管理層發言
Good afternoon, and welcome to the Elastic Third Quarter Fiscal 2024 Earnings Results Conference Call. All participants will be in listen-only mode. After today's presentation, there will be an opportunity to ask questions. Please note, this event is being recorded. I would now like to turn the conference over to Anthony Luscri, Vice President of Investor Relations. Please go ahead.
Thank you. Good afternoon, and thank you for joining us on today's conference call to discuss Elastic's third quarter fiscal 2024 financial results. On the call, we have Ash Kulkarni, Chief Executive Officer; and Janesh Moorjani, Chief Financial Officer and Chief Operating Officer. Following their prepared remarks, we will take questions. Our press release was issued today after the close of market and is posted on our website. Slides, which are supplemental to the call, can also be found on the Elastic Investor Relations website at ir.elastic.co. Our discussion will include forward-looking statements, which may include predictions, estimates, or expectations regarding the demand for our products and solutions and our future revenue and other information. These forward-looking statements are based on factors currently known to us, speak only as of the date of this call, and are subject to risk and uncertainties that could cause actual results to differ materially.
We disclaim any obligation to update or revise these forward-looking statements unless required by law. Please refer to the risks and uncertainties included in the press release that we issued earlier today included in the slides posted on the Investor Relations website and those more fully described in our filings with the Securities and Exchange Commission. We will also discuss certain non-GAAP financial measures. Disclosures regarding non-GAAP measures, including reconciliations with the most comparable GAAP measures, can be found in the press release and slides. The webcast replay of this call will be available on our company website under the Investor Relations link. Our fourth quarter fiscal 2024 quiet period begins at the close of business on Tuesday, April 16, 2024. We will be participating in the Morgan Stanley Technology Meeting and Telecom Conference on March 6 and the Wells Fargo Software Symposium on April 10. With that, I'll turn it over to Ash.
Thank you, Anthony. Good day, everyone. And thank you for joining us on today's call. Elastic delivered another strong quarter. Our results reflect our continued execution, our ability to deliver differentiated product innovations, our maniacal focus on our customers, and our attention to managing the business with discipline. The trends we saw over the first half of the fiscal year continued to drive momentum in our business in Q3. First, customer interest in generative AI remains strong, and as customers become more educated about what it takes to build generative AI applications, they're increasingly able to appreciate the advantages of our platform relative to pure play vector database vendors. Second, we saw continued momentum in platform consolidation as customers chose the Elasticsearch platform to displace incumbent solutions for multiple use cases, particularly in log analytics and SIEM.
Lastly, we saw continued stability in cloud consumption patterns as customers increased their consumption against their previous commitments. These trends reinforce our confidence in the business and in our future growth as more companies choose our search analytics platform as a core part of their IT infrastructure stack for log analytics, SIEM, and Gen AI applications. I'll talk more about these trends and share some customer success examples driving our growth in a few minutes. First, let me start with a review of our financial performance. I'm pleased with our strong momentum and execution that drove our third quarter results. In Q3, revenue grew 19% year-over-year, with Elastic Cloud growing 29% year-over-year, driven by continued customer traction in the cloud, and consolidation onto the Elasticsearch Platform. We also exceeded our profitability guidance with a non-GAAP operating margin of 13%.
Turning to the broad trends we saw this quarter, as I met with customers around the globe, I saw a strong desire to leverage AI to improve business processes and elevate customer experiences. At the same time, companies remain focused on finding ways to reduce costs. In a sense, it's all about gaining efficiencies without sacrificing innovation. This is where Elastic shines. With the Elasticsearch platform, businesses can unlock the full potential of all their data to gain insights that drive innovation, optimize customer experiences, and inform strategic decisions. In the area of generative AI, customers are starting to advance their capabilities in leveraging retrieval-augmented generation or RAG patterns for building GenAI applications using existing and emerging language models. This is leading them to fully explore the total set of functionalities needed to build enterprise-class GenAI applications.
Advanced features like hybrid search, personalization, and re-ranking with reciprocal rank fusion, document-level permissions, security, and much more. All of these capabilities, in addition to our leading vector search functionality, and the fact that we are already the incumbent platform for data for tens of thousands of customers means that we are able to significantly lower the barrier for businesses to build their own generative AI applications. This is our asymmetric advantage, and as Generative AI and RAG are becoming better understood, our edge is becoming more pronounced. In the third quarter, we once again added several hundred customers using the Elasticsearch Relevance Engine or ESRE. ESRE allows customers to build Generative AI applications quickly without complicated and expensive model training. This is a big reason why customers turn to us for Generative AI. One example of this is our long-time customer Consensus, an AI-powered search engine that aggregates and distills insights from more than 200 million peer-reviewed papers from the semantic scholar database.
They use Elastic Cloud for advanced artificial intelligence, semantic, and text search. Consensus 2.0 is powered by ELSER. Their users have benefited from increased accuracy and relevance of search results and new GenAI features such as a summarization of the top 10 studies based on integration with OpenAI's GPT-4 model. In Q3, we also expanded our business with Stack Overflow, the largest and most trusted online community for developers, which replaced its previous vector database with Elastic. The company chose Elastic Cloud for our scalability, differentiated functionality, and integration across the ecosystem. Stack Overflow is leveraging Elastic's vector and semantic search capabilities to deliver a more human-like generative AI-powered question-and-answer experience to developers via its Overflow AI product. We also continue to see customers choose our platform because they expect to use our ESRE functionality in the future.
