Confluent - Q2 2024
July 31, 2024
Transcript
Shane Xie (Head of Investor Relations)
Hello everyone, welcome to the Confluent Q2 2024 Earnings Conference call. I'm Shane Xie from Investor Relations, and I'm joined by Jay Kreps, Co-founder and CEO, and Rohan Sivaram, CFO. During today's call, management will make forward-looking statements regarding our business, operations, sales strategy, market and product positioning, financial performance, and future prospects, including statements regarding our financial guidance for the fiscal Q3 of 2024 and fiscal year 2024.
Alex Zukin (Analyst)
These forward-looking statements are subject to risks and uncertainties, which could cause actual results to differ materially from those anticipated by these statements. Further information on risk factors that could cause actual results to differ is included in our most recent Form 10-Q filed with the SEC. We assume no obligation to update these statements after today's call, except as required by law.
Unless stated otherwise, certain financial measures used on today's call are expressed on a non-GAAP basis, and all comparisons are made on a year-over-year basis. We use these non-GAAP financial measures internally to facilitate analysis of our financial and business trends and for internal planning and forecasting purposes. These non-GAAP financial measures have limitations and should not be considered in isolation from or as a substitute for financial information prepared in accordance with GAAP.
A reconciliation between these GAAP and non-GAAP financial measures is included in our earnings press release and supplemental financials, which can be found on our IR website at investor.confluent.io. Finally, once we've concluded our prepared remarks, we will post the Confluent earnings report to our IR website. With that, I'll turn it over to Jay.
Jay Kreps (CEO)
Thanks, Shane. Good afternoon, everyone, and welcome to our Q2 earnings call. I'm pleased to report a solid Q2, once again exceeding our revenue and margin guidance despite a continuing volatile macro environment. Subscription revenue grew 27% to $225 million, Confluent Cloud revenue grew 40% to $117 million, and non-GAAP operating margin was positive, representing approximately 10 percentage points in improvement.
Alex Zukin (Analyst)
These results underscore the power of our data streaming platform and our relentless focus on delivering success for our customers. Today, I'll start with providing an update on our consumption transformation. Overall, the vast majority of our rollout towards our transformation to becoming consumption-oriented is complete, a step change in how we run our cloud business. One success indicator for our transformation is new logo growth.
I'm pleased to share that we increased our total customer count by 320 in Q2, double from the previous quarter, and representing our largest sequential increase in 2 years. In addition to landing a higher volume of customers, we believe we have increased the quality. A strong focus on our target counts has increased the percentage of high-propensity customer lands.
Though this new consumption motion lands them earlier in their journey, we believe over time many of these customers will represent the next wave of high-consumption accounts for us. We remain as confident as ever in the strategic position of the company and the prospects for durable long-term growth. We have the industry-leading technology in an increasingly critical and relevant category that we believe will be as important as databases.
Over the past year, we've made a series of innovations to build out the full set of capabilities of a data streaming platform, enabling us to capture the full lifecycle of data in motion. As we progress along our consumption transformation, we will be better equipped than ever to rapidly acquire new customers, win new workloads, and fuel the adoption of our full set of product capabilities.
Our rapid pace of innovation and ability to land high-quality customers who have the potential to consume more of our product leaves me more excited than ever about the long-term opportunity to capture the lion's share of the data streaming market. Vimeo, a leading SaaS video platform with over 260 million users, is a great example of a customer using our complete data streaming platform.
Success for video platforms like Vimeo largely depends on delivering stellar user experiences, but traditional data warehousing and Batch ETL processes limited their ability to see real-time customer usage, hindering timely decisions, quick pivots on products and campaigns, and adaptive user experiences without buffering. So they turned to Confluent to enable real-time data flows that provided the visibility they lacked.
Fully managed connectors to Snowflake, S3, and others make it easy to instantly connect data through their business without the hassle of building and self-managing connectors. Stream Governance ensures data quality and security and allows Vimeo to safely scale and share data streams. And the team is looking towards Flink to enhance their streaming use cases while freeing up bandwidth across the team. With Confluent and real-time data flows, Vimeo can make quick decisions and deliver top-quality personalized experiences that drive growth and profitability.
Next, I'd like to discuss the critical role our partners play in our business and some exciting announcements in this area. Partners are essential to our strategy to expand our reach into new markets, improve sales efficiency, and complement our data streaming platform. We're pleased to share a number of notable achievements and milestones across all three pillars of our partner ecosystem. In Q2, we announced Build with Confluent and Accelerate with Confluent, which makes it easier for our global SI partners like KPMG, Accenture, and EY to build specialized offerings as they embed data streaming into their core business.
A wide range of use cases have already been established through the Build with Confluent program, including a GenAI bot for airline customer support, fraud detection against AI-powered voice phishing, automated limit increases for credit card users, and real-time telemetry analysis for freight optimization.
We're also seeing great traction with Connect with Confluent, a technology partner program that makes it easier for partners to build native integrations with Confluent Cloud. In Q2, we crossed more than 40 technology partner-built integrations, including SAP, MongoDB, Imply, and services at Google Cloud and AWS, giving us coverage across the major segments of the modern data and AI stack to help us drive consumption of Confluent.
The success of Connect with Confluent has tripled the amount of data traffic from our partner integration since the start of the year. And finally, on the CSP front, we were pleased to be recognized as Partners of the Year by both Google and Microsoft. This marks the third year in a row that Confluent won this award from Microsoft and the fifth year that we've been recognized by Google.
Together, these recognitions underscore our highly valuable and symbiotic relationship with our cloud partners. As we discussed last quarter, data streaming plays a critical role in fueling GenAI applications with contextual and trustworthy streams of real-time data. This was underscored by our 2024 data streaming report. 90% of the 4,000+ IT leaders we surveyed said data streaming platforms can lead to more product and service innovation in AI and ML development, with 63% saying data streaming platforms significantly fuel AI progress by building a real-time data foundation.
I'm pleased to share a couple of examples of this. A top five mortgage lender in the United States is turning to Confluent and GenAI to reimagine the home buying experience. This company uses generative AI to listen to client calls, transcribe them, analyze sentiment, and record client patterns and preferences to create personalized experiences for millions of customers and prospects.
They turn to Confluent as a key part of their RAG-enabled architecture to quickly build and scale GenAI use cases by tapping into readily usable, trustworthy data streams. Connectors allow them to connect data and systems from across lines of business, while stream governance and stream processing enable the team to create reusable data products that are easy to find, understand, and use. As a result of their GenAI initiatives, approximately 70% of servicing calls can be fully self-served without the need for full team member intervention.
