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

[Technical Difficulty]

Good morning, everyone. [indiscernible] a handful of slides giving you an overview of what Confluent does, and then I'll turn it over to Jason for Q&A. So I'm Steffan Tomlinson, CFO. When we look at the lineage of Confluent, it really starts with, like, the 3 [ engineers ] that blends in what we're trying to [ do ]. So our [indiscernible] founders starting with [ JPac ] was financial [ advantage ] trying to run their business in real time. And what happened was there was no software that existed that enabled all the applications and databases of what -- really didn't have data to be exchanged across [indiscernible]. It became a real problem, and they were able to withstand the overall environment [indiscernible] software, so they had to bring a code [indiscernible] called Kafka. And Kafka was all [indiscernible] released to the wild in the [ legal resource ] foundation and became 1 of the most successful [indiscernible] products [indiscernible]. Over 150,000 organizations use Kafka to run businesses in real time. So what does that mean? That means that, historically, data has been at rest. It's been resting in databases. It's stored in applications, and what Kafka did was effectively enabled data to float [ entirely ] amongst and between databases, applications, multi-cloud environment, et cetera. So when we look at the pervasiveness of Kafka and the [ go down ] with release, over 75% of the Fortune 500 uses it. And there's been [ developer-led ] events over the last, call it, 8 or 9 years, where literally hundreds of thousands of developers across the world use Kafka to help power businesses in real time. So when the matters of [indiscernible], let's say, ended up leaving [indiscernible], they [ bring ] it to Confluent and they used their skills and expertise to build a really strong team and we commercialized Kafka as a complement and we also created a cloud-based version, which is a big differentiator. So we run both on prem and in the cloud. And if you just take a look at some of the momentum, we've been public for a couple of years now. We started a $60 million total [ investable ] market. Our last quarter results were very strong for -- basically had about a $300 million a year run rate -- I'm sorry, $700 million a year run rate. Our revenue last quarter was $174 million with 3% year-over-year. Our cloud revenue is [ about ] $300 million run rate, that is 89% year-over-year. And we've been balancing growth profitability and the improved operating margin by 18 points. And the [ saviness ] of our products, we're an infrastructure play. We sit in the next-gen technology stack with high gross retention [indiscernible]. This is not something that you deploy in [indiscernible]. This has the stickiness that [ now ], with the current 30% net retention. And we have over 1,075 customers, we have spent over [ 100 ] in ARR. So this is not a small-ticket item. This is kind of a place where people go and they spend a lot of money and they continue to spend. So once we land an account, the expansion opportunity is very, very large. So you may ask, what are the use cases? So is Confluent used for. Confluent is applicable across every industry in the world. And if you think about some of the things that we interact with on daily basis, you're in [indiscernible] analysis, there is fraud detection, there is inventory management, there's point-of-sale transactions, there's curbside delivery and pick up. [ Because ] anything that an enterprise needs to do to both run an efficient back end and provide a rich customer experience, all of this needs to be done in real time today, and this revolution that's happened across the data landscape is only increasing. Think about the last, call it, 8 or 9 years where AI had not been very pervasive. Now generative AI is going to be another next level up in terms of more data needing to be transferred in real time, and Confluent this drive these use cases across every industry. The core problem that I alluded to is historically, data has been at rest in databases and is slowly [indiscernible] and end of the day, [indiscernible] no longer will companies [indiscernible]. They just won't -- I mean, people need the real time to [ leverage each ] of their businesses. And the tool that existed in the past have been legacy tools, data point-to-point solutions, [indiscernible], data integration vendors. Often times when I speak with investors [indiscernible], this isn't just like a typical or [indiscernible] or [indiscernible]. The reality is those solutions never delivered on [indiscernible] promise of running a business [ single ] time, which is part of the [ founding ] or origin story, which I covered them in the source center. So where engineers were tasked with [ branded ] businesses in real time, the existing legacy solutions just didn't work. And to make it work, if you look at this figure here, with the legacy solutions, all these point-to-point environments created [indiscernible] where you had SaaS applications, databases, data stores, et cetera, all over the place with, like, there's no rational way to run an efficient network with this [indiscernible] here. And so what Confluent did was really come up with a [indiscernible] which we call data in motion. We're continuously processing streams of data in [indiscernible]. We're seeking these transaction key events that are happening, and we're letting them flow through and connect with other databases and applications and data source. And so what this results in, taking that giant [indiscernible] of an architecture, we're [indiscernible] nervous system where applications and databases can be run in real time. And think about the leverage that companies get by having the central nervous system as opposed to that spaghetti mess [indiscernible] to deliver these great use cases and real value to their customers. So how do we serve this market? We have a data streaming platform, [indiscernible], but [indiscernible] protocol. And we put in a platform where we can serve customers on-prem, in the cloud and multi [indiscernible] markets. So data does have [indiscernible] on the core technologists, but data is everywhere. It's on prem, it's in cloud and multi-platforms, so you have every product that serves your customers wherever their data resides. So one of the key differentiators around Confluent is that we work everywhere. And that [Technical Difficulty] and it's become a very powerful platform that [indiscernible] customer base. There are future growth drivers here. We just try to [ consume ] the space of streaming applications through an acquisition that we did [indiscernible] space. So if you think about Confluent [indiscernible] data connective tissue that's driving real-time data, application developers are going to need to be tapping to that real-time data and building these applications [indiscernible] with our [ plain ] offering that will come out next year. By the end of this year, [indiscernible] and really see a benefit next year. And then, of course, gen AI will also be a catalyst in terms of [indiscernible] data that will be [ processed ] [indiscernible]. So where are we with where we serve our customers, again, on-prem and in the cloud? Start with our cloud product, the Confluent Cloud, it can start with [ real sales ] that we [ addressed ] becomes developer-led. [Technical Difficulty] to go. And as a great onboarding mechanism, [indiscernible] platform is our on-prem solution and so we can serve [indiscernible] not. We are going across every industry as we just [indiscernible] of our customer's logos that literally every [indiscernible] benefited from data [indiscernible]. Looking at the financial slide, we've had significant revenue growth at scale. Since 2018, we [indiscernible] CAGR. We've also demonstrated our ability to grow revenues from quarterly [indiscernible] basis. Our cloud business is the fastest growing part of our business. It's a [Technical Difficulty] growth [indiscernible] continues to demonstrate. Once we land [indiscernible], the consumption-based model that we have for our cloud business is a winning formula for us. We've also focused on managing both the growth side of the equation and the profitability side of the equation. We have been improving profitability. This is just a way to show our progress here. You can see from an operating market standpoint, you go through [indiscernible] points year-over-year in the last quarter. We have a target of [indiscernible] margin breakeven in Q4 of this year. And then we've been driving efficiencies across the rest of the company within sales and marketing as a percentage of revenue, R&D and D&A, including [ productivity ] and price efficiency, so we're building efficient growth [indiscernible]. So with that, I'll turn it over to Jason for any questions.

