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

What a delight to see everybody. Welcome to the Goldman Sachs Communacopia and Technology Conference. This year, as I mentioned at the keynote, we brought these 2 marquee conferences together. And I believe I'm told that it's officially the largest conference Goldman Sachs has hosted. On that note, a real delight to be able to welcome our guest, Steffan Tomlinson, CFO of Confluent. I just noticed that there's another Steffan in the audience. I won't say who it is, but Steffan, welcome to the Goldman conference. Steffan, you've got an amazing background. You were at Google before joining Confluent. But for those that are not familiar with the Confluent story, given that you're relatively new IPO, can you -- relatively new IPO. Can you just recap the Confluent story at a high level, the key investment points? And then we can jump into some Q&A.

Yes, I'd be happy to. And thank you for hosting me. So big picture, data streaming is a necessary category, is a critical part of every company's infrastructure. And when you think about the digital transformation that's going on across every industry, every vertical, data streaming is one of the linchpins around how companies are running their businesses. The lineage of the company started out really at LinkedIn. And the 3 engineers who created Kafka were ultimately the founders of Confluent. And so as they released Kafka to the open source community, it became one of the best all-time open source projects. And it took off like wildfire. Hundreds of thousands of companies use Kafka. And when Jay and Neha and Jun, our 3 founders, created and founded Confluent, they really were tapping into something that was pervasive across every industry. And what is pervasive across every industry is every business is moving from batch processing to real time. Everything is being run in real-time operations. And it literally cuts across every vertical. So the big picture is it's a $50 billion TAM, which is growing dramatically over the next several years. We're just getting started. We've been a company for about 8 years. We're turning 8 years this month. And we have assembled a team, and we have a technological differentiation, and we have the go-to-market structure and initiative to really attack and address this market. And it is all about efficient growth on a go-forward basis. So we feel very good about our positioning. We partner with the 3 main cloud service providers that are out there. And when you're looking for the play in data streaming, Confluent is the play.

Really, really, really well said. And congratulations on the upcoming eighth birthday. The youngest company -- I don't know if there's a ranking, but to get to this revenue scale within 8 years is truly remarkable. And personally, it's been a joy for me to watch the company evolve. We've hosted Jay when he was a private company CEO. To see the story come along to the point that it's a massive TAM is truly gratifying. You mentioned Kafka. Where are we with respect to Kafka adoption and the translation from Kafka as a funnel for leads into commercially paid for offerings? Where are we in that transition, especially given the macro conditions that we're beset with today?

Yes. So Kafka, if you look at the stats through the open source foundation, literally hundreds of thousands of organizations are running Kafka. And one of the things that is very evident is, especially in this environment where cost is a real issue, inflation is a real issue, cost reduction is the antidote to inflation. And when you have Kafka, which is a do-it-yourself, very heavily intensive, lots of people need to be put against it, lots of developers, lots of infrastructure, lots of computing power, and when you have a do-it-your-own strategy in this environment, it's actually not scalable for multiple reasons. When you look at what Confluent's offering is, we have a fully managed cloud-hosted offering, which is what we call 10x better than Kafka. And while Kafka brings a lot of greatness to the table in terms of fundamental data streaming, it lacks the security feature and functionality, the data governance and many of the other features and functionality that companies need to have in order to run it robustly across their entire footprint. Our offering, Confluent Cloud, is the perfect -- has perfect product market fit to what customers are looking for. It's the highest growing part of our business, and we differentiate ourselves in terms of being more resilient, having better elasticity, better storage, better latency than primary Kafka. So Kafka is great technology, and we're going to continue to foster that. But we're seeing the movement from Kafka users to Confluent happen. And we're on a journey. Like we said earlier, we're going to be turning 8 years old this year. And this is a multiple year, call it, decades-long journey around building out a core data infrastructure level technology company.

Got it. And we were covering Red Hat. And when they said they're going to charge for open source, people -- some of you in the audience might, if I remember, covering, owning Red Hat. People felt commercialization of open source would be a nonstarter, but they went on to do it magically. And there was a real differentiation in the features that you're getting for the paid version versus the unharnessed potential of the open source version. So as you look at the macro environment, one might wrongly conclude that, hey, it's going to cause people to downshift a little bit. Let me just go to the open source thing. It's a little bit cheaper. Revisit the plans to -- what are customers saying in this macro environment, how they are justifying their conversion of Kafka to a commercial version with all the inflationary pressures? Wage pressures, everything that we all face.