In other words, they see us as the right choice for their needs today as well as for the future. As an example, a cloud-based document management company for legal professionals signed a multi-year deal with Elastic this quarter. With over 6 billion searchable documents, the company chose Elastic for our innovations, performance, and ability to power search across their entire platform. Previously, they used a competitor's offering, but they moved to Elastic because of our market leadership and to modernize the search experience to align with their customers' expectations by leveraging Elastic's vector search capability to enhance the user experience. Beyond generative AI, we noted that customers are continuing to displace incumbents and consolidate onto the Elastic Platform for observability and security, which is the second trend I talked about earlier. The Elastic observability and security solutions are built on top of our Search Analytics Platform, with Elasticsearch as the foundation.
Our platform has the unparalleled ability to quickly ingest all kinds of data, index it at extreme speed and petabyte scale, and make it searchable, allowing for all kinds of advanced correlations, AI-powered searches, and ad hoc analytics. This gives customers the ability to centralize onto a single platform for all use cases that require this kind of advanced search and analytics. This is a big reason for our success in displacing incumbent solutions, especially around log analytics and SIEM. Our recent innovations, including our AI assistance for observability and security, are making it even more compelling for customers to move to our platform. This quarter, we closed several multi-million dollar deals where we displaced incumbent solutions for observability and security. Our value and our innovation resonate across industries as we continue helping customers save on their overall IT spend while helping them gain even greater value from our innovations.
As an example, a top U.S.-based bank chose Elastic as a desired SIM platform to replace their legacy vendor. The drivers behind the selection included platform speed, performance, scale, and Elastic's rich set of security features. They also chose Elastic for our innovative storage-tearing searchable snapshots capability, which allows the bank to retain important security logs for longer regulatory retention periods at dramatically reduced storage costs. We also expanded business with the U.S. state agency for elastic security. The agency initially started with Elastic for search and then in Q3 chose Elastic to replace their legacy security vendor for SIEM, enabling them to increase support for their security operations center. The scope of Elastic's ingest capabilities, search speed, platform functionality, stability, and reporting capabilities allows the agency to gain and maintain visibility across their environment and multiple agency branches and correlate alerts across a broad security tool set.
Now, turning to products and innovation, we continue to invest in capabilities that make it possible for customers to migrate easily from incumbent solutions to Elastic. On our last call, I mentioned the launch of our powerful new piped query language, Elasticsearch Query Language or ESQL. Even as we continue to enhance this new capability, we have seen tremendous interest from our customers with approximately a thousand customers trying it out since we launched in November. Our latest release of Elastic 8.12 in January also included several key enhancements, including scalar quantization, which is a vector search optimization technique that delivers performance improvements across several critical metrics. It dramatically improves enterprise-level scalability, allowing customers to store four times as many vectors in RAM while also improving overall indexing speed and reducing query latency by up to two times.
Additional enhancements to search concurrency resulted in speed ups across a very broad set of query types, including vector similarity and full-text search, as well as a 40% reduction in latency for analytical queries. In Elastic Observability, we brought several capabilities to general availability, including the enhanced Elastic AI assistant for observability, SLO monitoring, and mobile APM support based on open telemetry. In Elastic Security, we continue to enhance our market-leading Elastic AI Assistant for security, introducing real-time alerts with natural language interactions that enhance the way security analysts approach alert triaging. This results in a more efficient and effective security operation that is adept at navigating the complex cybersecurity environment. In cloud security, we added support for Microsoft Azure, in addition to our existing support for AWS and Google Cloud for cloud security posture management.
On the go-to-market front, we drove continued momentum in Q3 through our participation in AWS re:Invent and our ElasticON conference events that allowed us to connect with thousands of customers, prospects, and users worldwide. We have now concluded six of these ElasticON events and will be holding another six in our fiscal Q4. And as a final, but important highlight for the quarter, I was pleased to announce that Mark Dodds joined Elastic as our new Chief Revenue Officer. Mark leads all of Elastic's customer-facing functions, including global sales, customer success, solutions architecture, professional services, ecosystem and partnerships, and sales operations. Mark brings extensive go-to-market experience to our executive team as we continue to drive momentum in Generative AI and growth across all segments of our business. I'm excited about everything that Mark will do for us in the coming years as we push to capture the full opportunity ahead of us as a company.
In closing, I'm very pleased with how we performed this quarter. The innovations we are driving into our Search Analytics Platform, the momentum we are gaining around Generative AI, and the traction we are seeing in customers consolidating onto our platform is continuing to increase our confidence in our future. We see a future where our Search Analytics Platform becomes a core part of every IT infrastructure stack for gaining insights from all data, especially in the context of generative AI. I want to thank the Elastic team for their resolute focus and the entire Elastic community for their unwavering commitment. Now, with that, let me turn the call over to Janesh.