Thank you, Steffan. There's a lot to think about here. [Technical Difficulty] for nontechnical investors, will you explain the concept of data streaming. What is actually data streaming?

Yes. So data streaming, effectively is every application that runs in the organization is [indiscernible] data. And the [indiscernible] data will be put into a database, and [indiscernible] data will be [indiscernible]. And what Confluent has done is basically, we have unlocked the stream of data so it's flowing freely. It's a continuous flow, and we separated the compute from storage. And what we're able to do is when you think about the central nervous system analogy that I gave before, you have applications and databases that are proliferated across an enterprise to [indiscernible]. And historically, they were never really connected. So now, Confluent is the connective tissue that enables streams of data to flow freely across and amongst [indiscernible]. So like why is that important? Well, it's important because the speed at which customers [indiscernible] businesses is only increasing. The telemetry that we're using to deliver rich [indiscernible] experiences to their customers and run efficient back end operations is only increasing as well. And I [ don't think ] that the question, how pervasive is streaming in a company? Like, does every application have to be streamed? And the answer is today, that's not the case, but it will be in the future. I mean you've seen that team in customer examples where they start with a [ core ] project that they need to have streaming applications, so think like the financial industry fraud detection. Fraud detection, you don't turn on and off. It has to [ happen ] the time. Well, that fraud detection engine also needs to be basically linked to other elements within a [ bank's ] infrastructure like private wealth management arm of the bank, et cetera. And so as one department or organization starts to run in real time, there is a network effect that happens and more applications in the database [indiscernible] in there. And ultimately this, from a financial standpoint, if you look at our retention rates, we look at the consumption model, we're seeing customer onboard more applications and databases into our data platform.

So it's almost like a -- instead of connecting things on a custom base, you're creating a standard denominator, technology at every [indiscernible] [ launching points ] in the system than [indiscernible]?

That's right. Yes.

All right. And then you talked about the customers [indiscernible], what does a typical customer journey look like? Is it like start with a specific case base [indiscernible] source, and then they [indiscernible] over to you? Just give us some [indiscernible]?