Customers look at TCO and ROI. We skew very well on those metrics when we compare ourselves to a do-it-yourself Kafka implementation. We have a 250% ROI and the TCO is really [indiscernible]. So when companies do the analysis around having an engineering team that's dedicated to running Kafka, these are very expensive engineers. And why would you devote an engineering team to infrastructure-level technology when you can have a fully-managed hosted offering? So the TCO and ROI elements to our story are -- we put forward in this environment more so than ever. Also, when you think about all of the innovation that's going on, we have connectors that we're bringing out to the market that make it easy to use. We have recipes and use cases that we're rolling out to our customer base. This is all to remove the friction that it takes for do-it-yourself engineering organization to try to move from one use case to another. It becomes very cumbersome. So there's a slice of companies that are out there that we call large digital native companies that have large engineering teams doing this, right? But when you broaden the aperture and you look across Fortune 500, Global 2000, companies don't want to be investing in do-it-yourself Kafka. They just don't. And so when you have the math and the economics working out in Confluent Cloud's favor and when you have the overall efficiency and productivity and all of the things that drive top line growth, all of those elements drive to a decision that Confluent is the best offering that's out there.

Got it. Got it. You talked about lowering the paywalls on your earnings conference call. Can you just recap for us what was the objective of the company? And how is that playing out with respect to better close rate, conversion rates, better ASPs, whatnot?

Yes. So we have -- part of our offering for Confluent Cloud is customers can get started in a frictionless manner. They don't need to engage in a conversation with our sales organization. This is product-led. Companies can start out. And historically, what happened before we removed the paywall was a developer would have to swipe their credit card to get started. And then that's a barrier to start. So what we did was we removed the paywall and customers can -- developers can get started without swiping a credit card. And what we saw was a greater than 50% quarter-over-quarter growth in sign-ups from the developer community, and sign-ups is a great leading indicator of future customer adoption. So the desired outcome is when you remove the credit card swipe upfront, we're going to have more people adopting the technology on a trial basis. And when they ultimately get through a series of steps, when they ultimately swipe the credit card and become a pay-as-you-go customer, that will be indicative of someone who's really tried the product, who's seen the benefits of it. And we expect that customer, once they swipe, to grow and scale and ultimately matriculate from a pay-as-you-go customer to a committed contract. And those committed contracts and the conversion from pay-as-you-go to committed has been very robust. We've seen very strong matriculation there. And what does that mean over time for the business? It means that we are adding high-quality customers in an efficient manner. When we talk about one of the efficiencies that we're trying to drive is better sales and marketing as a percentage of revenue. Well, this product-led customer motion is one of the elements that we're focusing on. And over time, because it's consumption-oriented, we should be seeing less friction in terms of ongoing increases in consumption, use cases being adopted into the fold, and we'll see higher ASPs per customer on a go-forward basis. Some of the best metrics that we share with folks is the growth on $100,000-plus ARR customers and $1 million-plus ARR customers. Those have been growing very healthily. But I would still characterize our positioning in the growth trajectory as we're still in the early innings, if I were to use a baseball analogy. There's a lot of wood to chop in front of us.

Got it. Wood and baseball. [indiscernible]

Yes. Just a follow up on that. You mentioned the cycles for sign-ups, and that's a leading indicator. And what is a typical cycle for that? And especially in an environment where a lot of management teams, there's a lot more scrutiny and evaluation around any deals that are doing even in much larger software companies as well. So what does that typically look like? And how should we think about that when we look at Confluent?

By removing the paywall, we've actually removed another initial barrier for someone to even try the technology. And so people can try it for free. And so what does that mean in terms of a customer trying it for free and when do they become a paid customer? It typically takes a couple of quarters for us to really get material revenue from a pay-as-you-go customer. And now that we have third-party studies and we're leading with the TCO and ROI analysis in conjunction with the frictionless sales process upfront, we think that this is going to be a big growth driver for us going forward.