By saving the client servicing team 40,000 hours annually, team members can concentrate on cultivating strong, meaningful client relationships, while AI manages the mundane tasks. But this is just the start. In the future, their GenAI platform will learn homeowners' preferences and communication habits so team members can anticipate and solve clients' needs.
An international e-commerce company in Germany with €16 billion in annual sales is another example of companies turning to Confluent to play a key role in their GenAI stack. When e-commerce products from this company's iconic brands are promoted, their call centers receive massive spikes in call volume. But with only 100 call center reps, customers are often left in long wait queues, leading to failed conversions and unhappy customers.
So they turn to Confluent to help power GenAI applications like voice bots that can scale up in seconds and answer 1,000 calls in parallel. As customers interact with these voice bots, an AI order entry bot enables the order completion process and communicates with ERP systems in real time via Confluent, streaming transactional data such as product, order, customer, payment, and billing information.
A key part of this workflow is Confluent streaming KPIs from bots in the operational domain to their analytical systems, capturing data like real-time orders, call metrics, sentiment analysis, and how many AI tokens were used to analyze and measure the efficacy of calls to continuously make their bots smarter and more effective.
Powered by enriched, trusted data streams from Confluent, this e-commerce company can now handle unexpected traffic spikes to increase call center capacity on demand to reduce customer wait times and improve order completion rate. In closing, we're excited about the trajectory we're on. With our complete data streaming platform and our transition to a consumption-oriented business, we're well positioned to attract more customers who can drive even more value from Confluent as they capture the full lifecycle of streaming data from our platform. The future of Confluent is incredibly bright. With that, I'll turn it over to Rohan.
Rohan Sivaram (CFO)
Thanks, Jay. Good afternoon, everyone. In Q2, we delivered solid subscription revenue growth and substantial margin expansions, ending the quarter with positive non-GAAP operating margin and free cash flow margin. These results, coupled with continued adoption of our data streaming platform, demonstrate our relentless focus on driving long-term efficient growth in a volatile macro environment.
Alex Zukin (Analyst)
Turning to the Q2 results, subscription revenue grew 27% to $224.7 million, exceeding the high end of our guidance and representing 96% of total revenue. Confluent Platform revenue growth accelerated to 16%, ending the quarter at $107.3 million and accounting for 48% of subscription revenue. We closed two eight-figure multi-year deals consisting of renewal and expansion with existing customers in the financial services industry. This underscores Confluent as the platform of choice for data streaming among the world's most established enterprises.
Confluent Cloud revenue grew 40% to $117.4 million and accounted for 52% of subscription revenue, compared to 47% in the year-ago quarter. We are pleased with delivering high growth at scale and continuing to improve the margin profile for our cloud business. Turning to the geographic mix of total revenue, revenue from the U.S. grew 26% to $143.2 million. Revenue from outside the U.S. grew 22% to $91.7 million. Moving on to the rest of the income statement,
I'll be referring to non-GAAP results unless stated otherwise. Subscription gross margin was 80.8%, up 170 basis points. Gross margin performance was driven by strong Confluent Platform margin and improving unit economics of our Confluent Cloud offering. Turning to profitability and cash flow, operating margin expanded 9.7 percentage points to 0.6%, representing our eighth consecutive quarter of nine points or more in margin improvement.
Operating margin performance was driven by our gross margin performance and our continued focus on driving efficient growth across the company. Net income per share was $0.06 for Q2, using 354.2 million diluted weighted average shares outstanding. Fully diluted share count under the treasury stock method was approximately 362.9 million. Free cash flow margin turned positive and improved approximately 20 percentage points to 1.2%.
We ended Q2 with $1.93 billion in cash, cash equivalents, and marketable securities. Turning now to other business metrics, total customer count was approximately 5,440, up 320 customers sequentially, our largest sequential growth in two years. We are pleased with accelerating growth in total customer count and the quality of customers we have acquired. Additionally, we added 46 customers with $100K+ in ARR and nine customers with $1 million+ in ARR, bringing the total to 1,306 and 177, respectively.
Our 100K plus ARR customers continue to contribute greater than 85% of our revenue. Our new $1 million plus ARR customers include customers from a variety of industries, including energy, financial services, manufacturing, retail, transportation, and more. Turning to NRR, Q2 NRR was 118%, below our target range of 120%-125% for this year. This was due to continuing consumption volatility within our large digital native customer base.
After the stabilization in Q1 and a healthy start in Q2, we saw increased short-term cloud cost controls and focus on driving efficiencies in this customer cohort in the month of June, which impacted expansion of new use cases. While the green shoots we saw earlier this year haven't yet translated to the level of consumption we had expected, we are pleased to see continued strength in our gross retention rate.
GRR remained above 90% at a level consistent with the last two quarters, reflecting the sticky nature of our data streaming platform and our continued focus on delivering strong value and ROI to our customers. As we frame our guidance for the second half of 2024, there are two points I would like to call out before getting into the numbers. First, growth of our Confluent Platform business remains lumpy, driven by the timing of large renewal and expansion deals.
As a reminder, approximately 20% of our Confluent Platform TCV is recognized as upfront license revenue. This could create short-term variability for growth in Confluent Platform. Q2 was a good example where we benefited from this dynamic, which creates a tough compare for growth for the rest of 2024. Second, consumption in our digital native customer base remains volatile.
As mentioned, while the green shoots we saw over the last few months continue to take hold, they have not yet translated to the level of consumption we had expected. We have incorporated this consumption dynamic from our large digital native customers into our second-half outlook. Now, let's turn to guidance. For the Q3 of 2024, we expect subscription revenue to be in the range of $233 million-$234 million, representing growth of approximately 23%-24%.
Non-GAAP operating margin to break even, representing improvement of approximately 5 percentage points. And non-GAAP net income per diluted share to be $0.05. For the full year 2024, we expect subscription revenue to be approximately $910 million, representing growth of approximately 25%. Non-GAAP operating margin to break even, representing improvement of approximately 7 percentage points. Non-GAAP net income per diluted share to be $0.20.
Free cash flow margin to break even, representing improvement of approximately 16 percentage points. Looking out longer term, we remain well positioned to address our market opportunity. As Jay pointed out earlier, we're going after a highly strategic and mission-critical data streaming market, which we believe will be as important as databases. Over the last few decades, history in software has repeated itself many times that platform approach wins.
