally, they're not that dissimilar. I mean, Snowflake is a columnar data store with the separation of compute and storage, like we have with our cloud service. But I don't believe that it was originally engineered to power these extremely low-latency, high-concurrency analytical workloads. And that's what we're hearing from companies that are looking to migrate, frankly, off of technologies like that for more resource-efficient, less expensive analytical workloads. And so we've coined this phrase and we're embracing it, and the market seems to be responding pretty well to it.

And, for example, we launched our cloud service a little over a year ago, and we had this phased release where we had a private preview and a public beta and then made it generally available. And over the course of that year, we've now accumulated over 1,000 customers on ClickHouse Cloud. And it is largely self-service, PLG, whatever people call it these days.

It's not a new term. When I was back at Salesforce 22 years ago, we had a free trial that was for 30 days, and you could add your credit card and start using the service. But yeah, so that's kind of what we're seeing people come to the service for. And it's primarily driven around both cost and performance.

It's very broad. As you know, it's a database. So the use cases are very diverse, from fraud detection to sentiment analysis, sales and marketing analytics. I try to kind of bucket it into three broad categories. The first would be somebody that needs very low-latency, high-concurrency analytical workloads, typically to power, let's say, a B2B SaaS application. That's about a third of the use cases that we see. Another third would be traditional cloud-based data warehouses that are not providing the same performance or cost efficiency that ClickHouse Cloud does.

So we're seeing those workloads migrate to our cloud offering for those specific use cases where you think about a traditional data warehouse. And then the third would be observability workloads. So traditional logging, trace analytics, or metrics. And I think if you looked across our customer base and our user base, they would all more or less fall into one of those three.

I think a large number of these workloads are driven on latency requirements and getting these queries to perform in hundreds of milliseconds versus seconds or minutes. And we're seeing that in terms of the pull to our cloud service and why people are adopting this type of technology. And that didn't just start a few quarters ago. That started a few years ago, I think, in the requirement to have this very rich, immersive analytical experience where your data is being displayed almost simultaneously as it's being streamed in.

And if you think about some real-world use cases, like usage data, for example, a lot of the customers that we have are storing data in ClickHouse Cloud that is the use of their service, and their customers need to be able to analyze how they're using that service. Usage billing, for example: you prepay a certain amount of credits to a leading AI company. Well, that AI company, you need to be able to see your usage data to make sure that your costs are controlled.

That needs to be extremely performant. So you need to be able to see that and get the results of that billing usage data in several hundred milliseconds. And that's the power of ClickHouse. In those types of use cases, companies originally selected technologies like Postgres, for example, for that type of use case, and then realized that as these data volumes exploded, especially with the emergence of GenAI, that they simply weren't scaling and needed a more resource-efficient and performant analytical database like ClickHouse.

So there's a lot of e-commerce use cases. Instacart's a very prominent one and one of the more recent ones that have talked about their use of ClickHouse. And in an effort to obviously control their spiraling data warehouse costs, they've looked to ClickHouse to help solve that problem, but also in terms of performance. And it's not just fraud detection at Instacart, but A/B experimentation and other kinds of e-commerce analytical workloads that they're using ClickHouse for. We're really excited about that user.

So specifically, that's how I first encountered ClickHouse when I was at Elastic, the company behind Elasticsearch. And we started to see some pretty prominent logging workloads or metrics workloads migrate over to ClickHouse. For example, eBay had a large metrics project and they chose ClickHouse. Cloudflare is a very well-documented use case. These are all in the public domain. We started to see it emerge at Netflix, for example, at Disney, at Comcast.

Yeah, so when we were starting the company, and it took a little under a year to get the company formed, the business model to me was crystal clear. It was: we were going to build the next-generation real-time data warehouse in the cloud, and we're going to call it ClickHouse Cloud. And as you know, open-source companies typically start by selling support and an enterprise version of their product or proprietary plugins, and they bundle those and sell subscriptions, and they kind of bootstrap the company, generate early revenue through that mechanism, while they subsequently stand up a cloud service.

