The Exchanges

Q So what about, uh, self-hosted, uh, deployments? Are you, uh, strategically deciding to not support those? Are you supporting those in some way? What's the, What's this thinking?

A 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. And an example of that would be Netflix, for example, that that's using ClickHouse, um, 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 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 its costs, security concerns, and we think that's going to open up a big market for us, especially in the enterprise.

AI assessment note: “we do work with a handful of companies that are using ClickHouse in a self-managed way”

Q You know, which was going to be like while we're on the topic, going to be my next question, like, what license is Clickhouse under, and how do you, um, yeah, how do you plan on, on, on working with the, the hyperscaler, the cloud providers?

A So it's currently governed, governed under an Apache two license, which, you know, is extremely permissive, allows for redistribution, modification. You can build a managed service around it. It obviously aids in, in growth of the project and popularity, but it comes with obvious risks that I just described. There are a variety of different licenses that we have considered, um, that we have not yet adopted, like AGPL version three, uh, or SSPL, um, which other open source companies have, have deployed. We feel that today the right decision for the community is to stay with an Apache two license, and we're excited about that. Like I said before, you know, we're not moving away from open source, quite the contrary. We're going to double down and double the size of the team of the core contributors and recruit new people into our company that understand the technology and that have been contributing to the projects in the past. Um, and at the same time in parallel, we are going to build A multi-tenant managed service in the cloud, which will inevitably be deployed on a variety of different cloud platforms, whether it's AWS, GCP, Azure, whether we go to China, like I've done in the past and partner with companies like Alibaba and Tencent and enable them to, you know, take all of the orchestration framework that we develop, which will in some of which will be open source, some of wh…

AI assessment note: “So it's currently governed, governed under an Apache two license”

Q So I'd love to, uh, go down memory lane a bit and go back to the origins of the company. So it's, it's started, the open source project started at Yandex, right? Which is the, the, the Google of Russia. I'd love for you to, uh, tell us the story.

A Sure. So it first came on my radar a few years ago, uh, the technology itself, um, as I started to observe its increase in popularity, uh, in the market. And earlier this year, I was introduced to Yandex, the CFO specifically, who then quickly brought in the founder and CEO of Yandex, the co-founder, Arkady Velos. And that was literally the start of the calendar year back in early January. And Arkady and I Started to romanticize about what it might look like to spin ClickHouse out of Yandex, as well as the core engineering team and the creator of the project, Alexa, and form a new company around this extremely popular open source database technology. And I immediately reached out to two investors who I had worked with in the past, specifically Mike Volpe at Index Ventures and Peter Fenton at Benchmark, two people that you, I believe, have interviewed in the same forum in the past.

AI assessment note: “Arkady and I Started to romanticize about what it might look like to spin ClickHouse out”

Q That's actually a really interesting thought. Any, any guess why this hasn't happened as much at the application layer?

A I think the applications are so purpose built that the business logic behind them is so bespoke that, um, it's a good question. Like we at Salesforce, there is a open source product called Sugar CRM. And I remember when it emerged and I personally saw it as a potential threat, um, especially on the low end of the market. Which was a large part of Salesforce's business, small to mid-sized companies, uh, that were, 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, but for customer service and for marketing automation, the underlying platform. We just didn't see the same pace occur, uh, with the open source alternatives and who knows what the future will hold.

AI assessment note: “applications are so purpose built that the business logic behind them is so bespoke”

Q Are there any trade-offs? We was using ClickHouse, like things that, um, you know, there's a lot of things that ClickHouse is excellent at. Are there areas where, um, it's not the right choice or not the right choice yet?

A Of course. Right. I mean, if you've got a very, of course, there's going to be technologies that are better suited for certain use cases. And, you know, that's another thing that we're excited about is the diversity of the use cases we're seeing. We talked a bit about business analytics. We can talk about observability and metrics, for example, APM, right? If you are looking to deploy Um, a hosted APM solution, and you asked me objectively, my recommendation, I'd say use Datadog. I'd say that the product, the maturity, the experience is, is considerably better than trying to stitch together something using ClickHouse, for example. Now, perhaps ClickHouse could integrate with Datadog, and we see that in the market as well for analytics, but for the actual instrumentation of the application, deploying the agents, I would not recommend ClickHouse and today for that specific use case. So there's obvious exceptions where there's going to be a better suited technology in the market. I think one thing that we're equally excited about is the amount of integrations that we're seeing be developed with ClickHouse. We talked a little bit about, you know, ingestion and we're going to be investing heavily to make getting data from Kafka to ClickHouse or Kinesis or an integration on Data transformation with dbt, you know, a much more seamless experience than it is today. We talked a little bi…