Our industry-leading data streaming platform is the only complete platform spanning stream, connect, process, and govern, enabling us to capture the full lifecycle of data in motion at a lower TCO and delivering strong ROI to our customers. In Q2, we saw continued adoption of our data streaming platform. Connect, process, and govern of our DSP portfolio accounted for a larger portion of cloud revenue and grew substantially faster than our overall cloud business.
Multi-product customers remained our fastest-growing customer cohort, with NRR substantially higher than 130%. We believe DSP will continue to outgrow our core business for the foreseeable future. As it continues to scale, DSP is expected to be a strong growth tailwind to our business over a long period of time. Additionally, the rise of GenAI makes it clear that a modern organization's success is significantly influenced by its data strategy, particularly around data in motion.
With a complete cloud-native and ubiquitous platform and secular tailwinds for data in motion, we have never been more confident in our ability to sustain durable and efficient growth over the long term. In closing, we're pleased with our solid Q2 results. As we continue to execute on our consumption transformation, we remain focused on driving efficient growth and delivering break even for non-GAAP operating margin and free cash flow margin for 2024.Now, Jay and I will take your questions.
Shane Xie (Head of Investor Relations)
Thanks, Rohan. To join the Q&A, please raise your hand. Today, we will take our first question from Matt Hedberg with RBC, followed by Morgan Stanley.
Matt Hedberg (Analyst)
Great. Thanks for taking my questions, guys. I'm in the airport, so hopefully the background noise is okay. Jay, you know I noticed you know the strong customer logo add stood out to me. I think you noted on your prepared remarks, the quality was also higher. I guess, does that imply that you could see faster expansion from this cohort as they continue to advance with the platform?
Jay Kreps (CEO)
Yeah, over time. So one of the measures we, first of all, I would say, as we undertook this consumption transformation, one of our key goals was to accelerate customer acquisition. So that's played out as we desired. The second thing that we were aiming for was to be a little more targeted in which customers we're landing in.
Alex Zukin (Analyst)
And so we built a model to try and predict the opportunity in these customers and tagged a set of accounts we felt were high propensity. So the measure we use internally is what percentage of these new customer lands are in that high propensity segment. And so obviously, it's on us to go drive the growth in the accounts. One of the things we think is promising is that that percentage has gone up substantially when we look at the first half of this year versus last year in the customers we're landing in.
Matt Hedberg (Analyst)
Got it. Thank you. And then maybe one for you, Rohan. You noted some of the NRR pressure and the potential for lumpy Confluent platform business in the second half. It seems like that's contemplated in your guidance. I guess I'm curious, after you saw, it sounds like a little bit of slowness in June, how did things trend in July? I mean, have you seen things kind of stabilize a bit or just maybe a little bit more on the linearity on the cloud side? Thanks.
Rohan Sivaram (CFO)
Yeah. Thanks for your question, Matt. Yeah, I mean, you know from an overall guidance perspective, as you rightly pointed out, we've considered two factors. Our Confluent Platform business tends to be lumpy primarily because of how the rubric is and the timing of some of the larger deals. So that's incorporated in the guide.
Alex Zukin (Analyst)
And the second one that you mentioned was around the digital native customers. So for digital native customers, Q1, we saw stabilization. And as we entered Q2, we did see a continuation of that stabilization. However, I would say towards the latter half of the quarter, we did see cost efficiencies and focus on just driving some cost efficiencies from some of these customers, which actually continued into the month of July. As we think of the back half of the year, we think taking a thoughtful and prudent approach to guidance would be the right thing to do.
Matt Hedberg (Analyst)
Thanks. Likewise.
Shane Xie (Head of Investor Relations)
Thanks, Matt. We'll take our next question from Sanjit with Morgan Stanley, followed by Barclays.
Sanjit Singh (Analyst)
Thank you for taking the question. I want to return back to the sort of go-to-market transformation. I think there was an approach to utilize similar sales motions like MongoDB, who've done modern cloud sales. One of the things that they've been seeing is that as they try and incent for more volume and higher volumes of workloads, the quality 12 months down the road wasn't what they expected.
Alex Zukin (Analyst)
I was wondering, Jay, I'm happy to see the new customer logos really accelerate. But if you look at it on a workload basis, how are you guys thinking about ensuring quality of the workloads and not just seeing that increased volume?
Jay Kreps (CEO)
Yeah, you know that's obviously on us to ensure that we're kind of getting the growth and expansion out of this customer set. You know the first indication is just, are we getting into the right customers? I think that we've seen, you know as we think about over the course of this year, we want to see those grow.
Alex Zukin (Analyst)
Part of our objective was definitely to get into customers earlier in the lifecycle, earlier in the development lifecycle, where they're actually setting up the architecture, where they can use more of our DSP in the first application they're planning. I think that's been successful, and that's allowed us to take up the volume of customers. But yeah, really showing how those play out and spill into the $100K plus and $1 million plus, that will happen over time.
Sanjit Singh (Analyst)
Great. And then just as a follow-up, how are you guys thinking about when you look at the pieces of a real-time data infrastructure? I think OpenAI bought what's called like a real-time database. And you guys have sort of talked about the streaming platform, the stream processing platform. Where does a real-time database sort of fit in? And is that something that another service offering that Confluent may get involved with over time?
Jay Kreps (CEO)
Yeah. You know we actually partner very closely with a number of providers of similar functionality. So I think Rockset was the company that OpenAI announced having bought. There's probably half a dozen to a dozen companies in that segment where they're taking streams of data and serving different types of queries against those streams. And that's a great set of partners for us.
Alex Zukin (Analyst)
And we work with them around a whole set of applications, including some in the GenAI space, use cases around search, use cases around real-time analytics, use cases around aggregating and storing other key information sets that power workloads. So there's definitely kind of a nice integration there between some of the real-time databases or streaming databases and the rest of the streaming platform.
Shane Xie (Head of Investor Relations)
Thanks, Jay.
Thanks, Shane. We'll take our next question from Raimo Lenschow with Barclays, followed by Deutsche.
Raimo Lenschow (Analyst)
Hey, thank you. Two very quick questions. First, we had a bit of a change on the sales leadership, and I wanted to see if you kind of could have given us some additional comments there. And then, Jay, on Flink as well, we kind of have it in the market for a few months now, and you kind of mentioned some of that on the call. But what's your experience in so far in terms of adoption, what people are using for versus what you expected on the user side? Thank you.