Thanks, Ash. We are pleased that we delivered a strong quarter driven by our consistent focus on execution. We once again came in above the high end of our guidance for the quarter, both on the top line and the bottom line. The third quarter largely played out as expected. We delivered 19% year-over-year growth in total revenue, with Elastic Cloud growing 29% year-over-year. We continued our focus on profitability, delivering a non-GAAP operating margin of 13%. Our performance in the quarter reflects both our success in the market and continued investment discipline. We executed well to deliver strong growth in both self-managed and Elastic Cloud revenue. As in prior quarters, we saw a number of customers consolidate onto the Elastic Platform to lower their total spend without sacrificing innovation. We also saw healthy cloud consumption growth as customers continued to consume against their contractual commitments.
As Ash mentioned, we continue to see strong customer interest around Generative AI use cases. Customers express a strong desire to use ESRE to build Gen.AI applications since it provides a comprehensive and enterprise-ready platform to ground large language models with their private business context. While it will take some time for Gen.AI spend to become a significant driver of our revenue, we continue to expect that Gen.AI will present a meaningful revenue opportunity for us in the longer term. Let's get into the results for Q3 and our outlook. Total revenue in the third quarter was $328 million, up 19% year-over-year or 18% year-over-year on a constant currency basis. Subscription revenue in the third quarter totaled $308 million, up 20% year-over-year or 19% year-over-year in constant currency and comprised 94% of total revenue. Within subscriptions, revenue from Elastic Cloud was $143 million, growing 29% year-over-year on an as-reported and constant currency basis, reflecting the trends I mentioned previously.
Elastic Cloud represented 44% of total revenue in the quarter, up from 40% a year ago. Elastic Cloud revenue derived from month-to-month arrangements contributed 14% of total revenue. Professional services revenue in the third quarter was $20 million, up 7% year-over-year on an as-reported basis and 6% year-over-year on a constant currency basis. As we've said before, professional services revenue may fluctuate across quarters based on the timing of service delivery, and we do not expect it to vary significantly in mix over time. To add more context around overall deal flow, EMEA grew the fastest during the quarter, followed by the Americas and APJ. We continue to see a healthy balance across the business based on geography, solutions, and verticals, and this diversification reflects the breadth and popularity of our platform. Moving on to customer metrics. We ended the quarter with over 1,270 customers with annual contract values more than $100,000.
Looking at customer additions more broadly, we ended the quarter with over 4,290 customers above $10,000 in ACV and approximately 20,800 total subscription customers. We are pleased with this quarter's customer additions in the greater than 100K category, which reflects our focus on the highest value customers, who are the primary driver of the business. Our net expansion rate, which as you know, is a trailing 12-month lagging indicator, was approximately 109% consistent with Q2 and in line with our expectation for the quarter. Now turning to profitability for which I'll discuss non-GAAP measures. Our non-GAAP results for the third quarter exclude, among other items, a discrete income tax benefit of $207 million from the release of a valuation allowance for GAAP purposes. This did not impact any of our operating results, non-GAAP earnings per share, adjusted free cash flow or cash and cash equivalents.
Gross margin in the quarter was 77% versus 76.8% in the prior quarter, reflecting a slightly higher subscription mix. Our operating margin in the quarter was 13%, which was better than expected. The strong operating margin performance was driven by revenue outperformance and our continued focus on managing expenses as we invest with discipline to drive future growth. Diluted earnings per share in the third quarter was $0.36. Our free cash flow on an adjusted basis was $63 million in the quarter or a 19% adjusted free cash flow margin. For the full fiscal year, there is no change in our prior outlook, and we continue to expect free cash flow margin on an adjusted basis for fiscal '24 to be slightly above the non-GAAP operating margin for fiscal '24. Turning to guidance. While we are very pleased with our execution and our outperformance thus far in FY '24, our guidance philosophy remains unchanged.
We continue to be prudent in the near term. Despite the many moving parts in the broader macro climate, we anticipate that business conditions will remain largely unchanged. In addition, we believe it is appropriate to anticipate that consumption patterns may fluctuate in the near term. With that background, for the fourth quarter of fiscal '24, we expect total revenue in the range of $328 million to $330 million, representing 18% year-over-year growth at the midpoint on an as-reported and constant currency basis. We expect non-GAAP operating margin for the fourth quarter of fiscal '24 in the range of 7.4% to 7.8% and non-GAAP earnings per share in the range of $0.18 to $0.20 using between 105.5 million and 106.5 million diluted weighted average ordinary shares outstanding. For full fiscal '24, we are raising our outlook and now expect total revenue in the range of $1.260 billion to $1.262 billion, representing 18% year-over-year growth at the midpoint or 17% on a constant currency basis.
We expect non-GAAP operating margin for the full fiscal '24 to be approximately 11% and non-GAAP earnings per share in the range of $1.15 to $1.18 using between 103 million and 104 million diluted weighted average ordinary shares outstanding. Finally, looking ahead to fiscal '25, we're in the middle of our planning process. While it is too early to provide specifics, especially around revenue outlook, I will share our thoughts on how we are approaching next year from an investment perspective. We've already demonstrated that our business model has inherent operating leverage, and this year we've expanded non-GAAP operating margin even faster than we had initially expected. As we go through our planning process, we expect to balance investing for long-term growth against near-term profitability. Given the significant opportunity we see in Generative AI, we are planning to accelerate investments in fiscal '25 to help capture this large and expanding market.
