All right. Good morning, everyone. Thanks for being here on the third day of the conference. I'm Jason Ader with William Blair. I'm pleased to introduce Rohan Sivaram, CFO of Confluent.
Before we begin, I'm required to inform you that a complete list of research disclosures or potential conflict of interest is available on our website at williamblair.com. With that out of the way, Rohan, thanks for being here, and we're going to just do the fireside chat format here. Hopefully, we'll have some time towards the end for some audience Q&A, and then we're going to do the breakout upstairs. And I don't know which room, but I'll...
Maher.
What is it?
Maher.
Meher or Maher? Maher. Okay. For the IR in the audience. So Rohan, could you give a brief overview of the history of Confluent for those that don't know the story well and what problem that you originally aimed to solve?
Yes. Jason, great to be here, and thank you for hosting us, and good morning to everyone. Yes, Confluent is on a mission to set data in motion, and we want to be the central nervous system for every organization for data. So what does this mean? And what problem are we solving?
So for that, let me spend a couple of minutes just walking you through how the data architecture, infrastructure looks within organizations today. When you look at data architecture within an organization, it's typically in 2 different states. The first one is the operational estate, which is essentially used to run the business.
And the second estate is the analytical estate which is used to analyze the business. On the operational estate, what's prevalent is think headcount management systems, CRMs, ERPs essentially tied together with application integration tools. And now if you visualize it, you're essentially using a bunch of point-to-point integration tools to put all of this together and make the data work.
Now let's talk about the analytical estate. The analytical estate essentially includes data warehouses, data lakes, AI/ML platforms and essentially, you use some form of ETL to move the data from the operational estate into the analytical estate and try to piece together a view of the world there.
And ultimately, what you're trying to do is you're trying to join both these estates to run your business. And as you kind of visualize, you have all these applications in both of these estates that are connected with point-to-point tools. And if you start with a blank sheet of paper and try to connect all of these point-to-point tools, you get to what I'd like to call the big spaghetti mess. And that's the big spaghetti mess of data that we are trying to solve.
So what do we do? Kafka is an infrastructure data layer, where real-time data flows from within the organization and is flowing in real time and different applications can harness the data in real time. And it kind of bridges the divide between the operational estate and the analytical estate. So that's a little bit of the context around the problem we are solving. Why does it matter?
It matters because every organization today is a data organization and every organization today is a software organization. And how they harness their data actually differentiates between success and failure for these organizations with respect to running their businesses. And when you really think about Confluent, we're a 10-year-old company.
And our growth has primarily come from the streaming side, where we are essentially moving data in a real-time fashion within the organization. Looking ahead, we're making this transition from a single product company to a platform where we are not only moving data, but we'll be moving high-quality data and moving data that's enriched.
So that makes the value proposition even stronger for us. Finally, I'll leave you with one last thought on this overview, which is over the last decade, we've built a business by focusing on durable growth and profitability. And we've built a business which is a little over $1 billion in revenue run rate, ARR, as of the last reported quarter.
Okay. Thanks for that overview, Rohan. Just to kind of bring things down to earth for folks, can you talk about some example customers and how they use Confluent and maybe some of the use cases?
Yes. As I mentioned, every company is a data company today and how they harness their data matters. So with that as a backdrop, like we have mission-critical use cases in literally every industry. And I'll share some numbers, and I'll get into some examples. We have over 40% of Fortune 500 companies using Confluent within multiple different industry verticals.
And almost all of the top customers in each of these industry verticals are Confluent customers. So let's start. Let me give you an example of a couple. On the financial services side, there's fraud detection. And the best way to really make it real is when a fraudulent transaction happened, say, 5 years back, what would happen was each of you would get a mail, like a paper mail in your mailbox saying that these are the 5 transactions that are fraudulent. Why don't you call this number and either dispute it or confirm the transactions?
Fast forward to today, you get a text, right, real time that this particular transaction happened. That's basically real-time data infrastructure architecture working in the back end to make that happen. So in the financial services side, fraud detection, high-frequency trading, they are use cases that are very prevalent.
On the retail side, you have point-of-sale inventory, real-time marketing, inventory management, there are -- on the manufacturing side, how you manage your inventory. So as you go industry by industry, there are real-time use cases across the board.
