Olivia Hack
Welcome to Innovators and Ideas, an audio series from RBC Capital Markets where we engage with leaders driving change and innovation in today’s world. Today, we’re coming to you from RBC’s Global TIMT Conference in New York City speaking to Jay Kreps, CEO at Confluent whose data streaming platform is helping customers address the complexities of data management. Jay, thanks for joining us today. Could you briefly describe your business to us and the types of audiences that you serve?
Jay Kreps
Yeah, absolutely. Confluent is a company that's all about helping manage real time data. So, there's been lots of companies that focus on data storage or data at rest, but we're all about data in motion. As things happen, how can you process it and use it?
Olivia Hack
Real time. Love it. So Kafka revolutionized real time data streaming… as you transitioned to founding Confluent. What were the key learnings from the open source project that shaped your vision for the company?
Jay Kreps
Yeah, so the the original technology, Kafka, was an open source project, and it was developed at LinkedIn, where I worked, so it was a kind of internal infrastructure layer for helping the company work with data, and we released it as open source, and a lot of other tech companies started to adopt it, kind of Netflix's and Ubers of the world. And as we started to talk to some of these users, we realized it was actually kind of a big thing. There was a lot of opportunities, and not just in tech, but like companies all over. And when we thought about, well, can we accomplish this big transition from more kind of slow batch usage of data to real time. Can we accomplish that working kind of from the basement of a social network. We thought probably not... Probably there needs to be, like some company that tries to make this happen say something like this. Yeah. So that was what caused us to try and make it real. Yeah, and the idea behind the company was, make some, make something that was easy to consume. The idea of working with data that was up to date in real time, that's appealing, but it has to be as easy to use as all the other data technologies that companies have. We wanted to make it available as an enterprise product, available as a service in the cloud, make it something that was easy for developers to work with, kind of all the bells and whistles that makes it technology successful.
Olivia Hack
Fantastic. Ok. How are you helping customers address the complexities of data management and what are some of the important differences between Confluent and Kafka that's led to this over 40% of the F500 to adopt Confluent?
Jay Kreps
Yeah. The fundamental challenge that streaming helps with is, companies just have a lot of fragmentation. And yet, a lot of what companies want to do, to be intelligent, to serve customers better, to be relevant in how they interact to be efficient in their operations, it requires putting together lots of parts of that. And so the idea of streaming is to be able to, you know, connect all these little islands and allow data to flow between them, you know, transparently, quickly, in sync with the business, and allow the development of applications that work with that as it occurs. And so, you know, that's kind of the big idea. What we've done to make that accessible is really turn it into something, and it's a developer movement that people know how to build against and work with, which was kind of a new thing in the industry.
Olivia Hack
So on your recent earnings call, you'd noted Confluent is entering the third wave of growth. Do you want to walk us through this evolution, what it means for the company in terms of your long term growth and margins?
Jay Kreps
Yeah, yeah. Absolutely. So the goal of the company was really kind of shift the world to working with data in real time, dynamically. The first part of that was really built around this open source Kafka. So we started with the easiest thing to do as a very small company, which is a licensed software offer. We said, ‘Okay, we'll give people a product which helps them take this, put it in their data center, manage it, you know, run it reliably’, And that was our initial offering. As we were going public, that was the bulk of the business, yeah. And this was in 2021. But we had been investing in this second wave, yeah, which was our managed cloud service. There was actually a lot of work that had gone into building a world class cloud service that would work across the different cloud providers, that real customers could rely on and use. When we think about what's next I think what's most exciting is what I was kind of calling this third wave, which is, broadening from just having the stream of data and helping companies kind of get it from place to place, to really having, like, that rich set of capabilities. About, how do I work with data in real time? So how do I capture it? How do I process it? How can I govern it? How can I do all these things and have all those parts really come out of the box and work together? So then, the question is, well, okay, what does that mean for us as a business? And I think there's a there's a couple of different impacts. The the first order effect is like, well, we have some more things to sell. So that that helps to drive growth and drive expansion, but more than that, it actually makes this stuff easier for customers. So the set of use cases that they can go after, and the amount of investment to realize some value goes down. And so kind of the the opportunity then expands proportionally as a result of that. And we think that that's kind of the big opportunity here for us. The aspiration is to create something that's as important a data platform as data warehouse and operational databases have been something that's really kind of the central nervous system that has all the real time data flows across the company. That's an important base of application logic and the intelligence of a company, and that's not something you do overnight, but we're trying to lay down the kind of core foundation for that to be built.