We expect these investments to be broad-based across the business and are intended to drive meaningful revenue growth over the longer term. Therefore, we expect to expand non-GAAP operating margin only modestly in fiscal '25 relative to our current non-GAAP operating margin guidance for fiscal '24. We expect these investments are likely to be more front-end loaded in fiscal '25. We will provide more formal guidance around fiscal '25 on our Q4 earnings call. In summary, we are pleased with our strong performance so far in fiscal '24 and are confident in our outlook for the fourth quarter. And with that, let's go ahead and take questions.
分析師問答
We will now start the question-and-answer session. Our first question will come from Pinjalim Bora of JPMorgan. Please proceed.
Thank you for taking my question and congratulations on the quarter. I would like to focus on the Elastic Cloud aspect. Janesh, it seems this year we are experiencing a different pattern in seasonality. Q2 showed a significant increase compared to Q1, which is usually the reverse, and now Q3 appears to have seen a decline from Q2, which again is typically the opposite. Was there any one-time event in Q2 that didn't occur in Q3, or did you observe a lower consumption rate than last quarter? It seems like the 14% monthly rate may have introduced some variability as a challenge. Any insights on this would be appreciated, and how do you view the sustainability of that growth at 29% to 30% for total cloud?
Hey, Pinjalim. Yes, let me unpack that for you. So, you know, first off, if I just step back, we were actually very pleased with our overall performance in the quarter, including on cloud coming in at 29% year-over-year growth. Broadly speaking, I'd say it played out as we expected it would. If I think about the pieces in that, first-off, there was nothing unusual or one-time in Q2 that didn't recur in Q3 to just take that off the table. But if I think about the performance during Q3, as we look at our performance across the segments, we did really well in both commercial and enterprise, which as you know is our sales led motion and then we drive the consumption after we've secured the commitments. And then on the SMB side, which is predominantly in the self-serve motion, that piece remained soft, and that's been soft for a little bit of time now. So even that played out as we expected it would, but it was basically consistent with where it's been.
And so just given the relative growth in monthly cloud versus the rest of our cloud revenue, that's what drove monthly cloud to 14% of revenue versus 15% in the prior quarter or 16% in the year ago period. So I'd say broadly speaking, our sales motions are actually working quite well, both in terms of securing the initial commitments and then of course working with those customers to ramp their consumption. If I think about the seasonality and all the different elements you called out, in fact, in a way that's just reflective of the nature of consumption, which as we've said can fluctuate from time to time. If I think about consumption patterns across the quarter, though I'd say Q3 was generally stable across the board, as we mentioned earlier. And so overall, when I look back at all the moving parts, we were actually quite pleased with our cloud performance and we look forward to delivering a strong Q4.
Understood. Just a one follow-up, a little bit technical question for Ash. How should we think about the future of RAG in the context of the expanding context windows for tokens for LLMs? I'm thinking, does that latter kind of eat into some of the use cases for RAG or is cost going to be basically the primary driver of how people kind of architect these solutions?
Thank you for the question, Pinjalim. We are very confident, and our customers are confirming this, that Retrieval-Augmented Generation, or RAG, is a pattern that not only will continue but is expected to become the leading approach moving forward. The main reasons for this are twofold. First, costs increase as more tokens are sent to context windows. As companies develop real-world enterprise applications, this cost is a significant concern, and RAG helps manage those expenses. Second, when training or refining a model, there comes a point where training stops and the model is deployed. However, internal and proprietary data is constantly evolving, so to provide contextual and accurate responses, RAG is essential. These are why RAG is becoming the dominant method for building enterprise applications, and we are very positive about our role in this area.
Very clear. Thank you very much.
The next question comes from Brent Thill of Jefferies. Please go ahead.
Are you there?
Hey, Brent, are you there?
Operator, let’s move to the next caller.
Okay, the next question comes from Tyler Radke of Citi. Please go ahead.
Yes, thanks for taking the question. Hopefully…
Hey, Tyler, you're breaking up. We can't hear you.
Okay, how about now? Is that better?
Yes, yes it's better. Go ahead please.
Okay great. Janesh, can you just frame for us how we should be thinking about the moving pieces in Q4. Q3 saw stronger other subscription growth. How should we think about the relative growth between other subscription and Elastic Cloud? And then secondly, I apologize if I missed it, should we think about Q4 exit growth as the right top line growth for FY '25 to the starting point, realize you're not pinning guidance, but just any color on that would be helpful. Thank you.
Yes, Tyler, I'll cover all of those. So first off, I think about the other subscription of the self-managed business, and the strength that we saw there in the quarter. Maybe just to start with, I'll just remind folks that whether a customer chooses cloud or self-manage really depends on their own posture and deployment format. We've seen good momentum on both. And while we prefer cloud, and we think it's better for the customer and better for us, it is ultimately the customer's choice. So the mix of business is generally partly a function of deal flow in the quarter. And importantly, we did not see any kind of shift from customers trying to move from cloud to self-managed. I think it's just more a function of the deals that we saw across the geos and segments and timing of renewals and so forth. We had also shared on the prior call already that we actually expected to see an increase in self-managed revenue this quarter.