And do you have concentration with larger enterprises? Or are you very spread across kind of different customer sizes?
Yes. We -- obviously, from an industry vertical perspective, we -- when we started as a company, we were -- primarily, our first offering was this on-prem version of the product. And as you would expect, regulated industries were some of our earlier customers. And as we progress with our cloud customers, we've become a lot more diversified.
So from an industry vertical perspective, we are very well diversified. Financial services, technology, manufacturing, we're pretty much in all industry segments. And from a customer perspective, we shared at our last earnings call that when you look at individual customers, no custome is greater than 2% of our ARR. So not a huge amount of customer concentration as well.
Okay. Great. So data streaming market, which is basically a category you guys created, how do you help people size how big that market is and how fast it's growing at?
Yes. The way I like to think about it is, the market size is important, but like what drives the market size is the magnitude of the problem that you're solving. And the data problem today in enterprises continues to get bigger and bigger, it gets more -- it's getting more and more complex.
And with AI, I think the urgency is also really high. The modernization of data architecture to power your AI workloads, agentic workloads is critical. So in general, there are a lot of tailwinds that we are seeing from the broader market. And the tailwinds are in 3 categories. Obviously, there is more and more data getting produced.
There is this whole move to the cloud, which is -- which continues to happen and then there is AI. And these tailwinds underpin our market size of over $100 billion that we've called out at our Investor Day. And at the time of our IPO a couple of years back, the market size was $50 billion, we've seen this expansion. And a big part of this expansion has been driven by our product introductions that we've done in our core segment, which is streaming.
We've introduced new products. And then we've also kind of built the multiproduct platform with our data streaming platform, which includes our connector ecosystem, which includes stream processing, which includes governance, which is providing right access to the right people from a data perspective, and we recently introduced a new product called Tableflow.
So overall, we're making this transition from a single product company to a platform. And as we are doing it, we're also building our market size, which is over $100 billion with a lot of tailwinds of cloud, AI and data.
All right. Let's segue to AI because I'm sure everyone cares about that in terms of where you guys sit in the AI stack? And what is your sort of moat going forward in -- within that AI stack?
Today, Jason, every AI problem is a data problem. And ultimately, how you harness your data differentiates if you're going to be successful or not. And for any AI application workload you're trying to drive, be it if you're using an out-of-the-box solution or you're trying to build an application in-house, what you absolutely need is the right data architecture in the back end. That is, are you moving real-time data into the application? And that's critical.
So what we are seeing with AI in general, is there are a lot more applications that are moving to real time. There is a general trend of modernization of data architecture. And these essentially augur well for not only the broad real-time streaming category, but also Confluent in specific.
And specifically, when you think about Agentic AI, Agentic AI, these agents reside in the analytical databases. And the -- how they work, they work at conversational speed. So again, it goes back to the same point I'm making over and over again, which is for these agents to work, they need real-time access data. And that's the role that we play.
The back-end architecture of companies that are trying to move to real time that are trying to power these AI workloads, we play -- we will play an important role in that architecture.
So you're basically like a data pipeline layer. Is that the right way to think about it? So you're connecting real-time data into the into the AI systems. Is that...
We're moving real-time data into the data destinations that are eventually powering some kind of applications, which is using AI. And as we are doing it with streaming, we are moving the data with our DSP, data streaming platform, we are moving high-value data. We are moving enriched data, these data destinations.
Right. And it's right to think about you guys as kind of like a Switzerland, correct in the sense that the MongoDBs and the snowflakes and the data bricks and the Azures and the AWSs, they all sort of are either sources or destinations of data and you guys sort of sit in the middle of that. Is that the right way -- is that the right framing?
That is the right framing. I mean, from the onset, our product strategy, our strategy has been to be the Switzerland. And what I mean by this is when you think about form factors, we are on-prem, we're in the cloud. We're kind of agnostic. Then from a cloud perspective, we are in all 3 clouds, and we are agnostic. All our products can support all 3 clouds. And to your point around data destinations, we're agnostic, be it a data lake, data warehouse or any form of database, we're agnostic.
So that's been our product strategy, and that kind of puts us in a very unique position with respect to a pure-play player that is trying to be the Switzerland from an overall strategy perspective.
And from a competitive standpoint, is your main competition just the open-source Kafka that you guys are built upon? Or are there other competitors that are worth calling out?
