Confluent, Inc. (CFLT) Earnings Call Transcript & Summary

All right. Good morning, everyone. My name is Phil Winslow, Credit Suisse software analyst here. Very excited to have one of my favorite executives joining me, a longtime friend, Steffan Tomlinson, CFO of Confluent. So Steffan, thanks for coming down. It's great to see you in person.

Yes. Great to see you in person. Thank you for having me.

Yes. Great. So obviously, Confluent is a relatively new company to the public market. So I'm saying we'd start at a high level here. Maybe give us an overview of the business, obviously, the prevalent theme and just simply data in motion, the origins of Kafka and how the company is actually monetizing that.

Well, for starters, the origin really goes back about a decade when our 3 founders, Jay, Jun and Neha, were engineers at LinkedIn. And they were struggling with connecting applications and data stores to run their business in real time, and there was no commercially viable solution or even open source solution that was out there. And they ended up developing what ultimately became Kafka while at LinkedIn and to really process what amounted to billions of events per day. It's massive scale. And there was so much benefit that LinkedIn was getting from this and what Jay and Jun and Neha developed that they ended up releasing Kafka to the open source community. And the realization and the light bulb went off that hundreds and now hundreds of thousands of companies were struggling with the same dynamic where any of the legacy technologies that were out there weren't really delivering real-time data and motion characteristics. And what Kafka was able to do was really unleash that, the power of data in motion. And as the 3 founders took a look at the widespread adoption of Kafka, they decided to leave LinkedIn and create Confluent. And we've since commercialized Kafka and have built many proprietary features on top of Kafka to really give like enterprise-class scalability, reliability, security. And we have this concept of -- 3 concepts of cloud native, complete and everywhere. And these are differentiating pillars for us, and it's really driven remarkable growth in the business.

Yes. Well, those are part of my 3 pillars, too, of my thesis. So I like that we're synced up. But one of the things I always say is like fast data beats slow data any time.

And so when you think about -- to put a finer point here on sort of just the use cases of data in motion, what are the common ones that you hear about -- frankly, I actually thought your conference call was one of the best conference calls I've listened to. Wholly -- I advise listening to that just for some of those use cases, but maybe walk through some of them with us.

Yes. So the use cases are literally just boundless when you think about the horizontal applicability of the technology. But to put a finer point on it, if you look at the retail space, as an example, when you connect applications and data and run your business in real time, you have much better inventory management. You can do customer sentiment analysis. You have efficient both back-end and front-end operations. And you can deliver tailored solutions to your customer base. And that -- those are areas in the retail space that have a high level of resonance. In the financial vertical, early fraud detection, capital flows, market analysis, sensitivity analysis, retail and investment banking experiences are driven much better with a real-time data in motion framework.

Yes. One of the things I talked about with you all, and I think actually Jay used the term is just white space. When I think about Confluent, to your point, there are so many things that -- it's not that we're replacing some, we just didn't do this. So it's just all the applicability of Confluent just to actually deliver solutions that, frankly, you couldn't before. And so actually -- just to a finer point, and sort of like I always say, like, hey, if you give developers a platform, they'll use it to do new things. So when you think about -- from a developer perspective, what are some of the exciting use cases, too, that you're seeing today?

Well, first off, by adopting Confluent as a standard within a company's organization, it really frees up developer cycles to do the creative and inspirational things that developers like to do. For example, with other competing technologies or even open source Kafka, there is a high level of administration and complexity. And with our Confluent Cloud solution, with it being fully hosted and managed, it removes the cycles of administration. And when you combine that with the fact that we have over 120 prebuilt connectors, that enables folks to really write applications in a meaningful way that helps streamline the business. And when you also think about the ksqlDB offering that we have, that really marries data in motion and data at rest, and it does it in a developer language where they can use the SQL programming language that they already know. And so then they write applications. And I'll give you one example of something that Confluent would have never come up with on its own, but it was very creative. So there was a large banking institution that was running Salesforce.com, ZoomInfo and Confluent, and one of the biggest pain points for that institution were the employees having to automatically or manually update Salesforce.com, the contact graph, et cetera. So they ended up basically integrating Confluent, ZoomInfo and Salesforce.com, and they have made that updating of the contact graph in real time. And that is something that we would have never productized on our own, but it has revolutionized the day-to-day operations of that company.