And another asset that you guys touched on and has definitely been a highlight in your report like over the last few quarters has been hybrid and the use of both your cloud platform and your Confluent Platform. So can you just touch on what those use cases are, how you're going about that opportunity and how you're really managing even the investments across both cloud and platform?

To level set folks, we have -- in products, we have Confluent Platform, which is an on-prem solution. We have Confluent Cloud, which is a cloud-hosted solution. Our hybrid customers are those customers that are running both Confluent Platform and Confluent Cloud. Confluent Platform customers are typically the biggest hybrid customers that we have. And what does that mean? Well, it means that some of the largest companies in the world in the financial services industry, retail, insurance, health care, they started out with Confluent Platform. Now as more of their workloads are becoming cloud-based, they're adding Confluent Cloud, and that's why we call it a hybrid environment because they're running both Platform and Cloud. The important distinction to be made is it's additive. Confluent Cloud is additive to the footprint here. So we're really not seeing very much, if any, cannibalization that's going on. And if you think about the multi-decade secular trend that's been happening, which is -- and everyone's heard this, everything is moving to the cloud. Well, the reality is there's still -- vast majority of workloads are still on-prem. And so there's going to be continued growth in the cloud business over time for every company, including Confluent. And when we tie everything back to the platform, our Confluent offering is a central nervous system. Platforms win over time. You cannot run a data in motion infrastructure and only have that running in one part of the body. So if you use the central nervous system metaphor, you can't have the central nervous system for like the -- one arm and not having it for the other. Like you have to have one central nervous system for the entire body. So Confluent Platform and Confluent Cloud allow customers to have access to our platform wherever their data resides. And I'll also tell you that the 3 elements and pillars that are incredibly important to our story is we're cloud native, which means everything is, first and foremost, done with a cloud development cycle. We're complete, meaning we have a complete product offering, security, data governance, all the traditional data and motion elements there, and we're everywhere, which means we're on-prem and cloud. And when you have that type of setup, customers are really embracing the overall value proposition that we're bringing to the table. And so our hybrid customers, we think, are going to be the highest growth customers for us going forward. And they also have the highest net retention rates of any cohort in the company's infrastructure.

So I had a follow-up question on that. I was telling my team as we were going through our prep for the conference session that Microsoft launched its hybrid offerings. They made a version of Azure that was on -- sorry, version of their on-prem Windows server that could sync up with Azure and there could be consistency. And upon that release, their on-premise products and cloud services took off because there was consistency in being able to point your workloads that we're running on-prem to the cloud vice versa. In that regard, how effective is the level of synchronization between a hybrids customers on-prem deployment and being able to seamlessly work with the Confluent Cloud? Are you working on that? Or it's already there?

Oh, it's already there.

Okay.

Yes. And if it wasn't there, we wouldn't have the growth that we've been posting, and we wouldn't have the net retention rates that we have, specifically for that cohort of customers that are hybrid in nature. So the real-time nature is part and parcel of the reason why we're here. We need to be able to provide real-time on-prem, in the cloud, in multi-cloud and hybrid cloud environments. I'll tell you, of the biggest barriers to entry for any company that is trying to get into the space is the amount of engineering work that we have invested in to make this a multi-cloud offering. When you think about thematically what's happening, and I came from Google Cloud, I was the CFO there. Every company is looking at multi-cloud offerings, and they don't want vendor lock-in. So the fact that we are multi-cloud in nature is a very big differentiator for us and it drives a lot of business. And because we're multi-cloud in nature, we have to be real-time everywhere.

Yes. Makes a lot of sense. I wanted to get your thoughts on the consumption model, which was loved on Wall Street when things were going up and reviled now that there is a distinct downtick in the economy. How -- you structured your contracts to -- is it really to wax and wane with the actual consumption trends? Or maybe your business is still so young, you're not feeling the effect of the vagaries of the consumption model. Where do you stand on that?