Yes. The way that we typically start with customers is they -- they have a need. The need for the customer is they need to start running their business in real time and they are trying to go with source [indiscernible] which, by the way, a lot of people are like, open source is free, so it must be better. That's free upfront. It's actually very expensive down the road. In fact, when we run a TCO and ROI calculation for 10x, their 10x [indiscernible] better TCO and ROI proposition than the open source version of it. But when companies typically start with is usually one use case. They will look at it for [ discrete ] projects. And once they run in Confluent industry projects, they understand the power that we bring to the table and they want to expand. But very critical [indiscernible], anytime you take one part of the business and then you start to migrate it out to other parts of the business. There's a mission criticality that happens in secured and [indiscernible] functionality is incredibly important. And the product suite that we have enables customers to, without a lot of heartbreak, go through a process of transforming from their legacy solutions to Confluent, and they get a security feature functionality and they get the performance, they got the TCO. And so it usually starts with a -- probably like one discrete project and then builds over time. We have this 5-stage customer growth [indiscernible] journey. Many customers start without ever talking to a salesperson. They'll go to our website, they'll go to Confluent Cloud, they could start to [ approve ]. And then once they see the power, they'll swipe the credit card to become a customer, then they will [ contribute ] to what we call [indiscernible] contracts so that they'll come to us and they'll say, hey, we started with a [indiscernible] to go format in Confluent Cloud. Now, we want to do something more significant, we want to be more pervasive. And the beauty of the cloud product that we have is scalable, it's elastic, it's a fully-managed [indiscernible] offering. You don't need as many very expensive personnel running an on-prem version. You can use the cloud-based version. So think of like the tailwinds that happened just across the industry that benefits lots of companies such as Confluent. But the move of workloads to the cloud, the security future functionality, Confluent Cloud [indiscernible] that in a very good way. And with the new product offerings that we're going to be coming out with specifically for streaming applications, [ like ] that will be a growth vector for us in 2024 and beyond.

You talked about the rationale for [indiscernible]. You touched on it a little bit in your slides. But what [Technical Difficulty]?

Yes. Okay. So if you think about where Confluent sits in the next-generation technology stack, we're an infrastructure player and we're enabling all your databases and applications to be connected in real time and [indiscernible], okay? So once all that's flowing, you can -- how do we tap into all of the data that's in there to build streaming applications? So the initial person we're selling to for Kafka and Confluent is engineering or content, et cetera. With all of that data that's flowing, the next person we'll be selling to is the developer, the developer who's developing applications that can tap into the real-time data flows. And in today's world, you speak to any developer, of course, [indiscernible] principle is they're developing applications with streaming in mind. That is the start of the point today. And so the streaming applications that are happening, that can deliver the point of sale that deliver curbside pickup, that deliver fraud detection, et cetera, all of those applications are being primarily written in a language called Flink. And Flink is an open source project that didn't really have [indiscernible], and Confluent bought a company called Immerock to tap into the Flink open source project. So now, we have Kafka and Confluent, which is powering real-time data in motion. And then we have Flink that is all about the application developer and developing real-time streaming applications. So that is additional sale that we'll be doing. It will be a cloud-based offering which we'll offer in our Confluent Cloud ecosystem, and it's a very -- it's the next evolution of our company.

I guess the customers, what kind of [indiscernible]?

Customers were absolutely, with all the indications that we have, they've been coming to us saying, hey, you can help us [ allot ] more value. We can do so by the [ end point ]. And some of the largest organizations in the world are using Flink to register the applications like, [indiscernible] many of the workspaces of the world. Entertainment companies, they're already using us, but they're using an on-prem version which is very costly and it's very hard to manage. And so we're going to be providing a cloud-based version which basically takes a lot of the hard work out of creating this. So we're going to be delivering a lot of value just like we did with Confluent Cloud.

Great. [Technical Difficulty] lowest and [Technical Difficulty]?