I like that example too because those are 3 of my outperforms. So I like this. The -- but let's talk about TAM because, to your kind of point, it sort of -- I feel like your TAM is only limited by sort of developer's imagination. And that's usually a good problem to have. So how do you describe the TAM to investors because there is so much white space?

Well, we're a category-creating company. So there's no third-party resource that we just point to. So we had to go through an analysis of constructing the TAM, and we really looked at where budget dollars are coming from. And there are 4 categories that are defined by Gartner that we leverage. So app infrastructure and middleware, database, BI and data integration and data quality tools. Those 4 in total is about $150 billion market opportunity. Given our product feature set today, we address about $49 billion of that today. And when we evaluate our product road map over the next 3, 4 years, we see that -- our serviceable market growing from about $49 billion to about $90 billion.

And then I guess, how do you think about this, because this is a question I often get from investors since we launched this. Is Confluent sort of displacing another vendor or sort of taking share of that budget on from somebody else? Or is this just, call it, evangelical sort of sale and new spending?

It's really a combination of both. One of the powers of the model that we have at Confluent is this product-led adoption. So Kafka is an open-source software that anyone can download for free. It's a developer-led initiation point, an entry point into an organization. And what happens with that is the developers start using it. And when it moves from a single use case to multiple use cases or if it moves from a project to more of a platform, that's where an entry point for Confluent comes into play. And when you have those dynamics in the marketplace, it becomes a very -- kind of there's a viral sale to it. However, in order to go from really a project to a platform, you need to have the enterprise-class selling and go-to-market motion, which is critical. So we're doing both developer-led and product-led as well as the enterprise-class go-to-market because it's both a top-down and bottom-up sale.

Bottom-up sale, makes sense. The -- let's talk about Confluent Cloud for a little bit. Obviously, that's a big part of our thesis. And obviously, you've seen phenomenal growth the past few quarters in particular. I think it's 4 consecutive quarters of accelerating growth actually in Confluent Cloud. But the question similarly that I get from investors is how do you differentiate Confluent Cloud from some of the large hyperscalers out there that might have a competing solution? How do you differentiate? Why are customers choosing Confluent Cloud?

Well, confluent Cloud is -- it's one of the best-selling products that we have, right? Last quarter, it grew 246% year-over-year. It was accelerating growth. And so as we look at the overall ecosystem of partners that we have like at play, the cloud service providers are critical partners for us. And there's a level of coopetition with each of them. On the one hand, they absolutely love us because we're driving hundreds of petabytes of workloads across those environments. And they do have some competing products. AWS is the most pronounced. They have MSK and Kinesis, and Google Cloud and Azure have competing products. But I can tell you that all 3 CSPs incentivize their sales forces to sell Confluent because we're driving more workloads to their infrastructure. And from a product differentiation standpoint, none of the 3 CSPs have what we have relative to a cloud-native solution that's complete and that's everywhere. And just hitting briefly on those 3 points, on the cloud-native solution, we've rearchitected Kafka and put it into Confluent Cloud. And then we've added proprietary feature and functionality on it. The barrier to entry to do that is actually very high, and we are on all 3 of the main public cloud companies. On the completeness of offering, we have over 120 prebuilt connectors. We have ksqlDB. We have stream governance. We have a number of other kind of interoperability functionality that we've rolled out. And then on the everywhere side of the house, with cluster linking that we just announced, people -- developers are able to seamlessly link data sets and applications across multiple environments in a couple of clicks. And the ease of use and the ability to have a transparent infrastructure with data governance is a very powerful solution to companies who are running very complex systems.

Yes. No, exactly. Yes. I mean the thing I've been telling people too is that, okay, the Kafka API is not Kafka. It's like -- and Confluent Cloud, to your point, is the only cloud-native Kafka offering out there that's generally been rearchitected. Sometimes people just kind of miss sort of that kind of distinction there. But I describe it as if everybody supports Kafka API, it means Kafka is winning. And the only cloud-native Kafka offering is Confluent Cloud.

Confluent Cloud, exactly.

So the -- but let's stick with the idea of open source and differentiating Confluent Cloud because there's another question that I get not just for Confluent, but frankly, any company that has open source at the core. But it's like how do you sort of balance, I guess, the idea of fostering that open source community with contributions but also holding back some features, to your point, for the proprietary offering.