Yes. The competitive landscape, I like to think about it is in 3 pillars. The first pillar, as you rightly called out is open-source Kafka. And we have a very large, vibrant ecosystem of open source users. Specifically, we have 150,000 organizations using open-source Kafka. So you can call it competition, you can call it our biggest opportunity. That's number one.
And how we differentiate ourselves from open-source Kafka is through our complete cloud-native managed service that we provide. So that is category 1, open-source Kafka and the way I like to call it is that it's our biggest opportunity...
To top of funnel basically.
Top of funnel, exactly. Number two is the hyperscalers. And that's also very unique because there is an element of competition where we compete with them and we partner with them very closely. On the partnership side, their reps can retire quotas, selling Confluent. Their customers can use credits for Confluent and we are selling in their marketplaces.
And from a competition perspective, each of the -- each of these hyperscalers, they have their Kafka product. So however, the big takeaway is, irrespective, we are moving a lot of data into their respective ecosystem, which is why the partnership element of this is very strong.
And the third category is I'd like to call it natural, where you have the application integration players, the ETL players and then you also have the venture-funded startups. So that's category 3. And we pay close attention to it. But category 1 and 2 are probably the areas that we see a lot more.
And what is the -- I guess what is the playbook for you guys to convert an open source customer to a paid customer? I know that's sort of the primary sales motion in some ways. But what -- how does that work? And how do you convince a customer to move from the open source to Confluent Cloud?
Yes. When you -- there's this misnomer that open source is free. It's not. And when a customer is basically using open-source Kafka, how are they spending their money? What are they doing? Well, what they are doing is they're using open-source Kafka. They still need infrastructure which will probably be 1 of the 3 hyperscalers from an infrastructure perspective. They have expensive Kafka engineers that need to help build the product and put it all together.
And as you're doing it, you need the right security and governance to make sure the right people have access to the right data. Typically, in a managed service, you kind of package all of this together as a service. And your goal is to provide a low friction at the right ROI and TCO.
So that's one big area of differentiation with respect to our managed cloud product, which -- and a big differentiator of our managed cloud product is it scales up and down with usage. So you don't have to provision for your peak workloads. You can use our product, and it will kind of scale up and down with usage. So that's number one.
Number two, over the last 12 to 18 months, we've done a lot of product innovation, especially on the streaming side. And now we have multiple products within our streaming portfolio that cater to the unique use cases for our customers. We have enterprise SKUs, which is multi-tenant which has the right amount of private networking built in, so you can use that at the right ROI TCO.
You have freight clusters and work stream. They're kind of similar from a use case perspective, where they are targeting very high throughput, but relaxed latency workloads. If you want to do it in our cloud, you can use freight clusters. If you want to do it as a BYOC that is bring your own cloud, you can use WarpStream. And then you have dedicated.
So we've provided multiple different options to our customers to come into our ecosystem. So first, we have a managed service with the right ROI TCO, and we have multiple optionality. So we are able to essentially have add bats for all our customers' use cases.
Let's talk about the go-to-market side of it. I would say the sales execution has been a little bit inconsistent over the last few years for you guys. And I know you've made some changes to the team and maybe to the approach. But can you just talk about how the go-to-market has evolved and kind of your confidence level that you kind of have the right strategy in place to continue to be successful?
From a go-to-market perspective, our go-to-market efforts have -- if I were to categorize it today, there are probably 3 ways to think about it. Number one, we have this focus on product-led growth where the developer community, the technologists are a very important part of our ecosystem. So they can come to Confluent's website, try a product, spin a cluster and start using the product.
So that is -- again, goes back to top of the funnel. And over the last couple of years, we've continued to tweak that motion and fine-tune it. And the ROI, the results you'll see is our top of the funnel, how many total customers are we adding. That's kind of correlated.
The second pillar of our go-to-market is what all of us know, the traditional enterprise go-to-market motion where you're talking to the tech execs and you're really driving consumption at scale or selling Confluent platform at scale. And for that area, when you look at our 100,000-plus ARR customers, that cohort contributes greater than 90% of our ARR.
