Elastic N.V. (ESTC) Earnings Call Transcript & Summary

Hello everyone. Good afternoon. Tyler Radke, I co-head Citi's software sector. Welcome to the afternoon track of a busy day 2 at the Citi conference. We got the Elastic team here with us for a great discussion, Ash Kulkarni to my left, the CEO; and the recently appointed CFO, Navam Welihinda. I'm sorry if I got it right?

You got it.

Okay. So gentlemen, thanks for making the appearance. For folks in the room that maybe are less familiar with Elastic, could you just give a quick overview on the company and how kind of the evolution of AI has impacted the business?

Sure. So Elastic fundamentally as a company was founded on an open source project, Elasticsearch,; which was written by our co-founders. And the best way to think about it is it's a search platform. It's a technology that's designed for bringing in any and all kinds of messy unstructured information and then making all of it searchable. So through that, you can then analyze that data in all kinds of ways. And over the years, we started in that sort of basic search area, but then grew into observability, specifically starting with log analytics because logs tend to be extremely messy. They're voluminous. They're hard to analyze. And then getting into observability over time, we got into security as well, into cybersecurity, specifically starting with SIEM and in the security, event management and monitoring space, that became the cornerstone, but then we also expanded into other areas in security, specifically endpoint security and so on. As AI has become more and more prevalent in terms of how people are not only using AI for doing -- building conversational apps and so on, but also for building agentic workflows that are being used to automate more and more business processes. What is very obvious is these large language models, which truly are, in my opinion, like the operating system of the future, the way you program these large language models is using English, but these large language models only know what they've been trained on. And all of that is what is publicly available out there. So these large language models have no context about your private information. So if you're using a large language model, if you're building these kinds of agents within your company, you need to somehow provide context to these large language models. And that whole process of what's called context engineering or data retrieval for context engineering, that is how our product is being increasingly used today. So our vector database and everything else surrounding it, that is what gets used in building these agents, these kinds of applications. And that has also benefited us in security and observability because increasingly for those processes, whether it's a SOC analyst or a site reliability engineer, we, through our AI stack, are able to make their jobs easier through automation of those systems as well.

Great. It's a good overview. And I guess double-clicking on the AI opportunity, just for -- again, for those unfamiliar, like how do you -- there's a lot of different software companies talking about AI. You have application companies, some other infrastructure companies. Like how do you monetize AI today? How should investors kind of think about how AI impacts the numbers side of the equation?

So the biggest impact for us is on our search business because when we think about our search business, that encapsulates any time you're building any kind of custom application on top of our core platform. And this data retrieval, this context engineering that I talked about, that is all effectively a search problem. You're trying to find just the right set of data within your overall environment that is relevant to answer that particular question. And so we are seeing an expansion of our search opportunity, and that's resulting in search effectively becoming the fastest-growing part of our business. When you think about search, observability and security, these 3 areas, search has become the fastest-growing part of our business, and it's because of the tailwinds of AI that we have seen. For observability and security, it's helping us compete better. So we lead in, in the case of security, we have launched capabilities about a year ago, we unveiled something called Attack Discovery that instead of just showing you alerts, in your SIEM now looks at those alerts and is able to using AI, identify the cybersecurity attacks, the attack patterns that are happening within your data. So it's effectively doing a lot of the job that a SOC analyst does. So it's helping automate a lot of that work. It's helping simplify and actually acts as an aid, an accelerant for that SOC analyst. All of that is what our field teams tend to lead with because that's a big differentiator. That's something that we are able to do that others aren't able to replicate. And that's making our security business more and more competitive. It's making it easier for us to win. We're seeing that same thing also play out on the observability side. And that's why we feel that it's going to help us increasingly in all 3 parts of our business as we go ahead.

Okay. Great. And I think one of the unique parts of the Elastic story is, I mean, it's -- the product can kind of be interchangeably used across all those 3 use cases, meaning you buy a subscription or consumption credits and you can use it across a wide range of use cases. I'm curious, I mean, how do you think about opportunities also to amplify the product's reach, meaning like, yes, this is -- it's a great developer platform, but there are other vendors, both in the search space, whether it's Glean on the AI side. Obviously, you have a number of observability and cybersecurity vendors that offer something a little more packaged out of the box. And clearly, there's been a rise of Vibe coding start-ups that make developing software easier. So is there an opportunity to make this stuff easier to use and ultimately drive more consumption? Or just how do you think about that high level?