Olivia Hack
So I want to talk a little bit about some forces of change, and where companies that you know you're working with are struggling the most with their data management today, and how you are really positioning yourself to address some of these challenges and support customers and really optimizing their strategies.
Jay Kreps
For most of these organizations, there's a complicated set of what are effectively legacy technology investments they've made, you know, and this could be old relational databases, even old mainframes, old software systems that are just running big parts of their business, in some cases, big parts of the economy, right? Of course, they're very valuable, but they're not really the platform that people want to build on going forward. So that's the old, right? Then there's the new, yeah, there's a there's a ton of opportunities. And what companies want to do in terms of their next gen customer experience opportunities around AI, how they're going to harness data. So the goal of data streaming, you know, it's interesting. A lot of new tech companies, it's always like, hey, delete all the old and replace it with our new. That's kind of the message. And the challenge is, it turns out it's pretty hard to delete all the old. The nice thing about data streaming is you don't have to delete all the old. It really is about connecting to the different parts of the company. So connecting into some of the older systems that have critical data about the business, opening that up to some of these newer layers, the stuff that's happening in the cloud, some of the AI applications, I think that's a fundamentally, you know, more tractable story of technology, because it kind of works with companies where they're at and it unlocks value in the business. And, it doesn't require, you know, rebuilding everything from scratch.
Olivia Hack
How do we immediately enhance, something…
Jay Kreps
That's right, yeah, that's right.
Olivia Hack
Fantastic. Okay, great. So, we've talked a little bit about forces of change and how do you see these forces of change reshaping the data landscape and how are you responding?
Jay Kreps
Yeah, there's a lot happening. I think the migration to the cloud continues. That's been kind of the big background force. It's played out in an interesting way, where you haven't really seen data center spend come down that much, but you have seen most of the new stuff in the cloud. I think broadly, there's ups and downs, but that continues. And companies are kind of in this complex hybrid environment where there's stuff in data centers, there's stuff in different cloud providers. The addition of AI complicates that. A lot of that is harnessing bits of data and technology that are spread across a lot of systems. Yeah, so for us, what do we need to do to be successful in that environment? It's continuing to really make this hybrid architecture work well, right? Be able to connect through different parts of the business. It's integrating into the kind of next generation of AI technologies, allowing people to bring these language models to bear on the real time data as it occurs. That's an important part of these next gen applications that harness AI. And so those are some of the investments that we've been making. It goes along with the broader platform that we're building and it's not surprising that anything that's about real time data is going to be kind of a foundational part of that AI story.
Olivia Hack
And let's double click a little into that actually. We know data streaming has been increasingly critical for real time AI. How do you envision Confluent shaping that intersection between real time data and AI in the years to come?
Jay Kreps
Yeah. So the architecture that companies are pursuing is, how can I both gather context data, right, and be able to kind of store that in a way that's useful in these AI applications? This often has the acronym RAG retrieval, augmented generation. But, the part of the problem we would serve there is this kind of supply chain of data. How do I get it right? And then also, how can I bring that to bear. In my business. So, there's some something happens in the business, and there's some unit of work to do, how can the AI go do it, right? So an example of this would be like, we have a customer that's an insurance company. They do claims processing. So claim is filed, you know, they have some automated bits. But today, there's actually a fair amount of just humans doing work. Yeah, the humans may have some software help, but it's a lot of like, manual effort to get things right. And the opportunity in plugging in an AI model is, instead of, you know, landing something that's just kind of the raw information, like, really do a first draft with the model and see if you know from the examples you've got, can you produce something that's like, a pretty good guess of what's going to be the right outcome? And if you can do that, you can make the people who are working in that area much more effective, right? And so you can imagine that that's, it's very much a kind of streaming problem where, you know, there's, there's some new event that occurs in the business, which is like, a claim occurs, right? And then there's actually a fairly complicated set of things that have to happen and data that would have to be fed into the model for it to have context to do the right thing. And so the the opportunity for them is, take, gather all that real time context, buy the model, be able to produce an output that's ready and waiting for the human Yeah, to say, ‘Yeah, that's good or no, but at least it's a basis for me to get started with if I make these changes.’