So what that tells you is the traditional enterprise motion typically touches most of our ARR revenue, but both are equally important. And on this front, over the last couple of years, we've made some changes. In 2024, we kind of -- for our cloud business, we made changes to be -- to move our go-to-market motion to be consumption first and be it incentive structure, be it how we run the business.
And we are in year 2 of that. Every year, you make some tweaks, but broad brush, entering 2025, we've done some -- it's a continuity of what we've been doing last year. And the third area is something that I've spoken about recently, which is our partner ecosystem. We're getting to a scale and size where not only do we need our partners, but we are beneficial to our partners as well, be it the global system integrators, be it the regional system integrators, be it thinking about strategic partnerships.
So that's been a focus for us because it kind of helps amplify our message and provide scale to us. So these are some of the 3 puts and takes, but we've really focused on the changes that we made on the consumption side, and we are leaning into year 2 of the consumption transformation that we had.
Okay. Very helpful. I'm going to the earnings call now. So you talked on the earnings call about seeing some large customers focused on cloud cost optimization. But at the same time, you said you're not seeing really any macro impact in the quarter. So I guess can you square that for us if the large customers are leaning in on kind of cloud cost optimization but you don't think that's macro related and why did you call it out?
Yes. For what I said in the earnings call, I'll articulate it here one more time. What I said was for some of our larger customers across industries, we did see cost optimization and a slight slowdown in the net new use cases. And that was one thing. The other thing I said for the next year of customers, we did see some stability in our consumption patterns and our Confluent platform business was stable.
The reason we called it out because it's just articulating what we saw from our performance in the quarter and what are the dynamics and some of the puts and takes that we saw in our consumption business. And how that influenced with respect to our outlook for the rest of the year. So that's a little bit of the color of what we shared at the earnings call.
Okay. But didn't you also say that you didn't see an impact from macro?
Yes. I mean what we said was for our Q1 results. First of all, it's very hard to parse out what's macro. And in consumption businesses, you do see the sawtooth pattern of growth where if you look at the business and you draw -- you kind of look at 2 points in time and you draw a line, you see growth, but sometimes the growth can have the sawtooth pattern.
And so that's the dynamic that we called out. What we said was, typically, what follows these patterns is a recovery. And we've assumed that the rate of recovery, the slope of that line is going to be a little more gradual for the rest of the year.
So what gives you confidence that the updated guidance is derisked for the year?
Yes. I won't talk about the guidance. I mean, what we shared at the earnings call, it's in the transcript, so you can have a look at it. What I'll talk about is some of the growth drivers that we see what's ahead. And the growth drivers that we see fall in 4 categories.
We have this streaming opportunity where I shared, we have over 150,000 organizations using open-source Kafka. We are scratching the surface. So there is this large opportunity from a streaming perspective. And we've added on to our product portfolio. So that's growth driver one, focus area for us. Number two is our data streaming platform, where all the products within the data streaming platform, be it the connectors, stream processing, governance and Tableflow. They are in their earlier stages of their growth curve.
So that's opportunity number two. Number three is AI. As mainstream organizations adopt AI, we will have an important role in this new data architecture, data infrastructure. And the fourth one is our focus and what we are leaning in on the partner ecosystem side, with -- we called out OEM in Q1, we announced a Databricks partnership, which is a strategic partnership. That's an area of focus. So these are some of the puts and takes as we look at some of the growth drivers ahead.
Okay. Excellent. I guess last question for the session here. You meet with a lot of investors, what do you think is most underappreciated about the Confluent story?
Now I always like to think first principle and zoom out. And when you look at the data problem today, it continues to get more and more and more complex. And we are solving the problem with actually the industry's only complete data streaming platform from how we are approaching the problem.
And then you add AI where we're going to play a very important role. So I think just zooming out and looking at the big picture is super critical. And we have a large market opportunity, multiple tailwinds. 2024 as a year was a very important year from a product road map perspective. We had a bunch of new products come out.
And that kind of puts us in a good position with respect to taking advantage of the growth drivers that I laid out.
And I guess it's also hard to understand for some investors like the technical aspects of the story, right?
That's right. However, I'll say that we've come a long way from, say, 5 years back when streaming. I'm not joking, streaming was sometimes misconstrued to be streaming ESPN in your phones to what streaming is and what a real-time data architecture does today. So we've come a long way, but you're right, yes.
Okay. We'll...
This presentation has now finished.
