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

Thanks, everyone, for joining, particularly you, Rohan. Thanks very much for joining for day 2 of the Wells Fargo TMT Summit. I'm Michael Turrin. Pleased to be spending some time chatting through Confluent with you today. Hope you had a happy Thanksgiving. Thanks for making the trip down to the coast.

Absolutely. Thanks for having me. I'm looking forward to the conversation.

So this is -- I mean, there's a lot of different directions we could head. But I think just starting with the background on Confluent, the Kafka opportunity, it's still an area that I probably feel a significant majority of questions around is what exactly does Confluent do? How do I think about the market opportunity? Is there something they're replacing? So maybe just help frame the overall Confluent product vision and then we can dive into some of the metrics and other things.

Yes, we'll be happy to. We're living in a digital-first environment economy, and I'd like to say that every company today is a software company and every company is a data company. And how these companies harness data basically differentiates from success and failure for these companies. So that's the world that we live in. I'll just give you some examples. I mean, wake up in the morning, you only go to work, you need to call a Uber, you using a Rideshare app, that's real-time data. You're ordering a quick coffee, that's real-time data. You're swiping your credit card. The banks have tried to make sure there are no problem in transactions, that's real-time data. You're shopping in the afternoon during a lunch break, and you're trying to check out stuff, you get real-time marketing in it. You probably don't end up shopping, but then you get an e-mail saying that you have an unused cart, all of these are companies harnessing real-time data. It's like part of our fabric, part of how we live our life. So that's why we are playing a part. We're helping companies do this effectively. We're helping companies harness their data efficiently and we are helping companies harness their data in a cost-effective manner. So that's the broad framework. Jay, our CEO and Co-Founder, basically authored Apache Kafka about 13 years back, and the company is about a decade old. And in the last decade, we've kind of taken advantage of this opportunity with our Confluent platform on-prem product and our Confluent Cloud, which is a fully managed solution. And last quarter that we reported, which is Q3, we're roughly at $1 billion run rate for revenue, and we were profitable and we were generating free cash flow. So that's a little bit of overview.

That's great. Just 1 more on the market, and then I want to spend some time on Q3 and the most recent quarter. But you mentioned Jay started Apache Kafka and then I think realized there's a business opportunity behind it as well. So how does Confluent think about the base of open source Kafka that's out there? And if there are trigger points or ways to bring that base over to Confluent Managed Services. I know it's happened with some customers, but as you're looking at that as a potential play, how do you sort of think about that as a company?

Yes. Well, Kafka, from a technology perspective, as I said, is 13 years old. And we have over 150,000 organizations using open source Kafka. When I really think about our mission, our mission is to soak up the world's Kafka. And how are we doing it? So essentially, I think about like 3 pillars. Number one, when you really look at technology and technology differentiation, right? I think that is something that we started out with an on-prem product. But with any of these open source ecosystems, it's essentially a good managed solution that helps you take advantage of a bigger of part of this ecosystem. And with Confluent Cloud, we're continuing to differentiate where we are not only taking care of your infrastructure, we're taking care of expensive engineers, otherwise, you would be using. And we are taking care of other security needs that you have and providing it as a managed solution. So that's first like product differentiation. The second pillar is around just the completeness of the product and being truly multi-cloud. And with our data streaming platform, our vision is to move from a single product company to a platform and that continues to increase the differentiation that we have. So that's the second pillar. And the third pillar is all around ROI and TCO for the customers. And ultimately, what we are doing with our product strategy is essentially make sure that we have unique offerings for each of the use cases that customers might have. So we have a standard cluster. And the example I use is standard clusters is more like you're flying private. We have our enterprise cluster, where if you're okay with a multi-tenant offering, we have private networking, the right amount of security, you can use that. We have -- if you have large workloads, but your latency insensitive, you can use our freight clusters or you can use work stream. So our pricing strategy and packaging strategy is also part of taking advantage. So to summarize, it's obviously the cloud and product differentiation completeness of the product with our platform and pricing and packaging, which provides you with ROI.

That's great. Maybe from your perspective, the most recent earnings report, third quarter, the highlights, the key takeaways, the conversation points you're having with investors and meetings since that print, how would you characterize it?