If I just step back and I look at the AI opportunity, what I'm seeing is more and more applications being built that take advantage of all of the unstructured data that we have sitting within our organizations and automating business processes that depend on that unstructured data, right? That's really what's happening. Because we've always had the ability to write applications that worked on structured data in deterministic ways, right? So CRM systems or ERP systems, they all dependent on very structured data, and they had very well-defined processes that you could automate on top of that structured data. With LLMs, the big opportunity is to do that same kind of automation on unstructured data because that unstructured data, whether it's employee onboarding or it is customer support, there is a lot of stuff that you need to figure out on the fly. You need to reason on information that is not all structured. You're trying to respond to a customer that might have a particular problem. You need to look at so many different things. You need to look at what is the specific issue that they're facing. What do you know about the product and where you might have issues that might be known issues that exist in your product. You might want to look at other customer questions that might have come in to see, is there anything that we can learn from there. All of these are unstructured sets of information. You need to be able to automate based on all of that. That's where LLMs are really, really good. But now you need to provide this context to those LLMs. So the opportunity that I see for Elastic is that as more and more business processes get built to automate using AI, I see the opportunity for Elastic to be embedded as the vector database as the platform for doing that context engineering in as many of those AI applications as possible. That is the real opportunity for us, which is why we are so keen and so excited about what AI represents for Elastic.

Okay. Great. And as you think about that AI opportunity, I think clearly, it's going to be a huge market, early days, but there seems to be the opportunity both with AI natives, right? We're seeing a lot of these -- obviously, you got the model providers, the vibe coding platforms, cogeneration start-ups. And then you have the enterprise building their own platform AI. Like how are you positioned in each of those segments?

So we look at both motions as being equally relevant to us, right? At the end of the day, whether it's a software vendor that's building the next vibe coding platform or the next ERP solution or the next HRMS solution that's evolving out there, each of them is looking for some way to build an AI native application that's going to make that application more and more interesting, more and more exciting for users to use. We see that as the ISV play, right? So we want to be embedded in as many of those applications as possible. We've publicly talked about existing customers, whether it's a DocuSign, whether it's a Seismic. There are various examples that we have given of customers that have already done that with us, right? These are companies that have -- and we also have DevOps companies that have used us as part of their agents within their coding platforms. So we have lots of those examples. At the same time, we have enterprise customers that are using us for building applications, whether those are agents or just conversational chat style applications or implementing semantic search or what have you. In their environment where they're pulling data from different systems. So it's not -- it's a custom-built application that they are creating. It's for just their use case, it is not something that they're selling to customers. And that's perfectly fine, too. So we've seen banks. We've seen telcos. We've seen e-commerce companies. So lots of use cases. Fundamentally, Tyler, the way I look at it is AI represents a very different way to build applications in the future. Today, if you ask any major organization, if I just look at your company, your bank, Citi, you probably have dozens of agents that you are building or have built within your organization. And you probably, over time, will build hundreds, if not thousands, of these agents and applications. We are still in the early days. For us, the trick is getting into as many of these as possible right in the early stage. So we are part of that core fabric. We are part of that core infrastructure layer as they're building these applications. And as you build more and more applications, we just grow with you. I think that's the real opportunity here.

Yes. Yes. I think we're still a ways away from using agents at a large bank business.

I'm sure somebody within your company because we already work with various other banks. I'm...

Yes, for sure. For sure. Maybe there'll be an agent up here asking you questions next year, so better watch out. But Navam, I thought we'd bring you into this.

As long as they use Elastic for data processing.

Okay. Navam, I thought we'd bring you into the discussion here, just fresh off results, I think kind of maybe your first full quarter as CFO. Pretty strong beat across the board across both cloud and self-managed. I know there was a lot of questions around the price impacts in the quarter as well. So maybe just frame for us kind of the key puts and takes on the quarter from your perspective now that we're almost a week past when you released.

Yes. It's been a great first quarter. Let me tell you that much. So from what you said, that's absolutely true. Q1 was a strong quarter across the board, and we beat the top line by $18 million, and we had a strong bottom line performance as well. The way I think about the health of the business is around 2 aspects, which is how are commitments going and how is consumption going, right? And both of those in Q1 were very strong. Commitments, if you think about what happened there, year-over-year, we saw growth. The growth was across all geos. It was balanced, no outliers. That's great news. On consumption, it was the same thing, year-over-year growth, no outliers and strong commitments across the board in Q4. So overall, from a Q4 perspective, I thought it was a great quarter, and we were very happy with what we delivered. On your -- and the other thing is when you think about the underlying consumption level that we saw from Q4 to Q1, that increase was very strong. So we're happy with that consumption increase that we saw in Q4 to Q1 as well. So moving on to your price comment, I think it's worthwhile parsing back a little bit and understand most software companies do price changes, and we're no different from any of those other software companies. And this isn't the first one we've done. We did a price change last year on the self-managed side. We did a cloud and self-managed price change a couple of years before that, and we just did one in May. So this is more of a history of us periodically looking at prices and making changes just like most other software companies do. So it's more of a matter of course of business for us, right? And the reason we are okay with changing prices and the reason we feel good about changing prices is because of the way we introduce functionality into our product. So our customers, they don't get separate SKUs for new things that we deliver. We sell a platform and all the new product introductions go into that platform. So over time, what we're delivering is more value for money in that platform, which we capture periodically through price increases, right? So that's an important point to remember that this is no different from any other software company practice, and it's no different from the history of what we've been doing in the past. The second point I want to make is that when you think about price changes, what matters is how does consumption change with the changes in price, right? And when you think about consumption, there are multiple puts and takes you need to understand. In any given quarter, people are optimizing their usage. So there's data coming in, which increases their consumption. They're also optimizing their usage, which puts downward pressure on consumption. So that's happening in any given quarter. The second thing is we introduced feature functionality that is meant to drive more efficiency in our customers. Two specific examples of some things we've done in the past are searchable snapshots and Logsdb. If you adopted those 2 features, you're naturally going to be more efficient and consume less. So that's downward pressure on consumption. And then there's pricing, which adds upward pressure on consumption and also people react to price, there's elasticity and people change consumption. So what matters to us in a consumption business is very different from a seat-based model, which is a simple P x Q math. What matters is how does consumption change in relation to all these changes that I talked about earlier? And is the net consumption growing? And that's the biggest thing we care about. And what we saw in Q1, right, so that's the main point I want to make about pricing. You can't really in consumption models, isolate one individual variable across the many variables that I talked about and then do simple math and say, well, absent this, it was X, Y Z, right? So that's a big misconception that I think we need to clear about as we consume how to think about pricing in consumption models.