It's a really good question. And when we think about the care and feeding of that -- of the open source Kafka community, we take it very seriously. The reason why Confluent is so successful is the fact that Kafka has become the de facto standard for streaming. Hundreds of thousands of organizations use it. And so we want to not only foster that community, we want to nurture it. Now there are proprietary items around whether or not it's ease of use like the 120 prebuilt connectors that's proprietary. Cluster linking, which I mentioned before, that's also proprietary. And we have a number of other items that are proprietary in nature. And there is a determination that we make as a company around what we donate to the open source community versus what we build for a commercialized business. And we've -- I think we've struck like the right balance there over time.

And then actually, let's drill in this a little bit because, obviously, there's pure open source Kafka, kind of DIY, so to speak. And then obviously, there's Confluent, call it, on-prem, a Confluent Cloud. When -- is there a sort of a tipping point, I guess, that you see in terms of whether it be complexity or just scale, where, hey, the person that was trying to DIY it says, "Look, okay, I need some sort of commercial offering, whether it be deployed on-prem or in the cloud?"

Sure. So it starts with segmentation. For very large companies that are out there with the capacity to hire lots of engineers, the DIY approach historically, really before Confluent Cloud came about, was kind of a preferred choice for lots of the -- just very large companies that have big engineering teams. What happens though is the level of complexity when you add multiple use cases, when you add multiple geos, et cetera, it becomes unwieldy. And they're actually -- like if you look at the population set of companies that are out there, there are very few companies that can actually do a DIY approach in perpetuity. So that's one thing. Now when you look at like outside of the very large companies, if we look at large companies and also midsized companies, they don't have the time, inclination or money to hire an engineering staff to do DIY. And prior to Confluent Cloud's launch back in 2018, they were kind of left with either try to do a half-baked solution or not do it at all. And with Confluent Cloud, there's such a great product market fit because it's a fully managed hosted solution that companies can use and not have to have a full-blown engineering staff that is doing the updating, care and feeding of the system. And so over time, we believe that -- even the very large companies that have dedicated engineering staffs for a DIY approach, over time, they were going to be migrating to Confluent Cloud as well because the ease of use, the administration, the attributes that we bring to the table on being complete, cloud-native and everywhere, those are powerful things that most companies can't develop on their own. And so we see an opportunity for companies who have done a DIY approach to start migrating over to Confluent Cloud. We make it super easy. People can swipe with a credit card or do a free trial for Confluent Cloud just to get started. And then we're going to see -- what we're planning on was we'll see better adoption of Confluent Cloud even in those very large accounts that had previously been Kafka DIY.

And I guess, what I'm describing is sort of -- has something changed? Because to your point, obviously, phenomenal growth in the past multiple quarters. I mean it's sort of like you're hitting that inflection point with Confluent Cloud. Was it just that, to your point, obviously, product market fit? Was it just it had been out there for sort of like -- since 2018 so you'd had enough people sort of kicking the tires, so to speak, and then you had sort of Confluent progressing -- I'm sorry, Kafka progressing inside? Was it just sort of like all these things come together? Or like help me kind of explain this.

Yes. Yes. So it's a number of items. The first is the product has evolved since, obviously, 2018 when we launched it. And there have been lots of incremental feature functionality that we've added: security; governance; as I've mentioned before, cluster linking. And so the richness of the product has improved a lot since 2018. On the go-to-market side of the house, prior to about maybe 15 months ago, we didn't even have a head of commercial sales. So in the commercial segment, which we define as companies that are less than $1 billion in revenue, that is the perfect opportunity for us to leverage Confluent Cloud because -- for all the dynamics I discussed before, those customers don't have time or inclination to build their own engineering staff. So we've dedicated a team to do the commercial business. We've seen tremendous growth there. And then in the enterprise business, which are companies with greater than $1 billion in revenue, our sales team and marketing organization has had a focus on a cloud-first, not cloud-only, a cloud-first approach to our selling motion. And then when you have the wrapper of what Stephanie Buscemi has been doing from a Chief Marketing Officer standpoint around having this horizontal applicability messaging for Confluent and these marketing personas that we're going after relative to CxOs, also down to the data engineer. So we have a full-court press going on, and that's starting to pay off. So yes, all of those things. And I'd just say, like the last thing is with Erica Schultz at the helm of field operations, she's our President of Field Ops.