Jay Kreps (CEO)
Yeah. Yeah, those are good questions. So yeah, we did have one departure on the sales side. The structure of our team is the larger field effort, the reps, technical folks, BD folks, consulting, is all under Erica Schultz. So she's kind of at the top of the sales organization and reports to me. She remains in charge of that and has been at Confluent for over four years.
Alex Zukin (Analyst)
So there's a fair amount of continuity. Under her, the management structure for the reps has kind of bounced back and forth between something where she manages the individual theater leaders like APAC or EMEA directly, or has somebody who manages that set of theater leaders. And so we did make a change in that structure. I won't comment on the departure in detail, but we're not planning on immediately backfilling it. We're comfortable with the arrangement we've had.
That's worked well for us in the past, and we feel good about that in the near term. On Flink, yeah, that continues to go well. This is a ton of excitement from our customers. I think last quarter, we said on the order of around 600 prospects and customers had tried it out. That's up to over 1,000 now. We're starting to see really nice growth in the revenue from that.
But of course, it's still early days, small numbers. What we want to see is that ramp to a larger percentage of the business, see more of the kind of large-scale production use cases come online. That's what we're looking for over the course of this year. If we're successful in that, that will be a very positive tailwind for the business overall as it gets to a larger portion of the business heading into next year.
Raimo Lenschow (Analyst)
OK, perfect. Thank you.
Shane Xie (Head of Investor Relations)
Thanks, Ryan Moe. We'll take our next question from Brett Zelnick with Deutsche Bank, followed by William Blair.
Brett Zelnick (Analyst)
Thanks very much, Shane. Great to see the progress in new logo and the path to profitability. Rohan, you characterize digital native customers as volatile. What's the behavior that you're seeing specifically? Are they downselling? Are they churning? And is this more cloud or Confluent Platform? How much of a risk is this going forward and the extent to which you're embedding it in the guide?
Rohan Sivaram (CFO)
Yeah, thanks for the question, Brad. I mean, when you really look at the digital native segment and customers, and if you zoom out and look at it over a longer period of time, the growth has been linear. And growth has been upward, but not linear. So over a period of time, this segment, we've seen good growth, and the growth rates have been higher than the rate of the company.
Alex Zukin (Analyst)
But when you really look at the last 12-odd months, this segment has been under some kind of cost efficiency pressure. And you do see every now and then customers just focusing on the cost side of it and then adding more workloads. So that is a trend that we've seen.
But the big takeaway from my perspective is if you step back and look at it over a longer period of time, this segment has been actually a good grower for us. And that's the long-term outlook. But as you think about how we saw the month of June, we thought it's just prudent to bake it into our second-half outlook as we look ahead.
Jay Kreps (CEO)
Yeah, I think that's a good characterization. I mean, when we step back and think about the digital native segment overall, we remain very optimistic. That's an area that's grown for us very nicely over the last few years, and we think we'll continue that way. But as Rohan said, yeah, there is probably more cost pressure there.
Alex Zukin (Analyst)
And that does mean that if we look at Q1 to Q2, we did see different performance in that segment between those two quarters where we saw outsized growth in Q1 and a little bit more optimization in Q2. And I think that's kind of the balance of those two forces. We expect that to play out in the second half of the year as well.
Brett Zelnick (Analyst)
Thanks, Jay. Maybe just a quick follow-up for you. I see you have a number of open job recs for AEs in the U.S. federal market. Can you just remind us of where you stand in that market? And if you see upside in this Q3, that could come from federal, or is this more of a 2025 opportunity? Thanks.
Jay Kreps (CEO)
Yeah, yeah. By and large, open sales hires in the federal space that we're making in this next quarter would impact next year. But indeed, there's a very talented team that covers the public sector in the U.S. and elsewhere and is a good contributor to the business. So yeah, we're adding to that team, and we do think that's an area where we'll see growth.
Brett Zelnick (Analyst)
Cool. Thanks, and keep up the good work, guys.
Shane Xie (Head of Investor Relations)
Thanks. We'll take our next question from Jason Ader with William Blair, followed by TD Cowen.
Jason Ader (Analyst)
Yeah, thanks, Shane. Good afternoon, everyone. I just want to ask about go-to-market. It seems like it's been a bit of an evolution over the last several years, and you guys are still kind of figuring it out. Beyond the change to consumption and the management structure that you alluded to, what else would you call out that you've learned that could help move the needle on your go-to-market motion and where the best bang for the buck investments are going forward?
Jay Kreps (CEO)
Yeah, it's a good question. If I think about the first half of this year, it was mostly that shift to consumption. It's kind of rebuilding what you had on a new foundation around a different set of metrics, oriented on something new. If I think about the balance of the year and beyond, it's really taking advantage of that structure.
Alex Zukin (Analyst)
I do think that's the right setup for the type of cloud infrastructure company that we are. If you look at both peer companies and what the cloud providers have done over time to drive growth of workloads, it really hinges around that. And so I feel good about the progress we've made and some of the early results from that. I think it's even more exciting when we think about how we can use that to drive the adoption of these expansion product areas and the expansion of new workloads in the customer base.
Jason Ader (Analyst)
Just a quick follow-up on the consumption transition you talked about. I think you said smaller lands, Jay, is that right?
Jay Kreps (CEO)
Yeah, yeah. If I think about what we've shown, we've kind of taken up the volume of lands, but we will be catching these kind of earlier in the build cycle. We think that's largely good, but it does mean in that Q1, the contribution from a given customer will be smaller than it was in the past just because we're earlier in that project lifecycle.
Jason Ader (Analyst)
Thank you.
Shane Xie (Head of Investor Relations)
Thanks, Jason. We'll take our next question from Derek Wood with TD Cowen, followed by D.A. Davidson.
Derek Wood (Analyst)
Great. Thanks, Jay. In terms of the new motions of cross-selling DSP to the base, can you talk about what parts of that portfolio are seeing the strongest uptake when you look at governance, connectors, processors? Is there anything you'd highlight getting more traction? And when you sell into a core customer, what's often the ASP uplift when you cross-sell one of those components?
Jay Kreps (CEO)
Yeah, yeah. It's a good question. Yeah, I would think of each of those products as being in a kind of S curve where it's on some ramp. The earliest of those was our connector ecosystem, where we continue to make improvements and drive growth with it. We think there's a lot of opportunity there. The governance portfolio was next, and Flink is the newest addition to that.