So it was not a surprise to us. In fact, in Q3 in the year ago period, we saw a similar pattern where that self-managed business was a little bit strong. And so if I look at it even in terms of year-over-year growth, I think self-managed grew 13% year-over-year, compared to 11% in the prior couple of quarters. So it really wasn't a big deviation. And so as I think about then how that's factored into our guidance for Q4. One way to think about it is the guidance for Q4 sequentially is only $1 million higher than what we reported here for Q3. So there's just not a lot of moving parts in there, so I don't expect much movement. We obviously don't guide to the individual pieces around cloud and self-managed, but that's roughly how I think about it. And just keep in mind that in Q4, we've got two fewer days because we have 90 days versus 92 days that we typically have in Q1 to Q3. And so that presents a little bit of a headwind of, call it, roughly $6 million to $7 million of revenue in Q4.
And that's also what we've already factored into the guidance. And that's just something to keep in mind as you think about sequential growth rates. And then finally, in terms of thinking about growth for fiscal '25. Look, I think it's just too early at this stage, right. We've seen really strong performance from the business in '24 so far. We've been very pleased with our execution. We're still in the middle of the planning process for '25. So it's too early to have a concrete view on the revenue outlook. For now, we're just focused on delivering a strong Q4, and we will talk more about '25 on the next call.
Thank you. To follow up on Ash's question about the customer example that replaced the vector database, I'm trying to recall the name of the coding repository company. Can you discuss the decision-making process behind that? Vectors are still quite new in this emerging industry. How often do these kinds of replacements occur, and what were the key factors that led them to look for alternatives and ultimately choose Elastic?
Thank you, Tyler. As customers become more knowledgeable about building Generative AI applications, our core differentiators are increasingly evident to them. In discussions with customers, including recent examples like Stack Overflow and Consensus, four key aspects stand out. First, our vector database implementation, built on the SIEM, is exceptionally well-executed, particularly in scalability. Innovations like scaler quantization enhance performance and reduce memory usage. Second, the enterprise capabilities we’ve integrated, such as document-level permissions and advanced security features, are crucial for building enterprise applications, along with functionalities like personalization and geolocation. Third, hybrid search, which combines vector, semantic, and lexical search with re-ranking, is gaining traction as customers seek relevant answers and recognize the importance of these techniques. Finally, our established presence in the ecosystem, with tens of thousands of customers on the Elasticsearch platform, facilitates quick data integration, allowing clients to create vector embeddings efficiently for their Generative AI applications. These four differentiators are increasingly solidifying our competitive edge, and I believe they will become even more significant as the industry evolves.
Thank you.
The next question comes from Matt Hedberg of RBC Capital Markets. Please go ahead.
Great. Thanks for taking my questions. Congrats on the results, guys. Ash, obviously, a large firewall vendor talked about the plan to bundle offerings, including free trials. And I'm wondering if you think that eventually shows up as industry pricing pressures and things like logs or SIEM?
Yes. Thanks, Matt. So we haven't seen any change in the competitive dynamics. But what I will say is the whole idea of platform consolidation. As I talk to customers, what I'm seeing is that every CIO that I'm meeting is still very conscious about their overall budgets. So budget pressures are continuing, and there is tremendous excitement in terms of wanting to do more around Generative AI. So that's becoming a bit of a double bay, and that's incentivizing customers to want to do more on a few platforms that they trust, that they see as being innovative, that they see as having the ability to not only help them today but also future-proof their direction. And we are one of the beneficiaries of that. The fact that we can deliver tremendous innovation. We have been delivering innovations like Generative AI, our AI assistance for observability and security and doing it with our pricing model that makes it possible for customers to both consolidate and do that at a much better overall value to investment is something that's enabling us to win and displace incumbents. So that consolidation motion, we are leaning into it. We have been leaning into it for several quarters now, and we are seeing the benefits of it.
It's great to hear. There has been a lot of discussion about the use of vector and Generative AI, particularly in enterprise search or application and site search. I'm interested in how you see vector technology being integrated into broader observability or security search with Security Information and Event Management.
That's a great question. And although we very strongly believe that Generative AI and vector search and so on is going to materially increase the overall addressable market size for search in the long term. We are also seeing our ability to differentiate and compete better in the areas of observability and security. So just to give you some sense of how that plays out typically in terms of security, we've gone through several ElasticON events now. And the thing that consistently gets everybody super excited in the audience when we are talking to security customers is our security AI assistant. The fact that they can now see real-time alerts and really interrogate the system and ask questions about how to deal with those alerts and the AI assistant is able to help guide them through all of that, that is incredibly powerful because it reduces the bar on what a security analyst needs to know, the level of experience and expertise they need to possess. So that is making it that much easier for us to compete and win in these deals and displace incumbents. So we feel it's giving us that edge already. I expect that edge to only continue growing in the future. And that's the reason why we are so excited about everything that we're doing around Generative AI.
Great to hear. Thanks, guys.
The next question comes from Raimo Lenschow of Barclays. Please go ahead.
Hey, thank you. Congrats from me as well. Two quick questions. One, Ash, with the change in leadership now on the go-to-market side, are there any kind of changes you're seeing on the journey to kind of evolve go to market, but any changes we should anticipate as we're kind of getting ready for the new year? And then I had one follow-up for Janesh.