She's great, yes.

She and Larry Shurtz have done a remarkable job of building out a world-class, enterprise-class selling motion.

Yes. That's great. Yes. The -- now let's spend a little bit of time on Confluent Cloud versus Confluent Platform. And do you see companies that are using both. Okay -- and if so, do you see that staying that way? Or is it just sort of over time, people will migrate more to Confluent Cloud?

For companies that had started out using either Kafka or Confluent Platform, we see the dynamic of those customers using both Confluent Platform and Confluent Cloud. And the primary reason is in order for us to be delivering the most effective solution, we've -- we coined this term we are the central nervous system. You can't be a central nervous system to half the body. And so people have a lot of on-prem data. Also, this is not breaking any new news here, massive workloads are shifting to the cloud. And that's a secular trend that's going to go on for decades to come, right? And so customers have this hybrid environment where they have data in the cloud, data in on-prem. And so customers need to have both Confluent Platform and Confluent Cloud. I think over time, Confluent Cloud will be the dominant part of our business, but it's going to take time to fulfill that vision.

Yes. That makes sense. Now let's talk about one of our favorite metrics, DBNRR. It's -- obviously, you've seen a massive improvement over the past few quarters in that. There are sort of different levers that could drive that. Is it gross churn, just net expansion? Kind of walk us through the dynamics of what's driving once again that inflection there.

Yes. So back in Q1 of this year, our dollar-based net retention rate was 117%. It was a big focal point for us as a management team, figuring out, hey, what's going on here? And we fully understood the drivers. What we did operationally over the last, call it, 6, 8 months has really been focused on improving our gross retention and improving gross expansion. And so when you look at renewal rates and gross retention, they're hitting our internal targets. And so how are we doing that? We have a great customer success organization that reports up through Erica. And we did a full-court press on making sure that we have the proper care and feeding of our accounts, that they understand the product road map. And we also had future and functionality improvements, specifically in Confluent Cloud, as I mentioned before. The profile of dollar-based net retention rate, we believe, will drive higher NRR for Confluent Cloud over time because it's elastic, it's scalable. It's almost frictionless in terms of adding more workloads. And so that thesis is starting to play out. Last quarter, we saw our dollar-based net retention rate for Confluent Cloud be higher than Confluent Platform. They're both in good shape. But we think that over time, we should see higher NRR for Confluent Cloud.

Yes. I mean if -- my line I use in software, if you make something easier to consume, people are going to consume more of it. And so that's why I think of that sort of Confluent Cloud. Just take away all the administration of the clusters. Don't worry about that, just focus on the value. Don't focus on cluster management. It's -- let us do that. And so the -- I'm going to ask a couple more questions. I'm going to pause to open up to the audience. But let's talk about just go-to-market strategy. You touched on it a little bit. Obviously, there's a developer-led focus, but there's also the tops-down sales motion. How do you see this evolving over time?

Yes. We've been historically a developer-led organization. And over the last couple of years, we've really introduced enterprise-class go-to-market innovations that we've done, and that's led by, again, Erica and Stephanie. We have to continue to nurture the developer community. That is incredibly important for us, and we're never going to lose that focus. But in order to get true growth and scale across all of our environments in our customers' environments, we need to be at the CxO level as well and in areas in between, which is why I was talking about this persona-based marketing. So in the short term, that means an incremental like level of investment that we've shared with investors at the time of our IPO, et cetera. I'm like, "Hey, we have an enterprise-class go-to-market model at the top and we also have a developer-led sale." Over time, we think that's a winning strategy because it's going to mean more wallet share for us over time. If you look at customer lifetime value, you look at other unit economics that we assess, LTV to CAC, et cetera, over time, we think the strategy will work. And we should be seeing more productivity per salesperson.

Yes. That makes sense. The -- one more question, I guess, and then I'm going to pause. But when you think about R&D, because obviously, you've -- as you said, sort of Confluent -- sorry, Kafka's winning -- Confluent's winning within that context. When you think about where you're allocating your R&D dollars to make sure that, that gap stays there, how should we think about what Confluent is focused on?