And that's how Mongo did it, with Atlas subsequently becoming such a popular cloud offering. I think they just passed $1 billion of annual revenue. Elastic Cloud, Confluent Cloud. And having been through that journey, I just decided to skip that first step and go straight to cloud. And so, we started architecting essentially what we were going to build before the company was even formed. And that was really necessary for Yury Izrailevsky, who's my other co-founder, who I met when he was running platform engineering at Netflix, and he had since gone over to Google, and myself, to kind of align on that strategy.

So that was clear from day one. And we actually documented the process, and we released a blog about it. It was quite popular in terms of how we made these architectural decisions to get this product launched in less than a year, which, from my understanding in the market, was quite fast. And from my experience in enterprise software over the last two decades, it was definitely the fastest product launch that I've ever been part of by a long shot. And we sequenced it to where we had early design partners and then a public beta.

And there were a couple of critical decisions that we needed to make at the time. Obviously, we wanted a serverless architecture, so the separation of compute and storage. To accomplish that, we had to write a new table format, or a new table engine, that we call SharedMergeTree. And so Alexey and his team got to work on that, and they got that done. And we then had to decide what cloud provider we were going to start with, and we chose AWS, and we then subsequently deployed on GCP, and we'll be deploying on Azure and launching that service in public beta in two weeks, on March 21st.

And each one of those offerings has grown significantly over the past year. But yeah, I would encourage anybody that's interested to read that blog. We also wrote another blog called "Unbundling of the Cloud Data Warehouse," and that kind of further answers your question about how ClickHouse Cloud differs from traditional cloud-based data warehouses like Snowflake.

I mean, it ultimately comes down to the team. We're not building a mobile app, for example. This is a pretty complex piece of technology. And so we knew that it was going to take a lot of resources. And that's partially why we raised as much capital as we did back in the second half of 2021 to get the company started, is we knew that we were going to invest heavily in R&D, both in terms of hiring engineers, but also just the core infrastructure, the testing.

And the dev environment for CI/CD pipelines was going to be expensive. And it proved to be the case. But then frankly, the team that we were able to assemble, Alexey and his core team of engineers that joined us, and they're based in Amsterdam currently. My co-founder Yury has been building distributed systems on top of open source for 20 years.

So it's not his first time doing this.

And then Tanya Bragin, who joined us from Elastic, and she ran their observability product line to be our head of product. These are folks—a woman named Rupa, that joined us from Netflix—they just have extensive experience building these types of cloud services.

It's the bulk of the company. So we currently have about 170 employees. And 80-plus percent of those are in technical roles or in engineering roles. We've got a very small sales team, I think, relative to a company of our stage. And that again was very intentional. We didn't want to build on the go-to-market side until we had established product-market fit and had referenceable customers. And we also didn't want it to be a crutch for product and product quality and the ability for the product to distribute itself and for engineers to come to ClickHouse Cloud to create a free trial, to load their data, to run their queries, and to add their credit card without ever needing to talk to anybody in sales.

And I think that's what's really propelled the company to the current stage and has amassed 1,000 unique companies on the cloud service that frankly don't need to talk to anybody in sales if they don't want to. Now, these larger deployments and some of our larger customers obviously want to engage with our team, in which case we do. And there's a dozen or so folks around the world that are supporting those types of larger enterprise customers. But the long tail of companies, companies like LangChain, for example, they simply can get going without ever talking to anybody inside the company.

It's interesting. My background is not in open source, to be clear. I mean, my career started at Sun Microsystems in the late '90s, and then I went to PeopleSoft, and then I spent 12 years at Salesforce. It wasn't until Elasticsearch in 2014, 10 years ago, that I really understood the power of open source, the pace of innovation, the benefits of the distribution model, and the licensing complexity that goes into those types of decisions. So I guess I've been in it for 10 years, which may be longer than most or many.

And just how the community comes together, and how you can have hundreds of contributors around the world that improve upon the quality of the software, how users and companies like Instacart can also be contributors to the technology that they use, further driving faster innovation. So it's a highly disruptive technology model, I think, as everybody knows. I think it does represent really the future of software, especially in infrastructure. I don't think that's necessarily played out on the application layer as people thought it might.

