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

So thanks, everyone, for joining. This is Day 2 of the Wells Fargo TMT Summit here in sunny Southern California. Michael Turrin, software analyst here at Wells. Very pleased to have Confluent's CFO Rohan Sivaram with us for the next session. A good packed group of attendees here. So we may open it up for questions towards the end.

But I have a list to get to, and I want to start high level because I think in a lot of investor conversations there's still a lot of education going on around data streaming and Confluent, what Kafka is used for and why Confluent. So maybe you can set the stage with just high level what the technology does, who it's built for and where points of differentiation sit versus open source?

Yes. For sure. First of all, thanks for having us, Michael. Good to see you as well. And good morning, everyone. So it's a great question. And what I thought would be a good idea is actually go back to our founding story. And our founders, Jay, Jun and Neha were working in the infrastructure software team at LinkedIn. And they realized pretty fast that how complex the data environment can get. Primarily when you think about when a new person joins LinkedIn, there is like hundreds of different actions that need to happen in the back end in real time. And they were candidly just struggling to make it work. And back at that time, there was no out-of-the-box solution to make this happen. So that was essentially the inspiration for Kafka, which was written by them while they were at LinkedIn. And subsequently, they open sourced it. And what Kafka does is it's a layer where all data streams from the organization actually resides and applications can actually pull the stream based on what their needs are. So going back to my LinkedIn example, when, say, someone is searching a particular event or someone's getting added to the network or someone is adding another person, there are multiple triggers of events that get -- that happen. And everything is in the Kafka layer, and these applications can seamlessly pull the data. So that's essentially what it is. And fast forward, Confluent was founded, and we've continued to differentiate ourselves from the open source version of Kafka with our cloud product, with our on-prem product, which we can get into, but that's the broad story, Michael.

Okay. Super helpful. And then maybe just your background as well because you recently stepped into the CFO role. So another aspect that investors may be newer to. You spent some time at the Investor Day, and I think investors are starting to kind of build the profile, but maybe you can give a little bit of your background, how long you've been with the company and just your focus area as a CFO.

Sure. Well, before Confluent, I started my career with Morgan Stanley in their asset management. Spent a few years there, then decided to go back to school. And after school so the last 15-plus years, I've been in tech operating roles. And the last decade has been an high-growth deck, 6 years at Palo Alto Networks and about a little over 3 years here at Confluent. When I left Palo Alto, I was Head of Finance, SVP Finance there. And when I joined Confluent 3 years back, I basically started and built FP&A, investor relations, treasury, business operations team ground up. And I was essentially joined at the hip with Steffan with respect to running the financial operations. So -- and my focus area is actually go back to why I joined Confluent, and that's not changed. The reason I joined Confluent 3 years back was, I'd say, the 3 Ts. There was a really large TAM, $60 billion-plus TAM and the opportunity when you couple that with hundreds of thousands of organizations using open source Kafka, very, very compelling. How you are taking advantage of the opportunity matters. And we were doing that with differentiated technology, and we continue to build a more from a technology perspective. And of course, last but not least, like we have an amazingly talented team. So when you combine these 3 together, there's a compelling opportunity. And when I stand here today and I think about it, the opportunity continues to be the same. And my focus areas has always been how do you take advantage of this opportunity. And you do that by essentially high level doing 3 things. Number one, you're allocating resources and capital in an efficient manner. And you're doing that to ensure that you're driving durable and efficient growth over a long period of time. And the third piece is you can only drive efficient growth if you have this focus on unit economics and an ROI culture within the company with respect to how you invest dollars. So making sure that we're focused on the big opportunity and how we are driving it is my focus area.

Okay. I'll ask one more high-level question and then we'll get into some of the details of the most recent earnings results as well. But I think one of the key questions for Confluent, and this is something I've asked Jay in the past around too, is just -- I think the platform story is fairly clear because of the type of industries that you're going after. On the cloud side, I know there's a sort of just the temporal nature of the company started with the platform and built towards cloud. But how do you influence customers who are large Kafka users over to Confluent's Cloud? What are the kind of stats, metrics, ROI, TCO type of statements that you can make? And that's something that the team has gotten better at from your perspective, maybe just kind of talk through the value proposition of the cloud and the potential to bring over the larger open source users, that would be very compatible.