Olivia Hack
How do partnerships play a role in supporting Confluence innovation efforts and what qualities do you tend to look for in a strategic partner to help drive your software roadmap forward?
Jay Kreps
Yeah, there's a number of ways, you know, we have kind of a platforming technology. So what tends to be most relevant to us is, first of all, what is it that we sit on top of, which is the cloud providers, right? So the partnership with each of the cloud providers, how we go to market with them, how we work with them, technically, you know, all the aspects of how we integrate, and that's beneficial on both sides, like we send data off to a lot of their services, right? The part of our value proposition is connecting up the things the company uses. In many cases, these are services cloud providers are providing, and we're driving consumption for them. The other important part of this is, how do customers actually take this technology and build it into useful applications themselves? So sometimes they'll do that on their own with their internal teams, but very often there's some SI or other partner that's helping to guide them or even develop things for them. It's very important that those partners understand our technology, understand how to work with it, and have a set of patterns and practices that lead to success. If we talk only to the customer and not to the SI’s that they're working with, we really haven't done our job. And then last but not least, you know, Confluent is ultimately kind of a network that connects lots of technologies that kind of ISV partners of things we connect into. That's really important, right? So the, MongoDB and Snowflakes or other things that might be destinations for data, those are quite important as well. The saying in the data world is that there's some concept of, like, data gravity. This gravitational pull that you’re trying to escape the data center, escape whatever. You can't overcome the pull of of this data gravity that kind of sucks everything back, back down to earth. And so, I guess part of what we're trying to help these companies do, maybe, is kind of fight gravity, and, you know, allow this data to flow out into the environments where they want to have it. They want to take advantage of it. So it shouldn't be the case that your your next application, your next analytics platform, is inherently tied to wherever the source of the data is. It should be kind of transparent.
Olivia Hack
So Confluent has been actively investing in growth and innovation through M&A. What specific areas of innovation are you prioritizing when it comes to M&A, And how do recent strategic acquisitions such as Immerok in 2023 and WarpStream in 2024 form enhance value to customers?
Jay Kreps
You know, as we were public, we felt like, okay, there is some opportunity here, and there's a rich ecosystem around us. We're still not, you know, looking to add larger chunks of business. I would view this as a way of making strategic product bets, which is not the only strategy in this space. I think, the most important thing in my mind is that kind of active engagement, not being too immersed, not drinking too much of your own Kool Aid, being too immersed in the ‘not invented here.’ And then, kind of starting with the premise that anything that has some amount of traction, there must be something good there. And I think that's a very important thing I've found with both competitors, partners, other ecosystem components. And, then once you kind of have that cadence of activity, then I think opportunities present themselves when they present themselves. And I do think that individual companies, it's never exactly what you would imagine doing, there's always the very particulars of the those people and that product and what they've built and their customers, and yet, that's still often much better than what you would end up with if were you to do it yourself and kind of take care of it that way. And so we're lucky enough to have a really rich ecosystem of smaller startups, open source technologies that kind of exists around streaming and around Confluent. So we of course, we pay attention to all of them. And that's an opportunity to learn ourselves. It's an opportunity to inform what we build. And then occasionally it's an opportunity for an acquisition. And, there's not really…I think those acquisition opportunities are very much on- off decisions. Like each one is actually unique. The case for each one is different. What you get out of it is different. But broadly, we would look for things that are, kind of fit with that kind of unified platform vision. Can kind of directly be added into what we have. You know, it's usually going to be things that are pretty early in their journey of monetization, but which we think can be very successful products. Right now, the kind of data infrastructure space has a lot of startups, I would say, mostly fairly premium valuations relative to the public markets. And so I would say it's a target rich environment. But it only really works for things that you have extremely high conviction on. I could see that changing in the years ahead, when maybe some of the private market prices come down. Sure, at that point, maybe it's easier either way. I think the decision criteria are largely the same, regardless of the environment, and I think that just the kind of situational awareness is a very healthy thing for companies, even if we make no acquisition decisions like an active engagement in that ecosystem and understanding it tends to be pretty essential for our success.
Olivia Hack
Do you have any advice for entrepreneurs looking to successfully integrate new initiatives?