Absolutely. We are very excited about Mark's addition. Even in the short time he has been here, he has been a wonderful partner, and it’s impressive to see how he has truly embraced the role. He is collaborating closely with team members and demonstrating excellent leadership qualities. Our focus on the enterprise and commercial segments and our overall go-to-market strategy will remain the same, as we are seeing a lot of success in that area. Janesh mentioned this earlier, and we are thrilled about the expertise Mark brings in these domains. As the business continues to thrive in areas like Generative AI and platform consolidation, Mark's skills will only enhance our performance. At this moment, I do not anticipate any changes. I believe that the progress we have made and the direction we are heading in will benefit from Mark's contributions rather than require any alterations.
Okay, perfect. Yes, makes sense. And Janesh, for you, thanks for the initial look into '25, and I know you're not guiding yet. But if I break down the investment areas a little bit further, how should we think about it? Is that kind of a mixture between sales and you need to kind of think about building sales capacity, if you think about recovery, so you need more sales guys and they take time to get productive? And then on R&D, like there's obviously a lot of stuff going on with a lot of changes in the industry with AI. So is it just like across the board? Or are there kind of focus areas we could think about? Thank you and congrats again.
Thanks, Raimo. So as I think about fiscal '25, we're going to continue to obviously focus on growing profitably and specifically, we'll grow our revenue faster than our expenses. But as we are going through the planning process now, we do expect that we will balance investing for long-term growth against near-term profitability. And that's why I called out the relatively modest op margin expansion that we expect for next year. So over the course of this past year, we've seen a pretty significant opportunity ahead of us in terms of the increasing importance of GenAI and the massive opportunity that we can go capture. And so we really do plan to accelerate our investments in this area to go capture that market opportunity. And I think the investments will actually be broad-based across the business. We'll want to fund engineering. There will clearly be some go-to-market investments and support capabilities in order to unlock this opportunity and really execute well in the longer term.
But we do expect that these investments, Raimo, will support our longer-term growth trajectory. And that's why I indicated that we will be making the investments now, and that will have a little bit of an impact on non-GAAP op margin for next year. And the investments will also be a little bit more front-end loaded in fiscal '25. And we'll provide a little bit more color and formal guidance on that when we get to the next earnings call.
The next question comes from Andrew Nowinski of Wells Fargo. Please go ahead.
Thank you. Good afternoon. I have two questions. First one I want to ask was on ESQL. So you talked about one of the advantages of Elastic's vector search is that you already have a lot of the customer data on your platform. Do you think ESQL could be a tool for bringing more data onto your platform if it reduces the friction from leaving Splunk?
Yes. That's a great question, Andrew. And like I mentioned in the prepared remarks, I'm super excited about ESQL. We already have around 1,000 customers trying it out and it hasn't been that long since we introduced it. What it does is it makes us even more developer-friendly. In terms of how the language is designed, it's a very easy to use language. It's got the piping semantics, allows for iterative development and developers love it. And the other thing that it does is it makes it very easy for customers to move from incumbent legacy SIEM solutions onto our platform because some of those solutions have had similar capabilities with these kinds of pipe semantics and so on. And now that barrier is gone. Customers are able to quickly take that. And when you take ESQL and add to it our AI assistance, which are able to map other languages onto ESQL, it just makes that entire process very easy. And that's one of the reasons why I feel so excited about our continued ability to take share and make it possible for customers to switch from incumbent solutions onto Elastic, consolidate onto our platform and grow. So in the long term, I definitely believe this is going to bring more data onto our platform and allow us to become that much more critical to our customers' overall infrastructure.
Okay, thank you. That was great. Just my one quick follow-up question. Given all the functionality that you've added to the platform now with ESRE and ELSER and ESQL, et cetera, have you changed your pricing to account for all these new capabilities?
Yes. One of the aspects, Andrew, is the pricing structure that includes multiple tiers. Depending on which tier a customer is in, they can access different features. For instance, our top tier offers AI assistance, while the platinum tier includes some ESRE functionalities. Customers need to upgrade to a higher tier to access these additional capabilities. We periodically adjust prices for both self-managed and cloud services, but we only do this when it aligns with the enhanced functionality we provide. Let me check if Janesh has anything to add.
No, nothing, Ash. I think you covered it quite nicely.
Okay, thanks, guys.
The next question comes from Patrick Colville of Scotiabank. Please go ahead.
Alright, thank you for having me on the call and really great to be part of the Elastic story. I guess I want to ask about revenue; 29% constant currency in 3Q is pretty terrific. You called out some drivers; Gen AI was one, observability and security consolidation, and then consumption coming back to committed levels. I guess my question for you, Ash, is can you just rank order those three drivers you called out, which in 3Q was the most important.
Thank you for the question. Regarding Generative AI, it is definitely contributing to our revenue, but considering our overall size, its impact is still in the early stages. We are seeing consolidation on our platform, with customers moving away from incumbent solutions and opting for our observability and security services. This trend has been ongoing for several quarters and typically, it takes time for these new workloads to ramp up. Given that we've been focusing on this for a while, it has significantly supported our business growth, and the stability in consumption also plays a role. To summarize, the consolidation is likely the most significant factor, while stable cloud consumption helps balance everything out. Janesh, do you have anything you would like to add?