Yes. I mean, that's a great question. When we look at our opportunity, the open source community is a really large opportunity, hundreds of thousands of organizations using it. And Confluent differentiates from open source on two fronts. First of all, there is this meaningful technology differentiation that we have, and then there is the cost side of it. On the technology side, we like to say that we have a complete data streaming platform. And what that means is it's cloud native. It's complete, and it's everywhere. I'll double-click into each of them because these are important. When I say cloud native, what it means is our technology scales up and down. It's got high SLA. It can be multi-tenant, all the key features. And for engineers here, you know that it's not an easy thing to build, and you don't get that in open source. So that's number one. Number two, our data streaming platform is complete. What that means is open source is all about streaming, which is critical. But our view, our vision is not just streaming. We've got 120-plus connectors and the connectors ecosystem which brings a lot of data into our platform. We have governance products which make sure that the right people have access to the right data. And last but not least, we also have stream processing, right? When you kind of combine all of this, we feel that we have meaningful technology differentiation. That's one. On the cost side, obviously, we provide meaningful ROI and TCO, and we've shared that a bunch of studies out there, which shows that the ROI and TCO we bring. And I just want to simplify it. If you're an open source, if you are a company using open source Kafka, where are you spending your dollars? You're spending your dollars typically in 3 categories: The first is you have these infrastructure costs that you're spending your money. And given that this is one of many things that you are doing, your focus is not to optimize and make that infrastructure cost get better. So typically, that's number one. Number two, you are paying for very, very expensive Kafka engineers. And it depends on the size of your open source shop. It could be a handful to 10, 15, 20 engineers, right. And the third category is you're also paying for governance and security bolt-on applications. What we do with Confluent Cloud is we're basically providing all these 3 together as a service to you. And we obviously spend a lot more time in driving efficiencies and making it cheaper for you.

Great. So on the Q3 results. Before I get into some of the metrics and the puts and takes of what you laid out, I just want to ask kind of a high-level question around how you manage internally through the volatility that we're seeing in the market. I think the volatility has been staggering for many of us across earnings cycles and it in many cases has intensified. So just internally when you're managing the company and the culture and trying to explain what's happening and keep everyone on the same page, what does that look like afterwards? I know the company has seen volatility in a few different stages, but maybe you can just give us the sort of the on-the-field perspective of what happens within the company when Q3 results come out and the stock is just seeing sort of a meaningful change in where it's sitting?

Yes. Michael, I'll go back to the point I made around, really taking a step back and looking at the opportunity and reminding the team that we have a special opportunity, large TAM, differentiated product and we're, I'd say, in full position to take advantage of this opportunity, not only because of the current products that we have but because of the vision that we have. And that's important. That's really important to continuously share that perspective. So I'd put that number one. Number two, it's all about controlling the controllables and focusing on what you can control. And from our standpoint, over the next 12 to 24 months, we have, from a technology standpoint, very important technology unlock that's going to come our way. So the focus has to be there. On the go-to-market side, we're going through this consumption transformation which we might talk about. And it's not a business model transformation, but it's a go-to-market transformation we need to be focused, and that's in our control, and we need to flawlessly execute against it. So it's -- while it's not easy, it's important to take yourself out of the noise and focus on the big opportunity at hand and focus on what we can control.

Yes. Good answer. So on some of the Q3 metrics, in terms of cloud revenue, maybe you can just level set for investors who aren't as close to the mechanics of what happened. We're used to seeing that net new ARR sequentially build throughout the year. When we started this year, I think the message was that progression was what was expected. We've seen some puts and takes. I think Q2 was a little stronger and then Q3 was a little bit lighter. So what are the factors that are kind of altering the trajectory in cloud currently?