Alex Zukin (Analyst)
And so I think there's interest and traction around all three, but Connect is the bigger business out of those three just because it started earlier. When we think about kind of percentage growth, of course, you'd see the highest percentage off Flink because it's starting from the small space.
And so we want to continue to ramp all three of those, but we expect the attach from those components to increase quite substantially in the year ahead as we open that up to more customers. What drives that is, in part, maturity of the product.
A lot of the security, kind of private networking functionality, other things that are important for our largest revenue customers to use those components are either just added or just being added over the course of this year. And that really opens up a lot more of the large-scale production usage to the rest of our customer base. And so that's a critical focus for us when we think about the balance of the year. And then.
Derek Wood (Analyst)
Helpful color.
Jay Kreps (CEO)
Yeah, we haven't given an exact kind of uplift number. We shared a little bit of kind of sketch of the shape of that business so far. I think, to be honest, I think it's still early days as these components ramp. Even a customer that's adopted Connectors, there's more of them they can use. Even if they have their first few Flink queries, we would like to take over a substantial portion of their applications in the stream processing layer.
Alex Zukin (Analyst)
And so I wouldn't think of that as a dynamic uplift of like, oh, you'll see X%, and it's binary either on or off. I would see this is kind of unlocking an adoption cycle that drives growth, not just in core Kafka, but now in these components as well, which will themselves produce more growth in the underlying Kafka usage.
Derek Wood (Analyst)
Yeah, that makes a lot of sense. Thanks. Rohan, for you, last quarter, I think you highlighted targets of getting to 125% NRR next year. Given that NRR dropped below 120% and you're seeing some different behavior, how does this kind of change your view on where you think you can get to in fiscal 2025 on NRR?
Rohan Sivaram (CFO)
Yeah, Derek, I mean, when I really look at the medium term, as Jay highlighted right now, the outlook has not changed at all. I'll basically talk about three things. First, Data Streaming Platform. And as Jay just pointed out, the products within our Data Streaming Platform, be it Connect, Stream Processing, or Flink, or Governance, they're all early stages of adoption.
Alex Zukin (Analyst)
And 2025 is going to be the year where there is going to be monetization and material contribution to revenue. The second is around consumption transformation. By the time we finish the year, we'll have completed a full year of the transformation and a full year of fine-tuning our go-to-market, all things consumption. So we expect to get tailwind from that. And the third category is Gen AI.
We've been talking to customers, and Jay highlighted also in his prepared remarks, real-time data is going to be a key component and key part of this broader ecosystem. So when you really take a step back and look at some of these drivers and vectors of growth, our view around NRR getting back to our medium-term targets has not changed.
Derek Wood (Analyst)
Thank you. Appreciate the color
Shane Xie (Head of Investor Relations)
Thanks, Derek. We'll take our next question from Rudy Kessinger with D.A. Davidson, followed by Wolfe Research.
Rudy Kessinger (Analyst)
Hey, guys. Can you hear me OK? I'm having a bit of a connectivity issue.
Jay Kreps (CEO)
Yeah, loud and clear.
Rudy Kessinger (Analyst)
OK, great. I'm curious, on international, the growth has actually slowed there quite rapidly relative to U.S. growth. And in some of the sales attrition I've seen, it seems like it's been more weighted towards folks in Australia, APAC, more so than the U.S. relative to your revenue split. So any color you can provide there just on some of the slowdown in international?
Rohan Sivaram (CFO)
Yeah, do you want to speak to that, Rohan? Yeah, happy to. Rudy, candidly, not a whole lot to read into it. When you see the relative, I would say, difference in growth rates between international and U.S., more often than not, the driver is timing of some of these larger Confluent Platform deals. And I mean, Q2 was exactly the same.
Alex Zukin (Analyst)
We called out two large eight-figure deals. Both of the deals were in North America, and hence the delta, I would say, relative difference in the growth rates. So not a whole lot to call out. On the sales attrition, I'd say, again, no specific trends that we'd like to call out. Just two comments I'll make on that front.
First is, as we kind of went through the consumption transformation, obviously, as we stand right now reflecting back, we feel really good with the overall sales capacity we have on board. The first half of the year, we've had strong hiring from a sales organization perspective. That just sets us well as we look ahead for the rest of the year. I'd say we have the right capacity in place to deliver on our forecasts.
Rudy Kessinger (Analyst)
OK, and then secondly, just on the model, I know you guys are guiding to just total subscription revenue. It sounds like platform is probably tough comparing second half versus first half, maybe down a bit in Q3. Do I have that right? And then secondly, on cloud revenue, just any directional color you can give in the sequential expectations in cloud, it sounds like with maybe some of the headwinds in June and July, maybe the sequential add in Q3 is a bit lower than Q2. Yeah, you want to take?
Rohan Sivaram (CFO)
I lost you a little bit on the first part of your question, but I'll answer it. If it's not covered, let me know, right? Well, when we look at the second half guidance, I'll say I called out two drivers. The first is around Confluent Platform. And in general, when you have these large deals, the timing of these deals matter. And the timing of these deals, not only from a renewal perspective, but also from an expansion perspective.
Alex Zukin (Analyst)
Renewals tend to be a lot more predictable than the expansions. And hence my comment around the second half with respect to platform. On the Confluent Cloud side, again, what we are essentially modeling out is we're thinking that the dynamics that we saw in the digital native customer segment will kind of continue into the second half.
Shane Xie (Head of Investor Relations)
So yeah, starting this quarter, we've obviously stopped guiding total revenue, and the focus will be on subscription revenue. But just to help with the transition from an overall modeling point perspective, we expect cloud mix for the full year as a percentage of total subscription revenue to be approximately 53% in that zip code. And that's very consistent with the mix that I shared earlier in the year. So not a whole lot has changed, but I just wanted to provide that color as you think about the second half, Rudy.
Rudy Kessinger (Analyst)
Yeah, that's helpful. And you hit on both parts. Thank you.
Shane Xie (Head of Investor Relations)
Thanks, Rudy. We'll take our next question from Alex Zukin with Wolfe Research, followed by Goldman.
Alex Zukin (Analyst)
Hey, guys. Thanks for taking the question. It's great to be on with you. I guess maybe just on the first one, on the digital natives, if you can kind of dig in a little bit there on maybe did you see was the moderation in consumption or some of the, I'd say, modulation of the green shoots, was that tied to a particular vertical, a particular geography? We heard Microsoft last night talk about how some consumption trends were less than they expected in Europe. So just curious if you could maybe categorize where some of those modulations came from?