And who knows whether or not that'll occur. But look what's happening right now in the AI category in terms of the disruption of open source and how things are just changing overnight, frankly, with the release of some of these open-source models.

I think the applications are so purpose-built that the business logic behind them is so bespoke that it's a good question. Like, at Salesforce, there is an open-source product called SugarCRM, and I remember when it emerged, and I personally saw it as a potential threat, especially on the low end of the market, which was a large part of Salesforce's business: small- to mid-sized companies that were using Salesforce and paying per seat. Well, now they had an open-source alternative. But the reality was, the business logic that was required, that was built into Salesforce, both for Salesforce automation, customer service, and marketing automation, the underlying platform—we just didn't see the same pace occur with the open-source alternatives.

And who knows what the future will hold.

So we do work with a handful of companies that are using ClickHouse in a self-managed way, whether it's on-prem or they're deploying it in a public cloud provider, but they're not using our hosted service. An example of that would be Netflix, for example, that's using ClickHouse in AWS, but they manage that environment and we work with them. We provide technical support, and there's a dozen other companies that we also provide technical support around, but it's not our primary business model. Ultimately, we want to understand their use case so that we can improve upon the technology, make them successful.

Ultimately, we'd love to migrate them to our cloud offering, whether it's the current multi-tenant serverless offering or, in the future, a deployment model that we're going to launch later this year called Bring Your Own Cloud, where the data plane will sit behind the customer's VPC. We'll still manage the control plane. And so, for data residency requirements, data locality, whatever it is—cost, security concerns—we think that's going to open up a big market for us, especially in the enterprise.

So I think I mentioned Azure. Again, that's going to open up a big market. We have almost 500 companies on the waitlist for our Azure offering. It hasn't even been released in public beta yet. So we're excited about the demand we're already seeing from the market. I mentioned Bring Your Own Cloud, which will be a new deployment model that I just described. In China and other parts of Asia, we've partnered with Alibaba, specifically Alibaba Cloud, for them to redistribute our cloud binary.

As an enterprise version, and we have a revenue share agreement in place with them for the next few years. So we're excited for that product to get off the ground. It just launched recently, last month, after a year and a half of working with Alibaba. So we see that as a great way for us to work in the Asian markets. We're entering Japan, for example, and deploying on the AWS Tokyo region in early June. We're seeing a ton of demand come from the Japanese market.

From my experience at Salesforce, Japan was a huge market for that company. And I really see an opportunity for us to be successful there. Yeah.

That's right. Yeah. I often make the same point that people underestimate it, and also just how difficult it is to get it right. But one of the largest electronics manufacturing companies in the world is a customer of ours on cloud, specifically in Asia, and not even in Japan yet. And that's Sony. So we're really excited about leveraging that type of initial success already. And what's interesting about some of these transactions is how the marketplace activity works through both AWS and GCP, and what will be Azure as well, because when we simultaneously release on this cloud service provider, we also release in the marketplace.

And that's really opened up a new channel for us. There's so much more on the roadmap, especially around ClickPipes, for example, around data ingestion.

That's right. So the integrations that we do with Kafka, Confluent, or Redpanda.

I try not to—I don't want to call it a buzzword, this modern data stack. But if you've been in the industry as long as I have, you see a lot of things come and go pretty quickly and get disrupted even faster, which is what we're seeing right now with GenAI. And so the reality is ClickHouse has been used in AI applications, or to power AI applications, as a feature store for years. This isn't something that just emerged over the past year.

So that's a use case that's been very common. And now people are storing embeddings in ClickHouse, and it's used for vector search, for example. So we're seeing that grow. The way I describe it simply is kind of picks and shovels for this AI movement, where a lot of these AI companies need an analytical database, and they don't want to have a bunch of bespoke databases for each one of these use cases. They don't want to have a vector database and then an analytical database and a transactional database.