Well, it's important to note that Confluent is a hybrid revenue model. And we have a portion of our revenue that's upfront. We have a portion of our revenue that is ratable, and those 2 portions come from our Confluent Platform business. Our Confluent Cloud business is consumption. And Confluent Cloud right now represents roughly about 1/3 of our revenue, but it's the fastest growing. So how we think about this dynamic around the consumption model and what it looks like on a go-forward basis, you really have to start with the fundamental elements of our business model. This is not like an analytics play or a data warehousing play where people can turn things on and off pretty quickly. The workloads that are in use cases that are being funneled through our central nervous system, these are core ways that people are operating their business. These things don't get turned on and off. Think of retail. You think of real-time inventory management, you think about curbside pickup, personalized promotions. You just go down the list of use cases for retail. These are things that you're not turning on and off. These are workloads around how you run your business. Think about the finance industry. Credit card fraud, transaction monitoring, capital inflow management, real-time private wealth management. These are things that are running their businesses. These are workloads that you just don't turn on and off. So there's this other dynamic which you mentioned, which is we're still -- we're an 8-year-old company. We're building into this $50 billion TAM. Of course, there are going to be some perturbations up and down. But we feel really good about our positioning around driving incremental consumption on a go-forward basis.

Got it. This is a question that's just a natural segue from your comments and something I would normally not ask a CFO, but you sound very confident and competent that I think you could really take a crack at this. And then this question did come up in the days leading up to your IPO. And I think I asked Jay -- so real-time data streaming has been tried some 20, 25 years back. We've had legendary companies at one point. They were poised to really tip the balance in favor of real-time data architectures and then they went a different route. They went into private equity ownership whatnot. So what are the lessons you've absorbed but learning from prior attempts? And how do you feel about this time around real-time data streaming can be a legitimate platform? If you have views, and I'm sure you have views.

Yes. Yes. Well, I won't be sitting here if I didn't believe that Confluent was much different than the legacy companies that are out there. And it goes back to really the technical underpinnings of Confluent which the lineage, as I explained earlier today, is coming from Kafka. You look at Kafka, and Kafka is run by 70% of the Fortune 500. It's in hundreds of thousands of organizations. If the legacy technology could actually run real time, they would have done it. Kafka was born out of the fact that every other company that has tried to do real time were more point-to-point solutions or API-based solutions, and engineers and developers around the world were scratching their heads and were like, we need to run businesses in real time. There is no off-the-shelf software that works. And necessity is the mother of invention. Jay and Jun and Neha created Kafka to solve a developer problem. And I can just tell you, like when we go in front of customers, when we do bake-offs, none of the legacy players are there. They don't have a competing offering. So it's incumbent upon us to go out and win business every day, every week, every month and every quarter, deliver on the numbers. But the technological differentiation is night and day difference. And when you look at the overall thematics of what's going on, it's not -- there's no debate around, will streaming win this time? Streaming is out there. I think the discussion is, how big will the streaming category be? And when Confluent is leading the way, we need to go out and continue to show the innovation where we show the differentiation against Kafka and other companies and other solutions that are out there to take our fair share of the market and do it in an efficient manner.

Got it. I want to touch upon GTM. And I remember when we were doing due diligence in your company, customers were just lighting up, right? It was the most obvious thing. They were super excited about building new architectures. And I remember asking you, you don't even need a sales force. This product could just sell on its own, right? But you do have a fairly substantial GTM effort. Can you just level set us on what are the -- it looks like you're definitely building a GTM organization that can grow a very large company, multiples of your current size. How do you look at efficiencies in your GTM payback period, economics, et cetera? Because you have some very strong brand name type leaders, stock leaders in your GTM organization.