Yes. For our Q3 results, we -- at our earnings call, we shared essentially 2 categories. The first category is what I'd call fairly idiosyncratic to customers impacted and I'll get into it. And the second category was more around -- we've been seeing the slowdown in net new use cases, primarily in the digital-native segment. So let me start with the first category. And the -- 2 large customers, we called out that the first was a large gaming customer. And this gaming customer made an independent architectural decision to move out of the cloud and go on-prem. And Confluent was impacted by their independent decision. The reason they made that decision was because they had an outage, which is public, right? And what that does is it's a fairly large customer and you don't get revenue from a cloud perspective from that customer anymore. So that's impact number one which had a partial impact in Q3 and a full quarter impact in Q4 and 2024. So that's. The second customer is a large customer of ours who are going through a transaction and they have a commitment, but they're going through a transaction. So the forecasted consumption curve versus the actual consumption curve, there's been a delta. And that showed up in our Q3 numbers. And that's what we called out. So to your question, this first category is fairly idiosyncratic and it's customer specific. The second category, the digital-native customers who -- candidly, they've been under cost pressure over the last 12 to 18 months. And we're seeing that the deployment of net new use cases there slow down. And these are probably the 2 drivers that have impacted our results in Q3. And we basically said that they will be impacting our Q4 numbers as well. And that's part of our guidance.

Super helpful. So a couple of just follow-on questions on that. So the customer that's moving off of cloud entirely. Is that something you could recapture as platform? Is that a theoretical possibility for the future for Confluent or is that not something we should consider?

So good reminder, I should have mentioned it. We are in active conversations with them as we speak to help them with their on-prem workloads. Yes.

Yes. Okay. And then on the net new use case side, is there anything that you've gathered that would support the notion that just data architecture conversations are becoming more complex and longer cycle because of AI? And every company having, I think, more intense focus on prioritizing data-driven, decision-making, but trying to figure out what to do given how many different vendors are coming at them with solutions, which are not competitive with Confluent but are, in some way, adjacent.

I'll say that the second category is not pervasive across our customer base. It's mostly in the digital-native segment. And that's an important differentiation. And that particular segment in the last 12-odd months have seen cost pressure. And as a result of that cost pressure, I'd say that we are seeing a slowdown in the net new use cases. I mean taking a step back, the 3 drivers of consumption revenue when you think first principles is existing use cases driving more volume. That's one. Number two, just net new use cases, which candidly is the bigger driver. And the third category is I was talking about the data streaming platform, selling more componentry of the data streaming platform. And that opportunity is ahead of us because we'll have a series of unlocks in 2024. So when you think big picture and take all of this into account, it's the second category where -- and that's also a subsegment of our customers.

Okay. So just from there moving on to -- I think -- So on the go-to-market changes that you're making. Maybe you can just help characterize how significant or less significant those are because I think when software investors hear go-to-market changes, they tend to think of sweeping overarching significant changes. And I think there's more general caution when those changes happen heading into the end of an annual cycle. So you chose to make the changes now and maybe you can start by just characterizing what those changes are, how significant or insignificant those changes you think will prove, and then we can talk about like what the forward vision after those could look like.

Yes. When you think about the go-to-market change, I'll just say a couple of, I'd say, points I want to make before I answer your question. This is not a business model change. It's a go-to-market change. And the reason I say it is our cloud business is based on consumption, rev rec, how we sell, how our customers consume, no changes there. So what are the changes? Changes are today, when a customer is using and consuming Confluent, they focus on consumption, and they want to commit to the amount that they are consuming. When you compare and contrast that with how we are incentivizing our sales force, we're currently incentivizing our sales force to get the largest ACV deal. And there is this inherent friction in the system. With this change, we are trying to take this friction away by incentivizing our salespeople based on consumption and not bookings and ACV. So in a nutshell, that's the change. And when you think about the change, we're not doing the change in Q4. It's going to happen Jan 1, 2024. And most of our consumption peers at some point in their life cycle have actually gone through this change, and they've had good results. So I wanted to call that out as well. And from the nuts and bolts of the change, what we're going to do is this year, we actually dipped our toe in the waters. And for a cloud deal, about 15% of incentive was based on consumption. We're going to move from that to 100% incentive based on consumption. So the focus of our reps will be to drive that incremental dollar of consumption.

Thank you. I mean the reason I ask, and I didn't want to necessarily lead the question is because it would seem like that's very much aligned with what your customers would expect in maybe in an environment where there is increasing focus on cost. Better aligning their expectations with consumption just removes some friction from decision-making. Is that a part of the consideration?

Yes. That is a part of the consideration. And it's all about focusing on what is the driver of consumption and channelizing all our efforts to that versus trying to get the -- I mean, you can get the largest commitment from your customer, but we recognize revenue based on consumption. So after you do that, the reps focus will obviously come back to consumption and making sure your customer is successful. So why focus on driving and getting the biggest deal?