Jay Kreps (CEO)
Yeah, there may be some geographical effect, but I don't think that's the main thing. We actually saw this primarily in the larger set of digital native customers globally, right? So US, EMEA, et cetera. What is it that they're doing? It's not the case that they're cutting projects. It's not the case that they're switching to competitors or there's any kind of customer loss in that segment. It really was about kind of optimizing their use of the underlying infrastructure.
Alex Zukin (Analyst)
So in a lot of these customers, they have a built-out team that's FinOps or something like that that's kind of scrutinizing costs and working with infrastructure teams to figure out, hey, can we consolidate clusters across teams? Can we compress data more? What can we do to kind of squeeze a little bit more juice out of each dollar of cloud spend?
It's not unique to Confluent, but we did see a little bit more of that activity. Any quarter in that segment is a balance of that kind of optimization and expansion. We think about last quarter, the balance was definitely on the expansion side. This quarter, a little bit more on the optimization side. That means when we think about the rest of the year, it could play out either way, but we're factoring in more risk when we think about what we'll get out of that segment.
Makes a ton of sense. And then maybe the second one, super helpful, by the way, on the cloud revenue component for the second half. But maybe given the fact that you're signing more customers, that you're landing maybe a little bit smaller initially, you've got a number of product vectors starting to gain traction with Flink and a few others.
And then presumably some of these Gen AI use cases ramping. Is there the possibility that even though the second half implied guidance for subscription for cloud, there is a bit of deceleration in that guide to account for the conservatism? How do we think about the sequencing of those tailwinds to maybe drive reacceleration? Again, I'll point to Microsoft. You told us second half of their fiscal year, they're going to see Azure reaccelerate.
Jay Kreps (CEO)
Yeah, yeah. Yeah, I mean, obviously, we've kind of given the guidance for this year. It would be, I think, early to start describing next year. But just kind of broadly or qualitatively, as we head into that, yeah, we feel like there's substantial tailwinds. And all the things you just mentioned, I think will start to contribute more materially. And that's definitely a positive factor for us.
Alex Zukin (Analyst)
And then when we just talk to customers, I think this is an area they continue to really bet on and invest in. And it's increasingly, for these larger customers, becoming a very substantive part of their overall data infrastructure. And so I think, yeah, that gives us a lot of confidence heading into next year about what the trajectory is.
Jay Kreps (CEO)
Perfect. Thank you, guys.
Shane Xie (Head of Investor Relations)
Thanks, Alex. We'll take our next question from Kash Rangan with Goldman Sachs, followed by J.P. Morgan.
Kash Rangan (Analyst)
Hey, thank you, guys. So Jay, the motivation behind the transition to consumption was to ultimately get better growth rate acceleration, which was the right thing to do. So as you are halfway through the mark of the year, what are the things that you uncovered as unlocks? What are the things that are working, things that still need work?
Alex Zukin (Analyst)
Because when you look at the, if it were to accelerate, which I think is definitely the goal of the company and certainly the right thing, we should be starting to see some leading indicators get better. And maybe it's not the right way to look at the subscription revenue growth rate or the cloud revenue growth rate, which have been slowing down a bit, right? Maybe due to a whole host of other reasons. Well, your plans are certainly very impressive.
So what needs to happen more for the transition to unlock the real growth potential of the company, which is the acceleration that you have planned for sometime in calendar 2025? And I have a follow-up question. Thank you so much.
Jay Kreps (CEO)
Yeah, yeah. I mean, obviously, when we look at the performance end of the year, there's a number of factors in play, including just the fact that we're changing a lot of things on the go-to-market. The first step is kind of get back to where you were in the new system. And I feel like, hey, we've kind of done that.
Alex Zukin (Analyst)
It doesn't mean that everything is dramatically better, but it means we're now oriented around consumption, tracking each of the individual workloads in each customer, tracking the actual spend by component and the opportunities around that, which is what will drive the DSP growth. So that's a good foundation to build on. So what is it that we need to do? We need to really harness that and lean into it as the driver of growth.
I think that'll be a positive tailwind for us as we kind of head into next year, as well as all the customers we've landed. The nice thing is, hey, as we think about halfway into this year, we actually have a much broader base to grow out of in terms of new lands. Provided we're able to continue that throughout the second half of the year, heading into next year, that's kind of a cohort to drive growth that will be kind of coming towards their more production maturity, additional use cases, et cetera. That'll be a tailwind.
Kash Rangan (Analyst)
One for Rohan. So as you have the net expansion rates, I mean, certainly they came down a bit. But broadly speaking, they're still OK relative to the overall growth rate of the company. Does it mean that you could get more profitably because generating growth within the base is less costly, less time-consuming than going for growth outside? Or are there other puts and takes that offset that consideration, netting out on where you stand in terms of profitability? And could net expansion rates be a driver of better profitability?
Rohan Sivaram (CFO)
Yeah, I mean, Kash, thanks for your question. Higher net expansion rates and net retention rates are obviously a driver of leverage over a long period of time. So there's no question about it. But when I really I'll probably take the question in two lenses. First, when you double-click into net retention, in addition to the current trends that you see, we spoke about data streaming platform.
Alex Zukin (Analyst)
And for customers that are using multi-products for us, their net retention rates are well above 130%. So as you think about the next wave of growth, next vector of growth, that's going to be a tailwind to net retention rates. So that's one piece of it.
The second piece of it around the margin side, I mean, when we think about our resource allocation philosophy, it is always making sure you get the right balance between growth and profitability and you're driving efficient growth over a long period of time. That's our goal. Given the opportunities that we have in 2025 and beyond with respect to the new growth vectors, we're being prudent about it.
When you look at the second half, we'll be adding the right amount of capacity in engineering, in sales, and marketing to make sure we are ready and we are in the right position from an overall capacity perspective as we embark on next year. I mean, most of the hiring we're going to do between now and end of the year is going to benefit us next year.
I'll finish off by saying our medium-term goals from an operating margin perspective, which is in the zip code of 5%-10%, continues to be the same. We are going to manage that proactively by balancing both growth and profitability.
Kash Rangan (Analyst)
Thanks so much.
Shane Xie (Head of Investor Relations)
Thank you. We'll take our next question from Pinjalim Bora with J.P. Morgan, followed by Oppenheimer.