Yes, I'd agree with that ranking, Ash.
Thanks, Ash. And I guess, Janesh, my follow-up is for you, please. So 3Q non-GAAP op margins of 13%, nothing to snip out at all. Based on my imbalance, some investors nitpicking, but look, that's only the same as 2Q levels and guidance indicates a leg down in 4Q. Can you just double-click on what's going on in terms of profitability? Are there one-offs to call out in 3Q? Why this lag down in 4Q?
Yes. So happy to cover that. If I think about the change from the third quarter to the fourth quarter, we are obviously pleased with the outperformance that we delivered in Q3. But in the fourth quarter, one of the things that we had called out last time is that we have approximately $12 million higher in seasonal expenses in Q4. I had mentioned this on the prior call; this relates to the timing of employee benefits and our engineering all-hands events. And then also, as I mentioned a little bit earlier on this call, since we have two fewer days in the quarter, we've got a natural headwind of approximately $6 million to $7 million on revenue. So those two factors come together, and that's why Q4 is lower. The seasonal items affect only Q4. And so from our perspective, we really focus on managing the full-year number. And as you saw here, we've raised our outlook for the full year. We had raised it even earlier in the year, and we're very pleased that we can outperform even against our prior commitments, and we expect to finish the year strong.
Alright, nice to be part of the story, and thank you for having me on.
Thank you.
The next question comes from Brent Thill of Jefferies. Please go ahead.
Janesh, I know RPO bounces around, but it did decel anything underneath the surface that we should be aware of?
Hey, Brent, so overall, I'd say we are actually very pleased with the annual contract selling motion, including how sales successfully delivered a very strong Q3. We've talked about some of the trends in the business that we saw including platform consolidation and so forth, and all of those reflect that success. If I think about RPO, maybe the one thing I'd call out is that contract lengths were similar to what we've previously seen in the business at roughly 1.5 years, but they were slightly shorter than the year ago period. And so RPO can vary a little bit based on the terms of the deals that we signed in a consumption business. You've also got variations of customers trying to match the timing of their renewals and expansions with how quickly they are consuming. So tracking it on a quarter-to-quarter basis can be a little bit noisy. Nothing else I would actually call out. Fundamentally, RPO growth was in line with revenue growth and all the other measures. And was generally in line with the rest of the business. And as we've shared before, we look at revenue as really the primary indicator of the success of our business, and we are very pleased with the outcome there.
And then for Ash on AI, some of these early wins, you mentioned it's still obviously really early. But are there any common threads you pull in terms of verticals or geography or scope? What are the commonalities that you're starting to see? I guess is there anything you can do to kind of step on the gas in terms of awareness, and you've seen others start boot camps and do smaller deployments, and it seems to be resonating with others in the industry. Anything you're learning and taking and pushing forward as you go through this AI wave?
Thank you for the question. We are definitely observing a broad range of patterns across different industries. I've provided some examples in my prepared remarks. For instance, companies like Stack Overflow and Cisco, as well as those in the document management sector like Consensus, which focuses on scientific research, show this trend. Many use cases are still oriented towards internal operations, as our customers gain more confidence in managing the risk of hallucinations. The RAG pattern that utilizes Elastic helps in minimizing these hallucinations, but customers are still working towards fully integrating this. This reflects a second pattern of maintaining a lot of internal-facing staff, which will eventually evolve. Regarding our approach and how we engage with customers, I've mentioned in the last three quarters that we are adding several hundred new customers each quarter who are leveraging us for vector search applications, including storing dense vectors and conducting vector similarity searches.
I'm excited about the growing adoption of our ELSER Sparse Encoding model. We are driving this progress through our sales organization and marketing initiatives. We’ve completed six of our twelve ElasticON Events this fiscal year, with six more scheduled for the fourth quarter, where GenAI is a key focus. We are conducting boot camps and workshops with our customers, which have proven to be very successful. While we are in the early stages, I believe this will lead to a significant increase in our total addressable market in the long run.
Thank you.
The next question comes from Andrew Sherman of TD Cowen. Please go ahead.
Oh, great. Thanks. Hey, guys. Janesh, the annual cloud number seemed pretty strong and stable at mid-40s growth. So is it just the SMB pressure that's dragging down the overall number? What are you expecting in that segment going forward?
Yes, I believe the calculations you are making are quite clear. You are correct that we have been very satisfied with our annual cloud sales strategy, which relies heavily on sales efforts. We are pleased with the commitments we have secured, as Ash has mentioned, as well as the consumption aligned with those commitments. Regarding the small and medium-sized business segment, as I noted earlier, it remains weak and is consistent with previous trends. Looking ahead, while I won’t dive into segment-specific details about our cloud business, overall, we remain very enthusiastic about the opportunities in this area and anticipate it will continue to grow at a faster rate than the rest of the business.
And then, Ash, are you seeing anything yet in the pipeline potentially more customers reaching out to you because of potential disruption from your largest logs competitor?