Yes. If I had to summarize, we had a very solid Q3 earnings print. We grew subscription revenue 27%. Our cloud business grew at 42%, roughly, give or take, at a $0.5 billion run rate, that's high growth at scale. And we did that while driving 10-plus point operating margin and free cash flow margin improvements. So overall, from a numbers perspective, it was a very solid quarter. A couple of other data points. We spoke about our digital native customer segment. And that customer segment, we saw stabilization in some of the consumption patterns. So that's something is important to call out because these are customers who are technology forward, and we saw stabilization in Q3. And the third point I'll make is, from the beginning of the year, we've been talking about our data streaming platform. And I said that there are typically like 3 aspects to it. The first aspect is you need to have a product and the technology out there, which over the first 9 months, we've made good progress. The second aspect is around adoption of the technology. And the third is monetization. So on the adoption side, we did see good momentum and some of our larger customers starting to adopt the platform. Obviously, a lot of wood to chop, but pleased with the progress we made in Q3.

Is there any observation around what drove the stabilization, the improvements? Do you have anything that you've observed and kind of slicing the data, whether it seems like we're just getting further into the year and spend is coming back a bit or certain use cases or sort of opportunities to leverage the Kafka streaming platform that are becoming more top of mind?

Yes. From the digital native segment, if I really look at it over a period of time, it's been a growth driver for us. And I've said this before, has the growth always been linear? Probably not. You see a little bit of the sawtooth pattern, right? And in Q3, we did see stabilization in consumption patterns, not only for existing use case, but we are also starting to see some newer use cases being adopted. So that is good. And the third thing that I mentioned in my prior comment was around adoption of DSP. While very early days, I think that's something that we are pleased with, adoption of DSP. And as we look ahead, we're fairly underpenetrated and rather we have an opportunity in the digital native segment. So it's going to still continue to be a big area of opportunity for us.

How do you think about macro sensitivity to either customers' willingness to expand into new production level use cases at the pace that they might have previously or just other factors in the context of guidance, right? You mentioned that maybe things aren't always linear. And I think we appreciate that consumption models just have more of a kind of real point-in-time element to how revenue is captured. So just as CFO, how do you kind of roll that up? And how would you characterize the way that you're kind of framing the guidance that is still out there and how you've sort of guided throughout the course of the year?

Yes. My -- just broad-brush how I approach it philosophically. The -- our guidance philosophy, we try not to forecast macro. That's not something that we try to do. We try to focus on things that are in our control. So the simple assumption that we use is we'll continue to assume macro as it is, so -- right? So no movement in either side. So that's a framework that we use, and we've used for a while now. So we've been very consistent with it. From an overall consumption business, when you look at consumption businesses, the way I think about it is there are probably 3 drivers of consumption, if you really think first principle. The first one is there is like more data flowing through an existing use case. And there could be seasonal patterns. So that's number one. Number 2 is around adding net new use cases, which candidly is probably a bigger opportunity. And third is probably more Confluent focused specific, which is we have our adoption of our DSP. And these are all in our control, mostly in our control. So that's how I look at it. And from an overall predictability, it's -- we've been doing this for 5 years now. So we're very focused on making sure that we're trying to have a proactive approach to forecasting the business and predicting the business.

Okay. That will make sense. Can we tie that into expansion rates what you've seen? And I think you've commented that the expectation is that expansion rates are settling, at least holding stable is the expectation into 4Q. Can we talk about the drivers of expansion, different customer types and those profiles, and how we should expect or if there's a target range for us to think about in terms of expansion over a longer period of time?

Yes. So for net retention rates, I've not commented beyond Q4. And unfortunately, I'm not going to break new ground here, but I'll provide some color commentary, right? And the color commentary is when I -- there are 2 ways as the -- just if you look at the product angle and how customers are looking at it, and then there is a simple math exercise, right? So from a product angle, it is -- we have our existing customers, and we're selling our data streaming platform to our existing customers. So our connectors, our governance product and Flink. And I've said that this year, we're really focused on the adoption of it, and we'll continue to be focused on adoption will be a focus next year as well. But we'll start to see monetization that's going to benefit expansion. So that's 1 area. And then the second aspect is, I say it's a math thing, but it's more around, when you look at our net retention rates, we have company net retention rates, we have cloud net retention rates and you have net retention rates for customers that have DSP. And in that order, cloud is higher than the average company and customers that use DSP is higher, albeit very small numbers. So in general, it kind of goes back to the same point that as we look ahead over the next few years, DSP selling more products into the existing customer base is going to be an opportunity to drive expansion.