Pinjalim Bora (Analyst)
Great. Hey, guys, thank you for taking the question. One question for you. I'll just blurt both of them out. On the Flink market, as you try to kind of execute towards capturing the Flink opportunity, how does the maturity and kind of the understanding of the use cases compare to that of the core market?
Alex Zukin (Analyst)
I'm trying to understand if you're finding customers that have well-defined Flink use cases and there's a pent-up demand versus the use cases are not as defined as that of the core Kafka market, and there's a little bit of evangelization that maybe needs to happen. That's the first question. Second question, how big is the digital native segment for you? Thank you.
Jay Kreps (CEO)
Yeah. Yeah, I would think about the Flink as being maybe as an open-source project, call it maybe three years behind Kafka broadly. And so in terms of maturity, that's probably where it is. Does that mean customers have use cases or don't have use cases? Well, it depends on the customer, right? Typically, the way these technologies progress is from more technologically sophisticated companies to less technologically sophisticated companies.
Alex Zukin (Analyst)
And so in maybe some of the banks, digital native companies, they would have pretty mature Flink practices, maybe not as large scale as their Kafka usage, but certainly up and running. You get out further into mainstream enterprise, you'd see some of that, but maybe less, right? And so I would think about that as being kind of the vector of diffusion and very similar to Kafka usage.
So I think it follows a very similar pattern, but several years earlier in that flow. For us, I think the project is a lot easier than the initial Kafka land. If you think about what we had to do getting into these new accounts with a cloud offering, especially early on, a lot of these customers did not have a lot of cloud products for infrastructure outside of the cloud provider.
So that kind of initial land was definitely the harder lift. As you think about, okay, great, now they've got these streams of data expansion to things that use the stream of data, I think it's definitely easier than kind of getting Kafka in the first time. And I think that's a positive force for getting that product to ramp quickly. And so we think that's a positive thing. And remind me of the second question you had.
Pinjalim Bora (Analyst)
Just the size of digital natives for.
Jay Kreps (CEO)
Oh, yeah, yeah. We haven't broken that out. If we think about the business overall, it's not the largest kind of vertical. It's not even that well-defined kind of broadly in the industry, but we have a definition internally that we tag accounts with. So it's not the largest vertical for us, but it is sizable.
Pinjalim Bora (Analyst)
Thank you.
Jay Kreps (CEO)
Yeah.
Shane Xie (Head of Investor Relations)
Thank you. We'll take our next question from Ittai Kidron with Oppenheimer, followed by Mizuho.
Ittai Kidron (Analyst)
Thanks, Shane. Rohan, first question to you. Can you tell me what % of your platform revenue is associated with cloud customers as well? I'm just trying to think because if there's a significant overlap there, I would think that over time, the lumpiness, inherent lumpiness in platform would fade. So we'll have to get color on that.
Rohan Sivaram (CFO)
Yeah, that's a great question. Well, we've not broken that out, but what I can tell you is there is an increasing set of customers who are using both platform and Confluent Cloud. And a couple of quarters back, we've been kind of also providing updates that when customers have both cloud and platform, they tend to be more sticky, and their NRRs are higher than the company average NRRs as well. So that's one piece of it.
Alex Zukin (Analyst)
But to your other point around if you have both platform and cloud, the lumpiness will come down. Probably not always because most of these times, you do have complementary use cases that are working in parallel. So when a time for a platform renewal actually comes, there is an event that happens that drives upfront revenue recognition. That doesn't change. So the lumpiness will still be there.
But I think to answer your question, yeah, there is a growing number of customers that are using both platform and cloud. And actually, that's exactly the direction we want to go because that makes our customers a lot more stickier.
Ittai Kidron (Analyst)
Hear it. Jay, second question for you. You've talked about how the quality of the new customer additions is higher. Maybe you can dig into that a little bit more. What does that mean? Is this a Global 2000 customers more weighted? Any color that would be great.
Jay Kreps (CEO)
Yeah, yeah. It's not exactly global 2000. We have a couple of mechanisms that we include in that, but it'd be broadly aligned to something like that. We would also include certain types of tech customers that might not be global 2000 but have high potential. So it's a defined customer list that we've set aside. So it's something like that where we feel like, hey, there's potential for high IT spend and therefore kind of bigger growth for us over time.
Ittai Kidron (Analyst)
Yeah, I appreciate it.
Shane Xie (Head of Investor Relations)
Thanks. We'll take our next question from Gregg Moskowitz with Mizuho, followed by Needham.
Gregg Moskowitz (Analyst)
Okay, thank you. Rohan, you did a nice job explaining some of the deliberation by the digital native customers and what you're assuming going forward. That said, you beat Q2 subscription revenue by about $7 million, guided in line for Q3, which implies a fairly meaningful Q4 reduction. So I did just want to ask if either of these large deals closed sooner than you had expected, or if there's anything you're seeing in the pipeline that has caused you to be a little more cautious with respect to the Q4 specifically.
Rohan Sivaram (CFO)
Yeah, Greg, I'll say that I'll go back to when we think about the second half guidance, obviously, these two drivers that I spoke about on the Confluent Platform side. In general, you have Confluent Platform large deals that are either renewals and expansions and new deals. For the renewals, it's fairly predictable. You have visibility 12 months down the road when that's going to happen.
Alex Zukin (Analyst)
And more often than not, it ends up kind of in that zip code. For the newer deals and for the expansion deals, that's where you could see deals kind of shift one quarter plus minus. And that's the dynamic that basically I called out for the second half guidance. And it's also important to notice that our backdrop, what continues to be a fairly volatile macro environment.
We want to make sure that we are thoughtful and prudent with respect to as we think about the outlook. On the Confluent Cloud side, again, not a whole lot of new things to share, but we are basically assuming that the dynamics that we saw towards the end of Q2 will kind of have an impact to the second half of the year.
Gregg Moskowitz (Analyst)
Thank you. And then just for Jay, can you walk through the rationale of switching Confluent Cloud pricing for your basic and standard clusters from base and partitions previously to elastic CKUs? And then also, does this change have any impact on revenue over the remainder of 2024?
Jay Kreps (CEO)
Yeah. We made a set of pricing changes. Those were some of them. There was also some tune-ups on some of the throughput-related pricing, some of the connector pricing. We talked a little bit about that a few earnings back as being kind of aligned with this consumption transformation. So in general, we were looking at things that cost friction early in the buying process.