So that's another point I wanted to just hit on. Does this at all change the type of sales rep? Based on what you said, it sounds like no. But is it the same sales rep that was going after a bookings number, going after consumption number, mapped similarly? Or does it change the profile of sort of go-to-market sales reps that you would consider?

I wouldn't say a huge change. I mean in the last reported quarter, north of 40% of our revenue was coming from cloud and consumption. So that rhythm, that DNA has been building in the company for the last couple of years. So I wouldn't say that there's going to be a major change. But any time there is a go-to-market change, you can expect a small part of your sales force might weed out. And we've kind of baked in that implication of that change into our guidance for 2024.

Okay. Yes, helpful. As the CFO, does that impact the visibility that you have into a customer spend? Or do you have enough cohort analysis and kind of data to make an educated guess around similar customers, similar profile, consumption patterns and mapping those all together? Maybe you can just touch a little bit on if that alters visibility at all for CFO and the forecasting and then we can get into assumptions used in the prelim targets?

Yes. So when you look at our business, the part of the business that comes from on-prem Confluent platform, there are no changes. So that's give or take, just shy of 50%. That will continue to be there. For the Cloud business, I think over the last few years, we've gone through this big change within -- with respect to how we forecast, how we manage. And it's a big change because in a traditional world, you start with bookings and then you have your rev rec rules and you come up with revenue. In a consumption world, you actually start with revenue. And we've built that DNA within the company where we're forecasting consumption at a customer level. And I'll tell you, it's difficult to just forecast based on spreadsheets. We have our data science team driving predictive models to figure out what the consumption patterns could look like. And then you have the qualitative inputs from the field for the larger customers, again, going back to net new use cases, et cetera. So you triangulate all of these factors to come up with a point of view. And I'll tell you, as a CFO, I definitely focus on the health of our installed base. And our gross retention rates ever since we've been a public company has been north of 90%, which obviously helps from an overall baseline setting as we look ahead.

Yes. That's great. I mean at this point, it seems like the key question that investors are focusing on is just the prelim targets that you frame for next year, the 22% growth targets. What went into those if those are likely to prove conservative? And I think that's you're going to get a lot of questions on it for the foreseeable future. We are as well. And so maybe you can use this as a forum to just go back to the assumptions that you used in those forecasts, if there are layers of conservatism because of some of the things we touched on there today that might not have been there in years past and you're sort of confidence in what can drive you towards those targets or ideally above if the company executes?

Yes. I'll start off with my overall guidance philosophy, which is -- which has been very consistent with how we've done it in the past, which is making sure you're setting prudent but achievable guidance and you're doing that by providing good visibility into the drivers of the business to the investor community. And that's not changed. And that was our approach as we looked at our 2024 guide. And again, it goes back to the 3 drivers that we've been touching on. So I won't get into the details, but I'll just touch on them, right? Number one is the two customers. Again, idiosyncratic, but will have a full year impact in 2024. The net new use cases from digital native, we expect that impact to go into 2024. And the consumption transformation that we just touched on, that is something it's a change. Any time you have a go-to-market change, you just need to be prudent. So we've built in that adjustment period we are going to see probably the first 6-odd months of next year into our numbers. And so these are 3 numbers, and that's how we thought about. I'll say that exiting next year, we'll obviously have a series of product unlocks and that starts with our stream processing product, which is going to GA in Q1 next year. There'll be series of networking and security unlocks that are going to happen. And we'll have FedRAMP. We'll have, obviously, more momentum in the marketplace around AI. So we see a reasonable amount of tailwinds for us as we exit Q4 of 2024. And that gives us confidence that we can bridge to the 30% medium-term target that we have.

Very helpful. And so on the Flink stream processing opportunity, is it your guidance philosophy because it's not yet available to kind of remove assumptions from that in terms of contribution for next year?

Yes. We're not assuming a huge amount of contribution next year. And we expect to GA the product in Q1. And any time you have infrastructure software products, you need to build applications. And I touched on it during Analyst Day, rough and tough it takes about 6 months for you to get value and get up to speed. So if you do the math, you're in Q4. So exiting next year, we're starting -- we're going to start seeing momentum on the Flink side.