Yes. Yes. So there is a dynamic play where -- this, how do you get your foot in the door in a company? A lot of that heavy lifting has already been done by the universal adoption of Kafka. So that's actually a great starting point for the company. What the company realized, and candidly, this is even before I joined the company, that a purely developer-only led motion isn't going to really scale to get as much of the wallet share as they could. And why is that the case? Well, getting your foot in the door is one step and developer-led motion is incredibly important there. In order to move from a project-based approach to a more universal, like different type of architecture that needs persona-based marketing kind of up the management chain, there is an evangelism around the category like at a management level. And when we have architected our sales motion around not only devoting time to the developer, but also to the director level, senior director, CxO level across the org. And it's important for us to continue to do that and build that brand awareness of Confluent. Oftentimes, people use Kafka and Confluent interchangeably. And that's actually not a bad thing. But over time, we've created this differentiation. We've rolled out a 10x campaign that shows the differentiation between Confluent and Kafka. And then when you get to the efficiency side of the equation, we do skew high. Sales and marketing as a percentage of revenue is high, right? We've done that purposefully, and we are focused on looking at the unit economics of our business. So productivity per head count is one of the metrics we look at internally. That has been improving over time. For those of you that are familiar with the story, at the time of our IPO, we had mentioned that fiscal year 2021 was going to be a year of catch-up investments because we had hit the pause button on sales hiring in FY '20 when COVID hit and that in FY '22, the year that we're in right now, that we'd start to see the reps that we had hired in FY '21 become fully productive in FY '22 and beyond. And so we're going to have proportionally more fully-ramped, fully-tenured salespeople in FY '22 than FY '21 that should improve sales and marketing as a percentage of revenue. We're also in the early stages of building out our partner ecosystem. And we're looking at pipeline contribution from our partners, how partners help close business. We're in the early stages of that, and that's going to be another way that we increase sales and marketing productivity. And then last thing is the product innovation. This company doesn't exist without an incredible innovation engine that we've been building, and the new products and feature and functionality that we're releasing is improving and increasing the number of deals that we're winning, and it's increasing the average deal sizes over time. And for those of you who don't know, we have an investor track at our user conference called Current in October, where it's going to be very much a product- and customer-led Analyst Day. And because we're in our quiet period, we're not going to be talking about financials, but that's going to give investors an opportunity to really get a deeper dive into product road map, what we're releasing, hear from customers. So I highly encourage everyone to attend that either in person or virtually.

Wonderful. At this juncture, maybe I'll do a pulse check to see if anybody wants to jump in with a question. If you do have a question, we can send the mic your way.

Maybe while they're thinking about their questions, a question on $50 billion TAM opportunity. You mentioned that a lot of the legacy [indiscernible] are not there when you do bake-offs [indiscernible]. So who are you seeing? How do you think are the hyperscaler partnerships that you have helping you [indiscernible]?

We typically are seeing the open source Kafka as what we're replacing. And when we don't see that, then who else are we seeing? Well, each of the 3 cloud service providers, AWS, Azure and GCP, they all have competitive offerings. But we're way more friends than we are competitors. And the reason why that's the case is -- and having come from Google Cloud, the CSPs really want as many workloads as possible getting into their infrastructure, drive hundreds of petabytes of workloads to the cloud infrastructure providers. So they are very much focused on making us successful, even though they do have some competing products. Each of those 3 cloud service providers, their salespeople retire, get full quota retirement when they resell Confluent through the marketplace. And we've been getting typically deeper with each of the 3 cloud CSPs. So overall, the dynamic at play is there's a coopetition element with the 3 CSPs, but it's much more friendly than competitive.

Is there any other questions? Or we can keep going?

Can you quantify what [indiscernible]?

Well, I was talking about like the year-over-year growth. It's been very robust relative to like where they started, right? So oftentimes, those customers, $100,000 customers or $1 million-plus ARR customers, they could start with a $50,000 to $75,000 initial committed contract. And what happens is that usually starts with one use case in a company. And as companies add more use cases to the infrastructure that we're providing, the level of consumption goes up. And so this journey that customers are on, it's all about getting more use cases and more workloads into the mix. And there's a network effect that's happening where more data and more applications that get connected into the central nervous system called Confluent, there, it definitely gets more because then it's like, well, what else can be connected into the central nervous system? So we're an 8-year-old company this month. We've been selling for about 6 years. And if you look at the number of $100,000 customers, number of $1 million-plus customers, there has been robust growth over a relatively short period of time to get those customers to where they are.

How many salespeople do you have [indiscernible]

Yes. We don't give out quota-carrying salespeople, but roughly about half the company is in go-to-market -- in the go-to-market organization, which is a combination of sales and marketing and customer success. And so we don't give that specific number out, but there's been healthy growth year-over-year.

Yes. Go ahead, [ Jackie. ]

[indiscernible]

How does the hybrid revenue fit on the cloud bucket?

Yes. Yes. So as you think about the largest Confluent Platform customers that are hybrid customers, I would say that we're still -- estimate that like 70% to 80% of the revenue is still Confluent Platform revenue for those customers and the balance is Confluent Cloud. But Confluent Cloud, the growth rate is very high in those hybrid customers because we're literally just at the beginning stages of getting those cloud workloads for those large Confluent Platform customers to really grow. So over time, we think that -- and this isn't just for Confluent. This isn't just for hybrid. This is more broadly. Confluent Cloud will be the majority of revenues for the company over time. And in particular, for those hybrid customers, cloud will be the dominant revenue stream, but it's still early days for the hybrid customers.