Alex Zukin (Analyst)
We're kind of more aligned to a big upfront purchase than they were to a kind of use case by use case expansion. We wanted to pull some of that out. If you think about what we're trying to do broadly, there's a lot of open-source Kafka usage out there. We've monetized a small percentage of it, mostly sitting out there in open source. We want to go get it. There's a bunch of things we do to make that happen.
But part of what we want to do is really cover the spectrum of kind of TCO needs. There are very high volume, very price-sensitive workloads. There's very premium, high-quality workloads. You actually need some different SKUs to actually segment that market and go after it in the best way. So if you think about the additions to our portfolio, both the new offerings and Kafka for very large scale, some of the tuning at the low end, it really is about trying to properly segment that market and go capture it to kind of soak up more of the world's Kafka.
I think has proven successful, in part, as evidenced by the higher volume of customer lands. I mean, part of that's a kind of comp and sales practice thing, but part of it is also changes in the product that make it easier to land. Does it have impact on revenue? Yeah, it does have some. Whenever any of these things that kind of give customers more options, it does give them more opportunity to shift around. When we look at that, that effectively came out as expected.
So we had some model of like, hey, how much shift are we going to see? Typically, what you see when you make one of these changes is initially some shift that may be a little bit negative and then kind of better expansion as you have more options for customers to plug into. And so when we were checking ourselves against the model we built in terms of the impact of that, it was more or less exactly what we expected when you take it all together.
Of course, individual changes might be a little bit more or less. So the change on those SKUs was kind of part of that larger vision, which is like simplification, friction reduction, and just kind of soaking up all the Kafka. Is that helpful?
Gregg Moskowitz (Analyst)
Yep, super helpful. Thank you.
Rohan Sivaram (CFO)
Greg, I'll just add to what Jay said, that is, as we look at the second half of the year, some of these dynamics that he called out is essentially baked into our outlook for the year as well.
Gregg Moskowitz (Analyst)
Yeah. All right. Thank you, guys.
Shane Xie (Head of Investor Relations)
Thanks. Let's go to Mike Cikos with Needham. Mike?
Mike Cikos (Analyst)
Hey, thanks for taking the question, guys. But before I dig into mine, just to make sure we're clear on where we were going with Greg's last question there, as far as those pricing changes, that was contemplated when you guys provided us the guide for the full year two quarters ago. Or no, we're making changes now in real time off the Q2 result.
Jay Kreps (CEO)
Yeah, yeah. I think it was factored in early on. We had a model for what impact that would have, and we haven't seen it deviate from that. So when we think about like, hey, what are the factors impacting guidance in the second half? I think it's exactly what Rohan said, which is Confluent Platform lumpiness and what we've seen in the digital native segment. The price changes, yeah, we think it's net positive, but over time, but it's not a huge factor one way or the other right now.
Mike Cikos (Analyst)
Awesome. Awesome. So wanted to start with, I know the digital natives, we've highlighted that a lot. Is there any way to think about how their spend compares to other pieces or other segments within the Confluent customer base? Is it a 2X? What's the significance of the spend on a per customer basis versus the rest of the Confluent customer base?
Jay Kreps (CEO)
Yeah. Digital native itself spans from very small kind of tech companies to very large tech organizations with thousands of engineers. And so it'd be inaccurate to characterize that as a whole. So even internally, we break that into kind of large tech companies and smaller tech companies. What we saw was more concentrated on these large tech companies that are kind of more actively pursuing optimization. For those large customers, yeah, they do tend to operate at scale.
Alex Zukin (Analyst)
They have a lot of engineers. They tend to move quicker than the average customer in terms of adoption expansion, but also in terms of optimization. They're just a little bit more agile than a similarly sized company elsewhere in enterprise. That's probably broadly how I would characterize it. Their kind of use cases aren't vastly different from the rest of the world. They just tend to have a bit more data and somewhat more engineers than the average company that size.
Mike Cikos (Analyst)
Okay. And just thinking about, again, the quality of those new logos that we've highlighted a couple of times here as well. We're early days here, but is there any way to think about how the expansion of these new logos has played out versus what you've seen historically? And I guess the second part here for those new logos, does it skew more heavily towards digital natives, or is there any way to think about the composition of these new logos?
Jay Kreps (CEO)
Yeah. What we've shared is it's kind of broadly more concentrated in that high-propensity customer list, which you could think of as kind of Global 2000 plus high-propensity tech. I haven't looked at the tech composition specifically broken out, so I couldn't give you more color on that, but I'm not aware of that being a factor.
Alex Zukin (Analyst)
And then, yeah, we are catching these use cases earlier. It's just easier in this consumption motion for us to land. We're more flexible on the commitment amount. The product is better optimized for that. The comp is incentivizing it. And so, yeah, it is easier for us to get in early. But the key thing for us will be to kind of show that ramp of consumption over the course of the year with these new customers.
What we've seen so far is promising, both in terms of the numbers and that ramp time. I don't think it'll be exactly the same amount per customer, but the customer volume is much higher. And so we think overall that's a net positive contributor when we think about the base of new customers that we're acquiring.
Shane Xie (Head of Investor Relations)
Great. Thanks. A final question today will come from Michael Turrin with Wells Fargo.
Michael Turrin (Analyst)
Jay, great. Thanks for squeezing me in. Jay, between the go-to-market changes and the new product efforts, it seems like you're setting the stage for the next phase of Confluent. When do you think the foundation there is largely set? We go past talking about some of the transition impacts and more directly to some of the benefits from consumption focus and new products. Is that something we can start to build towards today? Is that end of year?
Jay Kreps (CEO)
Yeah. That's a good question. I mean, obviously, there's no moment in time where yesterday it was the old thing and today it's the new thing. But yeah, it's certainly the case that this year we've got a lot of moving parts that are part of kind of setting up where we want to get to, right? The consumption changes are part of that.
Alex Zukin (Analyst)
The new product introductions are part of that. As it becomes more of a tailwind for the business, I think it's definitely as we head into next year, that's where we have a more stable system that we're executing the go-to-market in that we've had more history with. And then we have products that have kind of gotten to enough scale of revenue where they can move the overall number.
Obviously, these newer offerings, you invest a lot of time in them, but the early results, you're just building that initial customer base. So it's relative to the amount of effort, you're not getting as many dollars out as you would with the core business. However, any of the big things start small. And so kind of getting those things seeded and started now,
I think, is very important, both in terms of what we want the business to do over the course of the next three years, but also in terms of what we want to be in customers, like the capabilities we want to provide them, setting ourselves up to really be a strategic platform in these early.