They want to have one data store where they can put all of this information and query it in real time. And that database needs to be extremely resource-efficient based on the volume of data that they're ingesting and extremely performant based on the latency. And so ClickHouse, I think, is the clear winner in that category for both of those dimensions. And so LangChain is a great example. They recently spoke at one of our meetups in San Francisco about how they're using ClickHouse Cloud in LangSmith, for example, and other leading AI companies are using our technology.

Yeah, it's a new connector that we've launched with Fivetran that we're excited about. And you'll see more there. There's some notebook integrations that we've also authored for those that want to have a data notebook experience integrated with ClickHouse Cloud. We talked about the ingestion layer, things like Kafka Open Source or Confluent Cloud or Confluent Enterprise, their on-prem product. Redpanda we're seeing emerge as a data ingestion format that people are excited about.

Data flowing in.

Then you've got the ETL providers, whether it's Fivetran or Airbyte or others that we're—

And that really varies based off use case. If you look at observability or metrics use cases, for example, Grafana is a very popular visualization tool. And so we've got a connector that we worked with Grafana on that's very popular. You think about BI workloads: Tableau, Power BI, Looker, for example.

Which company?

Yes, Rill Data. Mike Driscoll is excellent. So we're excited about our discussions with Rill, a company called Omni. Jamie Davidson, who is head of product at Looker, he and I met last week to talk about what they're building and how we could work more closely together. And then other open-source alternatives, Superset, for example. The company Preset's very popular. So making sure that we support those integrations.

Yeah, it's an interesting question. So we obviously measure this with all of our sales engagements and with all of our customers that migrate from alternative solutions, and it obviously varies based on the use case. So let's start there. I can't answer the question and give you our top competitors X, Y, or Z. It really varies on the use case. What we are seeing trends form around are companies that are migrating off of other OLAP databases. Druid is an example where Lyft, the ride-sharing company, recently migrated off of Druid and onto ClickHouse.

eBay is another example of a company that migrated off of Druid and onto ClickHouse. There's other open-source OLAP databases, but in the diligence that I did, ClickHouse was the winner in the category. We're seeing companies that chose a relational or transactional database like Postgres, for example. Perhaps it didn't scale, it didn't give the same performance that they needed. And so we see a lot of migrations from those technologies and companies that support those technologies. In the cloud data warehouse market, to answer your question specifically around Snowflake, we do see demand for customers to migrate off of technologies like BigQuery and Redshift and Snowflake, either for cost or performance or both.

And there's others, but those are the three predominant suppliers in the cloud data warehouse market that we see. And then for observability workloads, we see companies migrate off of, depending on the use case, Prometheus, Elasticsearch, and other popular backends for those types of workloads. So I think that's why we're so excited about this opportunity, is the database market is enormous, therefore it's very crowded. Nobody knows this better than you. And so the competitive landscape is very, very broad.

Yeah, we could spend an hour on that question alone, honestly, because it's very complex, as you know, in terms of how you take a product to market and how you distribute it, both on the application side but also in infrastructure. And then you add the dimension of open source, which adds its own set of challenges. And so, fortunately, I think that's potentially one of the most unique aspects about our company, is the three of us started the company—myself, Yury, and Alexey—and we couldn't be more different in terms of our respective backgrounds.

Alexey created ClickHouse. It's his life's work. He named it; it's short for Clickstream Data Warehouse. So he was thinking about data warehouse use cases when he formed this technology over a decade ago. Yury's been building distributed systems and open source for 25 years at companies like Netflix, Google, and Yahoo. And I've been a student of distribution, with most of that time being spent at Salesforce and then Elastic. And about a third of my career has been spent outside of the U.S.

I lived in Singapore for a number of years. I've lived in Europe on two different occasions. And really, there's no way to shortcut knowledge of how to go into those markets unless you live there. And so I understand the challenges and complexities with doing business internationally and the enormous opportunity of getting it right early. And already over half of our revenue comes from outside of the United States. And that's largely driven by this frictionless self-service motion, where a company in India can onboard onto our technology just as easily as a startup in San Francisco.

And that's showing up in the numbers, which we're really excited about. We use a PEO called Deel, or EOR, whatever you want to call it, to hire these folks remotely in countries where we do not have a legal entity. And Alex and his team have been excellent in providing support to help this international expansion, because we've been growing quickly. We basically went from zero to 150 employees in two years, and we're going to add another 100 in the next year.