Thank you for your question, Andrew. There isn't any specific disruption or change I want to highlight. However, our ongoing effort to encourage customers and prospects to move onto our platform by replacing existing solutions is something we've been pursuing for some time. We see this as a significant opportunity. Our strengths in log analytics, SIEM, and search enable us to establish a foothold, and once we are in, we can expand our offerings even further. The current market dynamics, alongside our innovations such as ESQL capabilities and the Frozen Tier, which greatly benefits our customers, as well as our AI advancements in observability and security, are all major reasons why customers feel confident and comfortable switching to our platform, which is very exciting.
Great, thanks guys.
The next question comes from Koji Akida of Bank of America. Please go ahead.
Hey guys, thanks for taking the questions. In the past, when we think about the business in terms of enterprise search, security, and observability, I mean in the past, you've talked about ways to think about the growth rates between those three categories. So wondering if you could maybe get a little granular here and let us know how you're thinking about growth rates between these three categories over the medium term to think about kind of your medium-term growth aspirations?
Koji, I'll start off and then let Ash add more. When I consider the performance across our solution areas, Q3 was in line with what we've typically observed over the past year regarding annual contract value. There are slight variations based on deal flow and the timing of renewal contracts. Looking at the medium term, all three solution areas are experiencing solid growth. Any potential shift in the mix will take time to manifest, so we aren't currently looking to change it. Regarding the future, we believe search will benefit from GenAI, which we've mentioned as an expansion of the total addressable market, offering us a significant opportunity. Additionally, we see considerable potential for growth in security and observability, driven by the consolidation trend and the strength of the GenAI capabilities we've integrated into our platform. These capabilities positively impact both security and observability, especially with the AI assistance we've introduced and discussed previously. That's my perspective on the medium term. Ash, do you have anything to add?
No, I think you nailed it, Janesh. Nothing more to add.
Great, guys, thanks so much. Thank you.
Thanks, Koji.
The next question comes from Brad Reback of Stifel. Please go ahead.
Thanks very much. Janesh, as part of these investments that you're talking about for next year and beyond, do you need to make COGS investments to support all this AI functionality?
Brad, as I think about our investment profile for the future, we will see a natural increase in COGS. I think that's more a function of the cloud growth that we see. We're not expecting to see anything structurally significantly different. That said, as we launch our serverless offering over the coming months. There may be some initial investments that we have related to that, but we will factor that into the model when we provide guidance more formally for fiscal '25. But there's nothing major I would call out at this stage.
Janesh, if I might just quickly add to that. One of the things that I want to make sure that everybody understands is, we don't need things like GPUs for us to be able to deliver the RAG functionality. We run our capabilities very, very efficiently on CPUs, and we have instance types in the cloud that are optimized for that, that are just regular instance types, time-memory instance types. And so from a COGS perspective, one of the reasons why customers really love our offering in the area of GenAI is because of the overall efficiency. So I just make sure that I add that because I want to make sure that if there are any questions around GPU usage and so on, there is some clarity around it.
Great, thank you.
The next question comes from Mike Cikos of Needham & Co. Please go ahead.
Hey, thanks for getting me on here, guys. A two-parter here. And I guess the first for Janesh, if I could ask, but I know earlier, we were talking about how the pricing changes are possible, but really, it seems like you guys are using the multiple tiers to drive adoption up tier for some of the new functionality. And I just wanted to get a better sense when we look at the revenue growth that Elastic demonstrating today, is there a reason to think that there's a more significant contribution coming from customers up tiering? Is that worth calling out right now? Or is that still relatively small in the grand scheme of things? And I'll just ask my second question now while I have you, but for Ash, I know we were highlighting ESQL as really helping streamline some of the migration issues from legacy incumbents. And I just wanted to make sure I'm thinking about that properly. And can you help me think through, if I'm using ESQL, is that a way for me to start bringing over maybe newly generated data or drive new use cases? Or is this actually porting over entire workloads and data estates that were previously locked up with those legacy incumbent SIEM offerings.
Maybe I'll start with the first question and keep it brief in the interest of time. In terms of the higher tier and the upsell there, we've generally seen a pretty steady increase in the adoption of our higher tiers over time, particularly the enterprise here. We've seen that for several quarters now, and we expect that this will continue to drive growth of Elastic Cloud looking ahead. So there's no particular inflection that I'm expecting our modeling. It's just consistent with what we've been seeing for quite some time already. We routinely add additional features to our higher subscription tiers. Our sales team demonstrates the value of that to customers. It’s just part of our regular upsell motion.
And let me quickly touch upon ESQL. So ESQL is not about bringing in data from incumbents. It's always been easy for us to do that. But incumbents have had query languages that allowed people to do interactive development using piped query semantics. So that just meant that for those users switching to the Elastic platform was a bit of a hurdle because they were used to doing things in a certain way. And now we are able to just translate their already built rules into ESQL and give them that ESQL query language which is something that they are very familiar with. That style of development is something that they're very familiar with, and that's what's enabling people to switch over that much faster.
This concludes our question-and-answer session. I would like to turn the conference back over to Ash Kulkarni for any closing remarks.
Thank you very much. Thanks for joining our call today. We are pleased with our Q3 performance and strong execution, as our focused sales motion in the enterprise and commercial segments is paying dividends. We are excited and confident about Q4 and the strength of the business heading into FY '25. Thank you again, and have a great day.
The conference has now concluded. Thank you for attending today's presentation, and you may now disconnect.