Is there a framework that you could help us with in terms of the DSP adoption curve? Meaning is that something that could kick in when a company hits a certain scale of what they're using Kafka for? Is it -- I mean, is it governance connectors first and Flink second? Is there a sequencing to consider? Just any kind of early examples of companies you've seen -- customers you've seen take that journey and what that could look like as we're just kind of thinking through just your evolution from a core, strong foundational technology to becoming more of this data platform?

Yes. It's difficult to answer because the way the products have come out, I wouldn't be able to say that what's a more ideal path. We've had our connectors on the cloud for a while. Obviously, this year, we've had product innovations on the networking and security side, which has kind of harden the product quite a bit. So when I really look at the connectors, there are large independent companies monetizing connect. It's an opportunity for us. And when you look at our DSP revenue today, that's probably the majority contributor given purely from when the product was started, et cetera. Then you have governance and governance is one of our fastest-growing products. And it's important because it's essentially making sure the right people have access to the right data. And when you have data moving internationally, it just is a product that we've seen good adoption. And then you have Flink. Flink is primarily -- the way to think about it is it's enriching the data to make it more powerful. And when you think about our role in the ecosystem, we've spoken about the operational estate and the analytical estate. And we're really trying to be the connective tissue and ensuring that you have a seamless flow of data between the operational and analytical estate. And doing some of this processing and governance earlier in the data life cycle is basically ensuring that as the data is moving to the destination, it's of higher quality. And that's where our DSP actually comes in. And we've also early accessed a product called Table Flow, which is essentially organizing the data into tables. And we're -- there's 1 technology that's really kind of becoming the de facto, which is Apache Iceberg. And when you kind of combine these together, it's governance, processing and organizing this data, you're doing it earlier in the life cycle. So as the data flows to the destination, it's of higher quality and higher fidelity.

So that sounds like a point that we're hearing from a lot of enterprise software companies as a pain point towards getting to the promised land of generative AI-focused applications. We hear Salesforce, ServiceNow, various companies talk about just the complexity of data within enterprise and the need to kind of help streamline that. Is there a way to think about Confluent's role within that? Like what are you seeing? Obviously, still very early in terms of generative related use cases or just that as a catalyst for more customers to look at how the data is organized and think about leveraging Confluent to help with that?

The generative AI landscape or in general, the AI landscape, we tend to look at it in 2 lenses. The first lens is you have these large LLM providers, and we think about them as large infrastructure companies. OpenAI is a customer of ours, and they can be large independent customers. So that's 1 set of the opportunity. Candidly, the bigger opportunity is more around as mainstream adoption of GenAI happens, you're going to see more and more applications move to real time, and we expect to play an important role in this ecosystem. So like not getting into technology, very simple, you have these RAG pipelines, you have a vector database, you have these LLMs. In the structure and this setup, you need real-time organizational context. And that's where we expect to play. So in general, what you're going to see is, there is this discussion and debate around applications that are batch, applications that are real time. Generally, you're going to see a tailwind to more applications moving to real-time, right, which is going to benefit the broader ecosystem and Confluent in specific. And then you're also seeing net new applications getting built as a result of it. So yes, we expect to play an important role in the ecosystem.

That's great. It feels like a very important point. Just back to the metrics for a moment. And I bring this up in the context of your consumption-based business, the Cloud segment has been the fastest growing, but the stock has reacted to kind of small shifts in cloud revenue from time to time. Is cloud the right metric? And maybe walk through your moving to subscription and the view from a Confluent and from a customer perspective on is subscription a more holistic way to evaluate the growth of the company and just your thoughts on kind of changing the metric a touch in terms of how you're guiding and focusing investors primarily.