Steffan, you were the CFO of Google public cloud. Does this mean Google is the largest of your 3 hyperscaler partners in terms of whatever, quantification of business, et cetera?

Say it again?

Is Google public cloud the largest contributor to your hyperscaler partnership revenues, given that you worked at...

Given that I worked -- yes. Yes. AWS is -- it kind of goes by market share. All 3 are on really robust growth rates. And to put more quantification on what robust means is there was a 600% increase in AWS transactions in Q2. We quantified that on our earnings call. And GCP and Azure are growing incredibly well, too. So this is -- if you think about thematically where we are, we're still in the early days of building out what we think is going to be a very healthy and big business.

We wanted to -- is there another question? Yes, please.

[indiscernible]

We do. Yes. The question is, do we have a specific strategy in the middle market? First off, we segment our business, enterprise and commercial. And commercial, I would put in the mid-market category. We are 100% Confluent Cloud-led in the commercial market -- in middle market. And we've seen the highest growth rates there as well. And why is that the case? Well, in the middle market, first of all, it's a very -- it's a large segment of companies that are out there. The profile of those companies, which are typically less than $1 billion in revenue, they cannot afford to have a dedicated engineering team building out a Kafka-based do-it-yourself offering for real time. They need a cloud-based offering, and we have perfect product market fit for that segment of the business.

Yes. Steffan, I just -- yes, go ahead, please.

[indiscernible]

Yes. So I'll refer back to what we talked about in -- on our Q2 earnings call, which was we actually saw a very healthy increase in sequential dollar growth. We ended up growing very helpfully quarter-over-quarter from Q2 to Q1. And so why is that the case? That is -- it basically goes back to these workloads that are running the businesses. You can't necessarily deprecate workloads that are core to running the business, which is a little bit different in the database segment or the analytics segment or the warehouse segment where you can spin up and spin down instances. In this type of software, you're really running primary workloads that are fundamental to operating your business. That's not to say we can't see perturbations. We -- if you look at the historical sequential cloud revenue growth, we've always been growing, but the amount varies by quarter. And part of that is based off of what projects are initiated, when they're implemented, ramp times. So hopefully, that gives you a sense.

Yes, please go ahead.

[indiscernible]

Yes. On our call, we mentioned that there was pockets across the world where we saw companies requiring more signatures to get deals done, not only in the back half of June, but that also precipitated through July as well. We haven't really seen a change in that environment. The good news is the vast majority of deals that we had -- we didn't call out by name, but those have closed since the quarter ended. So we feel really good about that. We haven't seen a change in extra signatures being required. And [ if you try ] to diagnose where is it most pronounced, it was really just pockets of it across the world. And so we feel good about the dynamics heading into this quarter.

To me, an extra signature is as simple as DocuSign, add another tag that says, sign here, sign here, and that should be it. I mean what are you guys talking about? I'm just kidding. Any other question? [ Kevin, ] go ahead.

[indiscernible]

[indiscernible] cycle where people are getting cash, bonus and equity. For company -- from employees that joined in a period where the stock is down precipitously, [indiscernible] the macro dynamics at play. [indiscernible] get it around, ensuring that some of the people who are most affected ended up getting a little bit of additional stock [indiscernible]. I believe that our [indiscernible] of the shareholders, and we're trying to build value over time. So the net of the answer is for a small subsection of employees that were dramatically impacted just by timing of when they joined, we wanted to take care of them. But for the rest of the employee base, it's all about building shareholder value over time.

[indiscernible]

[indiscernible] first half of the year.

We got 10 seconds.

[indiscernible] your guidance twice. So can you just talk to us a little bit about [indiscernible] how investors [indiscernible]

[indiscernible] employee hedged to our guidance [indiscernible] hedge that we put in, and that is reflective of what we saw overall [ dynamic ] [indiscernible].

We got to clear the room for the next presentation. I'm sorry, you could maybe ask the question off-line if that's okay. So thank you, Steffan, for doing this presentation. Thank you, everybody, for your attention.