Yes. The competitive landscape is really -- it starts with the open source Kafka [indiscernible] that's out there for complement. So we had over [indiscernible] 150,000-plus organizations that run in Kafka for free. We have [ 4,690 ] paying customers, so there's a very large opportunity for us to build after that open source community and convert them from Kafka to Confluent. So that, by far and away, is the biggest competitor. There is no real direct one-for-one, like, competitor at scale against Confluent. There are cloud service providers that are out there that offer [ competing ] solutions. And we have what I call a [indiscernible] dynamic with AWS, Azure and GCP. They all are [ big ] partners of ours. They all have some form of competing products, but ultimately, those companies are [ agnostic ] around which solution they sell because they are very much focused on getting as many workloads to their infrastructure as possible. We have driven hundreds of petabytes of workloads to the cloud service provider infrastructure. The [indiscernible] of providers have their sales [indiscernible] quota when they sell Confluent's. We have strategic collaboration agreements with them. And we are -- [indiscernible], we're living and breathing data in motion, data streaming and streaming applications. And for the cloud service providers, they have lots of stuff going on and this is just one part of what they're trying to do.

So I have to ask the gen AI, the [indiscernible] of gen AI [Technical Difficulty]?

In the space of gen AI, how I think about it is all of the large language models that need to effectively be updated in real time in order to provide a really good customer experience will be a catalyst for increased data in motion and data streaming. And Confluent is very well positioned and situated to enable the overall ecosystem of companies out there to have their data in motion and data streaming. And for companies that are really embraced in gen AI and have either their proprietary large language models or for internet companies, if they're more open sourced, it's all about real-time data and that will -- improves our benefit.

A little bit up to [Technical Difficulty].

So on [indiscernible] AI, you've seen in the last 4 months the initial [Technical Difficulty]. Have you seen any news for products or [indiscernible] stand?

So I think the question was in the last 4 months, have we seen --

4, 5 months.

4, 5 months, have we seen any increase in gen AI usage in our customer base? Is that --

That's the question.

Yes. So the short answer is we have -- like what we've seen as customers -- existing customers of ours have been improving their chat box that leverage data in motion. So we have seen a large travel company that is a customer of ours that has been [ building ] us over time, and that's something that we will continue to see them like [indiscernible] data in motion. There has not been any sort of like [indiscernible] around usage at the moment, but that's something that will be -- we think will improve our benefit over time.

Second question. Do you think as the -- Do you think as the AI workloads become -- so right now, it's been -- it's taking a [indiscernible] public information. But as the workloads start -- the AI workloads start to dig into proprietary [indiscernible] that you mentioned, they [indiscernible] internal value [ basis ], and then more product will actually become more strategic. I mean, [Technical Difficulty] more inflection in a [indiscernible] standpoint? Or is that [indiscernible]?

Yes, that makes sense. AI and ML have been -- for use cases that have powered our business. And so with generative AI coming into the mix and the proprietary models that these will need to be updated, that should improve our benefit. And I will tell you, the way that Confluent's revenue model has worked in -- for the purchasing parity of our customers, this tends to be a slow build over time because applications and databases often have to be replatformed, reconfigured in order to benefit from our solution. So this is something we're like, in AI and ML and now, gen AI, will play out over time, and it should be a catalyst of growth for our business.

I think there's another question [indiscernible].

[indiscernible] cloud optimization [indiscernible] business for [ cloud ]?

Cloud optimization, we're less exposed to massive cloud optimization than other companies because we sit lower in the technology stack. Oftentimes, the workloads that are put through our platform [indiscernible] optimized. So when companies are looking to save cost, they're oftentimes changing their storage requirements or they're running less queries or they're changing the frequency around observability. Those things typically are part of upper stack than we are. This is not to say that [ it's ] fully immune to cloud optimization workloads, but we've been able to manage through it. So I'd say we have a little bit of exposure [indiscernible] that's how it [indiscernible].

[Technical Difficulty]

[Technical Difficulty]

Yes, it's a combination both. I'd say, first part will be training because it's helping get the data into the model itself. And then the influence side of the house, this is something where we believe that the acquisition that we did at [indiscernible] to get us [indiscernible], streaming applications will be part of the gen AI space. And it's from an influence standpoint, we will be playing that [indiscernible].

Last question.

Just a question on the sales and marketing at 56% of sales. When would you expect that to come down? And could you explain how that takes leverage over time?

Yes. So sales and marketing as a percentage of revenue has been coming down. It's still high, but it has been coming down. A year ago, it had [indiscernible]. And our mid-term target model is to get sales and marketing down into the low to [ mid-30s ]. And that's something that we're going to be driving through by increasing sales efficiency and productivity. We also have this land and expand motion. Once we win the customer acquisition, once we acquire the customer, there are [indiscernible] [ effects ] that enable us to sell incremental workloads on lower cost of sales. So when you look at some of the [indiscernible] path that we've shown in the past, the profitability of customers over time increases pretty dramatically.

All right. Thank you, everybody, for coming. Thank you, Steffan.

Thank you so much.