And half of those folks are outside of the United States. We've got a ton of people in Western Europe. We've got a team in Australia. We now have a team in Singapore. And so it poses a huge opportunity, and it poses a lot of operational complexity. And then you think about, how do you then cater to these different market segments? And I try to simplify this for people that ask me questions about, okay, I need go-to-market advice. I'm like, well, think about it.

So, put product aside, which is impossible to do, but put the product quality aside. You need to think about internationalization. You need to think about industry orientation. So how do you cater to different industries? Because a bank's going to talk in a very different language than a telco. And then you need to think about market segmentation. So how do you sell to a small company very differently than how you sell to a large enterprise company? And you then need to structure your go-to-market organization to accommodate those three dimensions: internationalization, industry orientation, and market segmentation—the size of the company.

And you're going to have different comp plans to cater to those different segments. You're going to have different messaging to cater to the different industries, and you're going to have different internal requirements to support international expansion, which can be challenging in certain markets. So that's how I think about go-to-market, frankly.

So typically everything starts self-service. That's kind of the top of the funnel. Pretty much somebody could reach out to us.

It's embarrassing to say we really don't right now. We should. We don't. The sales team is very busy just responding to the amount of demand that's coming to us. And it really starts at the top of the funnel when somebody creates a new trial, and we provision credits, we give them a certain amount of time to consume those credits, and we monitor, through telemetry, the engagement. Or they come to us and they create a support case and say, I need help with data onboarding or optimizing my queries or things like that.

And we engage with them. Or somebody will come to us, a large company, and say, I'm looking to migrate off of another piece of technology. We need to do an evaluation, a proof of concept, blah, blah, blah. And so we engage with those folks that way. And I've gone on record to my company, and I'll probably eat these words at some point, and I'm going to contradict myself in the future, and I think I'm largely to blame for this, is I said, hey, success for our company will be if we never have SDRs and CSMs.

And again, I've formed these functions in the past, I've run these functions in the past, so it's not to take away from SDRs and CSMs, but in my experience, both of those functions can often be a crutch for a broader issue inside of the company. It could be around product quality, it could be around marketing and demand generation, it could be around the effectiveness of your sales team, it could be around your renewal rates.

But introducing all of these different functions into the sales cycle, in my experience, is clumsy for the customer's experience. And if you can enable your direct sales team to handle a lead from cradle to grave, I think you're going to drive efficiency as part of your sales organization. I think you're going to improve your customers' buying experience, and you're going to have higher retention and higher expansion because nobody is better equipped to manage that renewal cycle and that expansion than the original salesperson that sold that account or that manages that account, because they know the use case, they know the buying behavior, they know the procurement cycles, they know where the bodies are buried inside of that account.

And having that knowledge transfer occur constantly, it's very difficult to do when you're growing a company as quickly as we grew Salesforce and as quickly as you grow Elastic, because you're constantly carving up territories, and so you're constantly introducing new reps. So that's also intentionally why I've kept the sales team as small as we are, is to really push them to do as much as possible before we really put our foot on the gas with distribution. But if we can continue down this path of self-service the way that Datadog did so effectively in the early years, I think we'll be in really good shape.

No, support is its own function inside of the company, reports directly to me, and it's technical support. It's what you'd expect.

Now, what's different about how we're approaching it is our support engineers are as involved in pre-sales as they are in post-sales. And so, because a lot of the cases that come in are when somebody's trying the product out to determine whether or not they're going to use it, those technical cases are sometimes not that dissimilar from the cases that come after the sale. And so that team as well is also helping us not need to hire 100 solution architects, or what I used to call sales engineers.

And that was my first job at Salesforce, was an SE, by leveraging our support team. Yeah, we've made a couple organizational decisions that I think are a little bit unique and sometimes controversial. And you could argue they could sit in certain areas. Product marketing is a great example. Should that sit in corporate marketing? Should that sit in product management? Because of the technical nature of our sale, we felt that that should sit in product management.