We're definitely focused on being cloud first, but we're not cloud only. So we're cloud first. That means proportionately more investment is being made in Confluent Cloud, new feature and functionality. And then over time, the new feature and functionality that gets released on Confluent Cloud will also -- most of those items will be ported over to Confluent Platform. But we're leading with Confluent Cloud as the innovation engine. And with that said, we're still doing the care and feeding in Confluent Platform.

Yes. Got it. All right. I've got a bunch more here that I can go through, but I'll just pause to see if anybody has any in the audience. Just raise your hand and I'll point at you, or I'll just keep going. Do we have a mic or just -- I'll repeat it. Yes?

I just had a quick one on domain-specific languages in terms of the go-to market and the design of the product and use of that. Can you talk about other models that don't align with the developer community with domain-specific languages? Just getting back to the comment of reducing friction [indiscernible]. Maybe just [indiscernible] Confluent's choice there and competitiveness [indiscernible].

I'll just let you answer. [indiscernible]

Yes. I want to make sure I got the question right. So Confluent's choice of domain-specific languages...

Yes. Confluent streams...

Confluent streams and...

Yes. Exactly. So when we look at overall Kafka and then the derivations of what happens like within Kafka, we take a very clear-eyed approach around what our customers need. And one of the things that we did was we created like ksqlDB as an example. And this was an innovation that was done by Confluent. And that connects effectively data in motion with data at rest, these 2 different paradigms, and it puts it in a system that engineers are able to use because it's based off of the SQL language. And they don't -- so engineers and developers, they don't need to learn a new language. But what they do need to do is understand how to write applications in ksqlDB to then transform their business so they can have real-time streaming done with both legacy applications and current applications. This is a nascent part of our business, but we think that it has high growth potential over time. So that's one example of our decision to use a programming language like SQL and marry it with Kafka effectively.

Can you talk to your vertical strategy? Any vertical in particular where you see the most white space?

Well, first off, we are a very horizontal -- like horizontally applicable across all verticals. Where we've seen a lot of early traction, and this goes back -- Confluent is about 7 years old. Retail, banking, automotive, tech, those verticals have been kind of early adopters of the technology. But what we've really seen is this horizontal applicability. Part of that is being driven by the ease of use with Confluent Cloud. Prior to the introduction of Confluent Cloud and the resonance that it's had in the community, you would have to have like a do-it-yourself engineering team at some of these large companies, which would have adopted Kafka originally, like those 4 verticals I mentioned, and then transfer over to Confluent. Well, with Confluent Cloud, literally any vertical can -- any customer, any vertical can swipe their credit card and get going with Confluent Cloud. So we don't have a vertical-specific go-to-market motion other than public sector. Public sector, we have -- and specifically the Fed, we have folks that are focused on that. But our -- we don't have a segmented sales force that is vertical.

Last question for me here in the last minute, 40. Obviously, we think Confluent has a very large opportunity in front of it. But you as a CFO and just as a company, how are you thinking about the balance between sort of investing for growth and margin expansion from here?

Yes. I think about it a lot. And given the $50 billion TAM that we have today, the product market fit, the traction we've had -- hey, we have to go and prove it every -- my mantra is every week, every month, every quarter, we have to go and prove it, right? So provided that we're proving it, right, we have a growth and profitability framework that we -- that we're managing growth and profitability, too. So effectively, what we've said is when we're in high-growth mode, meaning revenues growing greater than 30%, our focus is not going to be to maximize profitability. We're going to be delivering annual operating margin improvement. That's our plan starting in FY '22 and towards our midterm target of effectively 5% operating margin and 10% free cash flow margin. Our long-term target model is to get to about 20% to 25% operating margin and greater than 25% free cash flow margin. And so this is a top line growth kind of driven profitability model. As growth slows, we would plan to deliver more operating margin expansion towards the midterm and long-term target. If growth increases, then we also use that growth and profitability framework. But that's what we're using to guide our investment level.

I'm hoping that 25% free cash margin is like way out here, though. It means you're growing for a long time.

Yes.

Awesome. Well, we're out of time, my friend. That went very fast. Well, thank you for coming here, brother.

Hey, thank you so much. Yes, really appreciate it. Thank you.

Appreciate it. Well, thank you, everyone.