Helpful. We have 7 minutes left. I'm going to turn it over to the group and see if anyone has a question. You could raise your hand. I think we have mics and that capability, if you'd like. Upfront. You can go ahead.

[indiscernible]

Yes. Should I repeat the question?

Yes. Yes.

I mean I think the question is around how do you get the quotas right, given the change? Yes. It's a great question. And when you think about what are the quotas going to be, the quotas are going to be consumption. And like I mentioned to Michael, we've been -- we are currently forecasting consumption at a customer level, purely based on -- that's how you need to do it. So from an internal operations perspective, how we are forecasting the business does not change. So our visibility will not change from a business perspective. So we feel good that we are in a good position operationally to set the right quotas. And that's -- that obviously we're going to learn as we go. But where we stand today, it's not going to be a huge change from an internal operations perspective with respect to how we set quotas. Having said that, how we run the business, it's going to be different. Like this year, definition of a pipeline is bookings. And next year, the definition of a pipeline is going to be net new use cases. So those are the changes that we are working on and making sure that we'll be ready as we get into January 2024.

Last one on margin. You've driven -- I mean -- so that became a point of focus when sort of the macro changed and Confluent shown good progress on free cash flow, I think on the cloud gross margins and what we get to from implied shows significant improvements there as well. How do you think about sources of future leverage there? And do some of the changes you're making on the go-to-market side impact the ways that we should think about margin progression at least in the -- sort of the near term?

Okay. If you look at the last 7-odd quarters for us, we've improved our margins by 35-plus points. And when you look at our guidance for Q4, we've said that we'll be op margin neutral. And when you look at our guidance for 2024, we've said that for the full year, we're going to be -- we'll break even from an op margin perspective. And when you compare guidance versus guidance, that's another 9-point of improvement full year. So we've seen a big trajectory from a margin improvement perspective. And -- you don't see this kind of improvement if you're not intentional about it. And if you don't have the buy-in from the company. So I think that's something that we've bought in that we need to be driving efficient growth. So that's how we are seeing a trajectory. So going back to your question around sources of leverage, I mean, clearly, if you look at last 7 quarters, go-to-market has been a contributor of sources of leverage. As you look ahead, we expect to see that go-to-market will contribute from a leverage perspective. And we're getting to a size where the partner ecosystem over, say, the next 12 to 24 months will contribute from that perspective because you get the attention of DSI and GSI. So that's one area. Obviously, through -- by removing the friction from the consumption transformation, you expect to see productivity improvement. That's going to be another source of leverage. And I'll be remiss if I don't mention -- you touched on it, like our cloud margin improvements have over the last 12, 18 months has been great, although we don't report the number, but again, making sure that we continue to drive that focus will be important.

Yes. That's great. Just a few minutes left, so I'm going to turn it back to you for closing thoughts. But I think what's important is to kind of think through the 3-year, the 5-year tailwinds that keep you enthusiastic around the Confluent opportunity. You touched on AI. We haven't touched on it in this conversation. So maybe between Flink and stream processing and potential other tailwinds that you see, what are the tech trends that have you excited and kind of viewing Confluent as this growing into a significant and expanding market opportunity as you initially characterized it?

Yes. I think today, when I look at the world, every company is a software company, and every company is a data company. So success of these companies will depend on how well they are able to harness their data. And the power of data will show up. And we're a key player in making sure companies can do that efficiently, companies can do that economically and companies can do that in a safe manner. So that's the biggest trend that we see. Now if you kind of look under that, like a couple of micro trends, clearly, cloud and the move to cloud, the digitization there is going to help. AI is going to help. If you go one level deeper, we're not just about streaming, we're about a platform. And we didn't spend a lot of time, but stream processing is very, very critical. It's about how you use your streams of data matter. I'll give you a simple example. If you are a travel company and your agents have a stream of data on flight times, and then there is another stream of data coming based on weather, and there's another stream of data based on customer information. You can combine all these 3 streams with stream processing and provide meaningful insights specific to specific customers around flight times, correlating to weather, correlating to what -- how it could impact them. So it's very powerful. And spend today, when you look at company, there's a large amount of spend happening in stream processing. So that's another area of opportunity that we see.

That's great. Lots to keep busy with. I appreciate all the details on the conversation, Rohan. Thanks for joining.

Thank you. Appreciate it.