Yes, that's a great question, Michael. I mean one of the reasons we decided to change the guidance and how we look at the business from total revenue to subscription revenue, and we did that over a period of 12 months. So there's a big overlap. So that was not a surprise. But the reason we did it was because we feel that subscription revenue is a better gauge of the organic momentum of the business, right? And when you double click into it, you have a Confluent platform, which is a traditional license maintenance where approximately, give or take 20% of our revenue is recognized upfront as license revenue and the rest is maintenance over the period of the term of the contract. And then you have our Cloud business, which is where the rev rec is more consumption-based, right? So that's a broad structure. And of course, like we -- I like to say that Confluent needs to be wherever our customers' data and infrastructure reside. If it's on-prem, we need to be on-prem with Confluent platform. If it's in the cloud, we need to be there with Confluent Cloud. And we have customers that are -- have hybrid deployments, and we have customers that have multi-cloud. So that's the entire architecture. Like we need to be wherever our customers data is. So both will be important. I'll just finish off by saying, when you look at Confluent platform, historically, the broader drivers of that business has been the regulated industries and Confluent Cloud has been a lot more broad-based. And that will continue to be -- so long story short, both are very important part of our product portfolio.

Can you touch on where WarpStream and Tableflow fit within that conversation of kind of wanting to meet customers where they are and just what you saw from a product portfolio perspective that made those kind of important priority efforts on the product side?

Yes. I'll start with WarpStream first. Really, going back to my comment around Confluent needs to be wherever our customers data and infrastructure reside. We are playing on-prem, playing in the cloud. And over the last 12-odd months, we were seeing this BYOC category develop where essentially, you want a cloud product, but you want to do it in your infrastructure and not the vendor's infrastructure. And that was an opportunity. And if you go back to my prior comment, we have 150,000 organizations using open source Kafka. And we need -- from a pricing and packaging perspective, we need to make sure that we have an option for you what you want to -- how you want to deploy a product. So WarpStream kind of plays into that. But our M&A philosophy has been always around really high-quality technology and really best-in-class teams, right? So obviously, not groundbreaking, but that's our policy. From a technology perspective, WarpStream is what we call BYOC native. A lot of BYOC technologies where you might still have to get into customers' infrastructure and what we call a break the class. WarpStream does not do that. So from a security perspective, customers love it. So that's why it kind of provides that optionality for our customers if you want large logging, observability or you're moving data to your data lakes, those kind of use cases, which are relaxed latency, and you want to do it in your own infrastructure, WarpStream is a great solution and will provide you with the right ROI. If you want to do it in our infrastructure, we have a product called Freight, we can do it. So it just adding to optionality. Part 2 of your question on Tableflow. I think Tableflow is, again, goes back to -- for most of these analytical use cases, you need to organize the data in the form of tables. Apache Iceberg was becoming the de facto standard. So we kind of aid Tableflow. Ultimately, the vision is, we can kind of enrich the data earlier in the life cycle. And that's the play we spoke about it. So I won't repeat.

Yes. Great. I'm going to pause for a second. We have a good size group in the room and just see if anyone has a question they'd like to loop into the conversation. I have like 15 more that we won't get to in 10 minutes. But if anyone has a question, feel free to raise your hand. I think we have mic runners if we need them. In the interim, I want to go to just some of the go-to-market changes that -- I mean transformation is to be a word, the evolution of your go-to-market, right? We've seen other consumption-focused models make similar changes. But maybe level set the changes that were made in terms of incentivizing go-to-market and where you say we are in terms of progress there? If there's anything to be mindful of just in the 4Q time frame still around that transition or just any other pieces of that, that you would highlight?

Yes. Our go-to-market changes were implemented Jan 1. So what were them? Like Jan 1 of this year. So we've been 3 quarters into it. So there -- I'll say there are probably 3 categories. One is kind of ongoing and 2 that we've implemented. The first one is around the incentive structure. We changed our incentive structure from comping our go-to-market teams on Commit or ACV for the cloud business. We moved from Commit ACV to consumption, and that was 1 change. Like if you think about it, that's just a compensation change. But the second derivative implication of that is actually changes how we run the business. The example that I'd like to give is the definition of a pipeline last year was bookings ACV. The definition of a pipeline this year is the use case that we are trying to build, right? So what also happened parallelly that went live beginning of the year was we built our in-house systems, operating rhythm. So we're 9 months into both these changes, and we feel good where we are. There are always tweaks that you make, but just the broad set of changes are behind us, right? Okay. The third aspect of our go-to-market change, we spoke about us making this shift from a single product company to a platform. And there are, again, 3 aspects to it, I mentioned, there's a technology product aspect, building a product that's integrated. And then the next 2 are essentially how as a go-to-market team, you're selling a platform. That's new muscles that you are developing from a single product to a platform. And that's the journey that we are in right now. So that's probably the broad set of go-to-market changes that -- and 9 months where we are.