AI assessment note: “for the actual instrumentation of the application, deploying the agents, I would not recommend ClickHouse”

Q was going to be, and obviously Kafka and Confluent, uh, Confluent You know, built a, a, a multi-billion dollar, uh, company, uh, doing this. Um, but, uh, so what, what, for that first use case, what do you see, um, is, is, are we in the world of real time now? Is that an accelerating trend? Uh, and, uh, do you think that bit by bit everything becomes real time?

A I think a large number of these workloads are driven on, Latency requirements and getting these queries to perform in hundreds of milliseconds versus, you know, 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. Use billing. For example, you, you prepay a certain amount of credits to a leading AI company. Well, that AI company, you need to be able to see your use of that data to make sure that you're 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. And those types of use cases, companies originally selected technologies like Postgres, for example, for that type of use case, and then realize that as these data volumes exploded, …

AI assessment note: “we're seeing that in terms of the pull to our cloud service”

Q a clear plan in mind, um, anything that really led to, uh, being able to do this faster? Was it a question of, uh, resources and more people, or was there like a certain angle or, I don't know, product building strategies? Uh, cause like you said, uh, that's pretty much record time. Like, you know, looking around at the industry typically takes It's a lot longer, a lot more.

A I mean, it ultimately comes down to the team. We're not building a mobile app, for example, like 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, um, back in the second half of, uh, twenty-twenty-one to get the company started, is we knew that we were going to invest heavily in R&D, and both in terms of hiring engineers, but also just the core infrastructure, the, the testing And the dev and the environment for CICD pipelines was going to be expensive and it, uh, and it, and it proved to be the case. So, um, but then frankly, you know, the team that we were able to assemble, Alexei and his core team of engineers that joined us in their based in Amsterdam. Currently, um, my co-founder Uri 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 Bragan, who joined us from

AI assessment note: “it ultimately comes down to the team. We're not building a mobile app”

Q thought, right? Like no customer success, uh, department. Like I, I, um, I found that, uh, fascinating. So the, the way that works, Mechanically, is that, so the reps are in charge of, as you said, what was the expression? Cradle to grave or, um, so the reps are in charge of the selling and the upselling. If there's a problem that goes into support, and support is with product?

A No, supports 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, cause a lot of the cases that come in is when somebody's trying the product out to determine whether or not they're going to use it. And 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 a hundred solution architects or what are called, what I used to call sales engineers. And that was my first job at Salesforce was an SC, um, by leveraging our, our support team. Yeah. We've made a couple, Organizational decisions that I think are a little bit unique, um, 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.

AI assessment note: “No, supports its own function inside of the company, reports directly to me”

Q as, okay, well, this is, uh, you know, a fantastic, um, Tool for, uh, you know, all out, but like more specifically like real time analytics. And I, I think what you're, um, saying is that the ultimate goal for the company is to be, um, much more than that. I mean, basically be the, the data store for all, all things. Is that a fair way of putting it?

A Why? I point to how the market is, is reacting and we're seeing this convergence today. So, you know, traditional OLAP tools where you're running analytical workloads on top of aggregated data. There's a lot of options in the market. There's nothing new. It's been around for 20 years plus. The, the, the transactional database systems that we've talked about in the past, you know, uh, are very effective for, let's say financial transactions, credit card transactions, et cetera. And we're starting to see those two converge. And I don't think it's just click house. So I'm not going to go on record and saying that we're going to be the be all end all for every analytical workload on-prem or in the cloud. I think we're very, very attractive. Piece of technology for the majority of those use cases. And I described just a subset of those around observability, but we've got banks using us for fraud detection. There's purpose-built solutions in the market for anti-money laundering and fraud detection, specifically the financial services. You wouldn't necessarily turn to a general purpose column or data store like ClickHouse, but companies are. And so I think it's more of a trend in the industry than it is specific to ClickHouse that you're going to see this convergence occur. And again, I'm going to, I'll bring it all the way back. To, to Steven O'Gretti's article that was posted. It's …

AI assessment note: “I'm not going to go on record and saying that we're going to be”

Q competing with a bunch of different, uh, competitors. So all of this is super impressive, obviously, but also very complex. I'm curious, um, you know, especially with your background as a former head of sales about like how, uh, you, you, you, How, you know, how you sell and how you have, uh, equip a team to be successful selling a product in different verticals, different countries, different use cases?