Brand, demand gen, events, DevRel probably would be the four ways to think about it. So community, we're very event-heavy. So I was just in Vegas for Google Cloud Next. I'll be in Vegas for AWS re:Invent.

It was great. Yuri and I, in our late 40s, are at the booth handing out socks, scanning badges, chasing down leads. It's a lot of fun. And yeah, we'll be at Microsoft Build here shortly. So very event-heavy. We do a lot of the regional events, the regional AWS events, and then we do a ton of meetups. We just had a meetup here in New York where we had over 150 people show up. And that's a great way to foster the community.

No, we do. We were just up in Seattle meeting with large hyperscalers about this specific topic around go-to-market, aligning with their industry leads, talking about the references that we have, the migrations that we're leading. And a lot of these large CSPs are highly incentivized to capture competitive workloads, so moving from one cloud to another or moving from on-prem to cloud. And we have a great story to tell for both of those types of migrations of companies that are running ClickHouse on-prem, or they're running another piece of technology on-prem and they want to move to ClickHouse Cloud because we've got all of these operational benefits around backups and version upgrades and security and integrations.

And so they've got a ton of incentives to align with us on the go-to-market side, but also some of the technologies that I also referenced before. Confluent is another example where we're excited to partner with them on the go-to-market side.

I think as early as possible. These relationships take time to form, and especially in the world of open source, if you don't do it, someone else might. And so you want to be perceived to be the company that is really driving the innovation in your category and not let other people capture that message.

Well, I wouldn't say I have this illustrious background. Let's start there. I grew up in a very modest middle-class environment, the son of an immigrant, went to public school all the way through university, was just very lucky with some of the decisions that were presented to me in terms of my career, and surrounded myself with what I think are very smart people that helped me make those decisions. The reality is the only reason I went to work at Salesforce 22 years ago was because I did not get into business school.

I always wanted to go to a leading business school, so I applied to Stanford and Harvard and MIT, where I got waitlisted, and I didn't get in. It was back in 2002 when a lot of people thought that it was a good idea to go back to school.

Because the dot-com boom had just occurred and busted. And so I met with Marc Benioff and a few others, and I got offered a job. That's really—otherwise—

I honestly was waiting tables up in Lake Tahoe at a Mexican restaurant.

I was unemployed. I had not gotten into business school, and I didn't know what I was going to do next. I needed a job, basically. And my sister recommended Salesforce as a company. She knew the company. She was in technology as well. And so I thought, well, I've got to get in front of this guy, Marc. And what's the best way to do it? I need to activate my network. So I basically just started talking to anybody that would listen in San Francisco that had a relationship with this company.

And through a friend of a friend, I got my application in.

And I met with Marc. I remember that interview very vividly. Yeah. Nevertheless, so that was my time at Salesforce. I've completely forgotten the question that you asked.

Was it like a three-year-old startup? Started in '99, so about 150 employees, maybe a little bit less.

Yeah. So I spent 12 years there. I just—

Yeah. But then I quickly moved into sales management, honestly, six months later, running the West for their small business sales organization. The company went public in '04. In 2005, Marc asked if I would move to Singapore and help open up their Asia Pacific headquarters. I did that, ended up staying there for almost four years, helping expand Salesforce across Asia, then went out to Europe. Spent time in Dublin and London, helping expand Salesforce across Europe, and then came back, ran Latin America for a few years.

So I just was kind of one of his have-bag, will-travel sales leaders for a long time.

Nevertheless, 12 years later, I just needed to leave home professionally, so to speak.

And I was at that point ready to take on a bigger job at a smaller company. And two investors, Peter Fenton and Mike Volpi, helped recruit me out of Salesforce and into Elasticsearch at the time. Spent six years there.

That's a pretty accurate description. Fifty, 60 employees.

Very small. They had a head of sales who's excellent, Justin Hoffman, who's a good friend of mine today, and he and I spent six years growing that company together. So they already had a great sales leader. They just needed somebody maybe that had the international background, who knows, to help scale it. So I joined that company, brought some people with me on the go-to-market side. We took the company public four years later in 2018, and I spent a few more years then before stepping out during COVID. That's incredibly quick.