Does that change at all the type of sales rep you look to bring on in the future? Does that change the structure? Like how natural is that move from more of the single product to multiproduct. The consumption piece, I think, makes sense. It's very aligned with how you should be driving customers and kind of nudging them towards expanding their use cases. But on the journey from Streaming Kafka to the sort of the broader platform, what are the lessons learned? And does it change the structure or the attributes of a rep that you're looking to bring on board in the future?

Yes. You mentioned 2 aspects of it. First is just focusing on the consumption side of it. And we've been doing that for 5 years now, selling cloud. So not a whole lot of change. Yes, how am I getting incentivized? That's a change. So on that front, I just want to make a comment that last year, of course, in an ACV-based model, as Confluent, we are incentivized to kind of close the largest deal, right, versus this year, that's not what we want to do. We want to make sure that we're unlocking the net new use cases. The commitment is still going to be there, don't get me wrong, and that's going to be more of a pull from the customer versus a push from Confluent. So that's the shift that we've seen. And in general, when you look at it, that kind of takes away some friction from the system. The second aspect that you called out selling multiproduct, I wouldn't say there's like a new set of different skill sets that you need to bring in the same skill set, it's essentially articulating a more complete story around data and the journey that we are in. And we've always had this structure where there are sales engineers who are part of the selling process. And they're a very important part of the selling process, and they'll continue to be. So as a combination, I think it doesn't imply that there are going to be changes or different skill sets we need, just building the muscle.

Is that evolution from to the multiproduct platform, does that change the competitive landscape either in terms of you're hitting different areas of the data stack than you were previously or just reinforcing position within the core and kind of flexing your muscle and just having more to offer than other solutions that might be in the market?

From a competitive landscape perspective, I think the short answer is, of course, the platform strengthens it with the why, what's the why, right? So you look at our competitive landscape/opportunity set, again, like 3 lenses, okay? First is we have 150,000-plus organizations using open source Kafka with differentiation in your product with building a platform, you're building a more complete product, and you're continuing to add to the technology differentiation with open source. So that's good. That's good. And when you combine that with the pricing and packaging optionality, it's getting to a spot that we like. The second aspect is if you look at our competitive landscape, we have this dynamic, I'd like to say as competition with hyperscalers. And when you look at that, it's primarily we're competing with them, but we're also partnering with them very closely. We have go-to-market partnerships. We have technology partnerships with all of them. So from that landscape, again, building a more complete platform, typically makes it a lot more stickier. So it can only get better. And the third aspect is like a catch-all category. You have these small startups. You have more legacy players. And again, with the platform, you continue to differentiate. And specifically on the start-ups, I called out in my earnings call prepared remarks that we've seen really good traction in Q3. We have won more than 90% and that's something that we wanted to call out as well that we're seeing good momentum.

No, that's great. I think we have time for just 1 more. So I'm just going to turn it over to you for closing thoughts. I think investors are starting to think about next year, you're probably planning for the next 3 years. But as we're continuing to have these conversations, how should investors expect Confluent to evolve? What are the key priorities that you as CFO are focused on?

I'll go to my mental framework which is the 3 Ts, I call it, like TAM, technology and team. And from a TAM perspective, we can argue like what the number is. But ultimately, when you really look at the data landscape, it's getting complex. It's very, very, very important for companies, and we're playing an important role. So there's a problem, and the problem will get more complex with years to come. So that's probably the broad set when you look ahead. On the technology side, we're transitioning from a single product company to a platform and 2024 was a big year where we kind of put some of the pieces in place. The journey of innovation will continue, but it's an important year. And then you have the team and the execution. Personally, if you ask me over the next couple of years, that's a big focus. How do you take advantage of this opportunity and execution as a focus. So that's where I leave it.

You've got plenty to keep busy with. Rohan, I appreciate you making time for us before the end of the year. Great to see you.

Absolutely. Always good to see you. Thank you.