Jay Kreps
Yeah, yeah. You know, I think it's a really hard thing to get a company of almost any size to do something new. There's a certain amount of ‘Hey, an object in motion tends to stay in motion.’ So, I do think the most important thing is kind of applying that outside force, like, continuously until it happens. You know, it's interesting… as we were undertaking our cloud offering, sure, it was very different from what we'd done before, in terms of how we would sell in terms of how we would build in really software. It was actually like, almost like a different company. I do think that that commitment right is very important, because, for optional things, they probably won't happen. There's a fair amount of inertia or momentum or whatever, of the core business, anything that's kind of nice to have, tends to fall away. And there's some reason. And the more complicated and larger the company is, the more kind of veto points there are, the more teams that have to contribute their bit for anything to succeed, of course, and so I do think that it's very important for the management team to be on the same page, right, and just, you know, be extremely obstinate about it, yeah. And to only undertake this kind of thing if it's really something that they feel that level of commitment for, right? I think where often companies go wrong is there's, you know, there's sort of a moonshot approach to innovation, which is like, ‘Hey, you know, 1000 flowers will bloom, and we'll see what happens.’ Yeah. And mostly, they don't bloom, you know, you plant them, but nothing really happens. And so I think it's, you know, it's important to kick around the new ideas, but, but I think it does require a fair amount of top down buy in to take the thing that you think is critical to the next wave of the company, and then just kind of push it all the way through.
Olivia Hack
All right, fantastic. I'd like to hear about how your experience as an engineer working at Kafka helped inform your strategic leadership as Confluent’s CEO?
Jay Kreps
Yeah, it's, it's interesting, like in many there's a lot of tech CEOs who are former engineers, in many ways, it's like the least relevant possible training for what a CEO actually does, which is a fair amount of communication and a broad overview of different parts of the business, yeah? I would say the one thing that is helpful is, I do think software engineers, probably engineers broadly, but software engineers in particular, you do tend to get trained in a certain type of system thinking, And it turns out that works for software modules, it also works for parts of a company. Software engineering is a very dynamic field, yeah? And so it's kind of expected that you are continuously learning stuff. I do think that that's another carryover for CEOs, where you're, CEOs are kind of perpetually unqualified for everything they do. You kind of sit atop whatever it is seven or 10 different functions. Probably somewhere between six and nine of those functions you never yourself, had depth in running and yet, it's very important to know What makes a good sales leader? What makes a good sales effort? Where is your sales team falling short? What makes a good product leader? What makes a good product effort? So, the process of learning about that is probably a little different learning software engineering. You don't just go read a bunch of books, yeah? But nonetheless, that idea that you're going to just your job is going to be this, never really knowing enough and always trying to learn. I think that's probably a useful mentality, probably in all parts of business, but particularly for CEOs, because you have this kind of untenable breadth that you're never really going to be able to learn everything about everything about.
Olivia Hack
Yeah. What major challenges are facing software founders and CEOs over the next five years, and what advice do you have for them?
Jay Kreps
Yeah, it's a good question. I mean, it's an interesting time right now, because we're in a market that's been relatively tight on software spend. So, and we're coming out of a time period that was quite loose in terms of software spend. And so companies that grew up in kind of a, call it a loose money era, both in terms of funding, but also in terms of buying. The company is tuned a certain way. And so the difficulty over, the last whatever call it two, three years, has been kind of retooled, right? You can call it getting fit, or whatever you want to call it, it's a pretty significant reorganization of the company. So like in the midst of everything else we were doing, Confluent improved operating margins by about 40 points, which is like, actually a lot. And the amount of effort and attention that goes into that is very high. So then the challenge is, in a pure contractionary time period, that's all you got to do is get efficient, yeah? But we're not in a pure contractionary time period. There's also a fair amount of innovation, and there's new things happening around AI, it's a very dynamic landscape. So you're kind of both need to become much more efficient, but also completely orient yourself to capture the new opportunity. And so I think, it's a very dynamic industry. And so I think the challenge for anybody leading a tech company today is, how do you navigate through that, still attached to the big vision you had, but in a much harder environment, both on the buying side and how you're deploying capital, that's the dilemma.
Olivia Hack
Fantastic, well thank you so much for your time.