A Yeah, it's, well, 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, and both in application side, but also in infrastructure. And then you add the dimension of open source, um, which, uh, which adds its own sets 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, Yuri, and Alexei, and we couldn't be more different in terms of our respective backgrounds. Yeah, Alexei 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. Um, Yuri's been building distributed systems on open source for 25 years at companies like Netflix and Google and Yahoo, and I've been a student of distribution, and 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 oppor…

AI assessment note: “we could spend an hour on that question alone, honestly, because it's very complex”

Q And how does that work? Uh, so you, you, you mentioned the, the, the technical aspect and the sales aspect of, of, of building a global company, but culturally, so you're in Silicon Valley or the Bay Area, uh, you have a big presence in Amsterdam. That's nine hours of, of time difference. Um, how do you all make it work, uh, logistically, but more importantly, culturally?

A Well, the company was born during COVID. So we were kind of distributed a bit out of necessity. Um, and it was intentional. And when we started the company, Alexei and his team, uh, moved to Amsterdam. And so we had a hub there and my head of Europe, Arno is also based there as well. So it was obvious for us to use that as a bit of a hub. Um, the reality is because of how much importance I place on the international markets, we're 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, you know, almost every major time zone. Um, so that was, that was an obvious decision for me, how it's working out, uh, logistically, like practically day to day is because I think if you took a look at our management team, and even if you went down, you know, 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 the generation of the people that are making decisions inside of the company. And, um, 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, 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 t…

AI assessment note: “on average, we've got more years of experience than your typical Early stage startup”

Q How do you think about the open source project these days as you become more and more of a successful commercial company?

A It's interesting. My background is not in open source, to be clear. I mean, I, my career started at Sun Microsystems in the, in the late nineties, and then I went to PeopleSoft, and then I spent 12 years at Salesforce, and then it wasn't until Elasticsearch in 2014, 10 years ago. So I guess I've been in it for 10 years, which may be longer than most or many. Did I really understand the power of open source and the pace of innovation and the benefits of the distribution model and understand the licensing complexity that goes into those types of decisions? And I think that in just how the community comes together, And, uh, how you can have hundreds of contributors, uh, around the world, uh, that improve upon the quality of the software. Um, how users, uh, and companies like Instacart can actually also be contributors, uh, to the technology that they use, further driving a faster innovation. So it's a, it's a highly disruptive, uh, technology model, I think, as everybody knows. Um, I think it does represent really the future of, of software, especially in infrastructure. I don't think that's necessarily played out On the application layer, um, as people thought it might, and who knows whether or not that'll, that'll occur, but look what's happening right now in, in the AI category in terms of the disruption of open source and how things are just changing overnight, frankly, with t…

AI assessment note: “My background is not in open source, to be clear.”

Q Is there a real-time data stack? Is that like how you think about it? Uh, I don't know, maybe Kafka to ClickHouse to something or not?

A You know, 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 is what we're seeing right now with Gen AI. And so the reality is ClickHouse has been used for 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 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, um, or they want, 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 in real time. And that database needs to be extremely resource efficient based off the volume of data that they're ingesting and extremely performant based off the latency. And so ClickHouse, I think is the clear winner in that category for both of those dimensions. And so, you know, LangChain is a great example. They recently spoke a…

AI assessment note: “I don't want to call it a buzzword, this modern data stack”

Q So maybe, uh, going in a different direction. Um, I'm curious about your, Personal experience. So you have this, uh, very, uh, interesting and illustrious background as a sales leader, Salesforce and Elastic. Um, how has the transition to being a CEO been? What was surprising, different, challenging, or perhaps not so challenging?

A Well, I wouldn't say I have this illustrious background. Let's start there. I grew up in a very modest middle-class Uh, environment, the son of an immigrant, um, went to public school all the way through university. Um, was just very lucky with some of the, the decisions that were presented to me in terms of my career and surrounded myself with, 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 cause 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 bust. And so, uh, I met with Marc Benioff and a few others and I got offered a job. That's really otherwise.

AI assessment note: “Well, I wouldn't say I have this illustrious background. Let's start there.”