That's right.

It was a lot of fun.

Yeah, it was a lot of fun. And it's—anyway, so then took some time off and did some soul-searching on what I was going to do next. And I had a few different options. One, I could just stay as kind of a mercenary sales leader and go be another CRO at another kind of later-stage private company. I could do what you do. It was kind of in vogue at the time for washed-up go-to-market people to become investors.

I tried that for a period of time, but it wasn't my calling, at least at that stage of my life. And I just kept coming back to the fact that I wanted to be a CEO and run a company. And ClickHouse came and was presented to me, and here we are.

I mean, outside of my early years at Salesforce, and it was such a different stage of my life, this is the happiest I've ever been professionally and the most fulfilled. And it is entirely driven by the team that Yury, Alexey, and I have been able to assemble. It's just really high quality. And I've worked for some great companies in the past, but being able to do it on my terms is probably the most refreshing part of the job.

Well, the company was born during COVID, so we're kind of distributed a bit out of necessity, and it was intentional. And when we started the company, Alexey and his team moved to Amsterdam. And so we had a hub there. And my head of Europe, Arnaud, is also based there as well. So it was obvious for us to use that as a bit of a hub. The reality is, because of how much importance I place on the international markets, we were going to be a global company very, very early.

Plus, when you're supporting a managed service around the world, you're going to have SREs in almost every major time zone. So that was an obvious decision for me. How it's working out logistically, practically day to day, is because if you took a look at our management team, and even if you went down a layer or two, I would say on average we've got more years of experience than your typical early-stage startup, is how I would diplomatically state it.

The generation of the people that are making decisions inside of the company. And so we've had the benefit of working for other companies in an office setting in the past, and we've been able to take, I think, what's best about that and leave what didn't work necessarily, both about an office setting and a remote setting. And I've worked in both of those environments. Honestly, I'm getting ready to go back to an office on a more regular basis, two, three times a week.

So I would predict that at some point in the next year or two, we may have some sort of setup in the Bay Area where there's a concentration of folks. But again, we've been hiring people around the world, and we're not going to have some return-to-work mandate for people that are obviously distributed. And we've been able to tap into some amazing talent as a result of that. We're containing that, however. So I think we have employees right now in 15 different countries.

And we're not adding to that list except for Japan.

So all I can do is draw upon my experiences to answer that question. And what Salesforce did, which I think everybody can claim is one of the most successful technology companies of the last 30 years, is they established market dominance in one product line while they started to incubate new product areas. And it started in Salesforce automation, and then it went to customer service and support, and then it went to marketing automation, and then the underlying platform. And they've obviously expanded past that.

They didn't go into those new areas until they had demonstrated success as a clear market leader in the first product line, then the second. And so while a lot of it was run in parallel, there was a very logical sequencing to that. And the first product that we've launched, ClickHouse Cloud, supports a variety of different use cases as a database, is growing very quickly. And so I would say for the next year or two, it's really just doubling down on the formula that is clearly working.

There's sustained demand for this offering. That was my biggest concern, is we're going to tap into this latent demand, and then we're going to see this deceleration that hasn't occurred. In fact, the opposite has occurred. We're seeing reacceleration almost at every level of the business. And so what I don't want to do is say, okay, well, all of a sudden now we need to pivot the company and go into this other attractive use case or this attractive use case. We can incubate ideas inside the company.

We can make some small acquisitions, obviously, to try to accelerate that inorganically. But if I look out for the next few years, I'd say that's most likely going to be more of the same, and then just rapidly innovate. And you asked about the roadmap, and I just scratched the surface in terms of what we're going to be delivering over the next 12 months. There's so much more coming that if we can execute against that, the market opportunity alone just for this first product that we've offered is absolutely enormous.

The company is less than three years old. So if you ask me to look out five years from now, it's a very difficult exercise to do. But if we continue to grow the way that we're growing, five years from now, things will look very different than they do today in terms of the health of the balance sheet and the income statement.

You bet, Matt. Thanks for the invitation.