Right, right. And some of those efficiency capabilities that you highlighted, the searchable snapshot and some of the login functionality, was there any like significant additions to the portfolio in Q1? Or was this kind of more of a comment on like...

No, no, no. Logsdb was introduced in Q4. So, yes, yes. So this was -- but again, like...

In conjunction with the...

Well, not in conjunction. We don't think about it necessarily in conjunction, Tyler. The best way to think about it is we introduced those kinds of features. Searchable snapshots was about 5, 6 years ago. Then another thing that we did some time ago was supporting some of the newer chipsets, ARM-based chipsets, Graviton chipsets from AWS. That made the system more efficient. So it brought down consumption. We introduced better compression, and we have done that in the past. That brings down consumption. The whole goal is to keep making the platform better and better. So competitively, this looks like the absolute best choice for customers, and they keep bringing more and more workloads to us. And what we see is when we do that, for any given workload, we might see a near-term pressure, but very quickly in the mid- to long term, people bring more workloads and we see growth, right? Searchable snapshots when we introduced it 6 years ago, it immediately put pressure on existing workloads. But as people started retaining data longer, they started bringing other workloads. We started to see -- that was one of the features that drove more and more people to use us as compared to Splunk, even in those days. That pays off very nicely in the long run, right? And that's why we do these things. So to Navam's point, trying to disaggregate any one factor, is just meaningless.

Okay. Very clear. So you feel very good about the consumption growth in Q1. And I assume based on the raise of the year, you're expecting that -- those trends to continue.

Yes. We expect -- because of what we are seeing, both in commitments and in consumption, we feel really good about the strength of the business in the year.

Great. Great. No, super helpful. Speaking of sort of the changes between Q4 and Q1, I think you sounded a bit more upbeat on the macro environment, maybe upbeat, a little strong of a word, but obviously, the federal weakness that you saw in Q4 was pretty widespread across the software space. Like just remind us, what did you see get better in Q1 versus Q4, both in federal and across other industries?

Well, let me tell you about federal and then Navam can also add to it. In federal, so I spend a lot of time with customers. I'm generally out on the road a fair bit. Next week, I'm going to be at the Billington Cybersecurity Conference in D.C. On the federal side, the way I describe it is in Q1, it felt like the new administration has sort of settled in, right? So clearly, they have a greater focus on efficiency than we have seen from prior administrations. But like there is -- the fact that they are settled basically means that we are not seeing a lot of personnel changes on an ongoing basis, like things are just more settled, people are making decisions. And even in that new normal, even though there might be a greater consideration and focus on efficiency, that's a world that's more stable, and we know how to operate in that world, right? It's a world where we know the strength of our product is going to continue to hold very, very well because, Tyler, as you know, when you look at value for price, like we've always had an outstanding value proposition. And so from a competitive standpoint, we feel really good. And now that people are making decisions, it is an environment that we feel really good about operating in. So from that perspective, like we said, like the public sector environment, especially in the U.S. felt stable, right? It felt good compared to the uncertainty that we were experiencing 90-ish days ago. But I don't know if I missed anything.

I would echo that. Going into the year, we had detailed out the headwinds we were -- or the pressure we were seeing and then assumptions on what could happen beyond the U.S. public sector. civilian ag, just to Ash's point, throughout the quarter, we saw that the U.S. pub sector was much more stable. We know how to operate and the teams are primed to operate in that environment now. Our products resonate in the administrations focused doing more with less, right? So we have a good footing in this new normal. That's number one. And then clearly, we did not see the scenario play out where things spread across other geos and unrelated sectors to create more headwinds. It was a much more stable environment than what we had feared or guided to in Q1 and -- or in Q4. And the result is we had a great quarter. We feel good about the year, and we reflected that by raising the guide more than what we had beat. So at the bottom end of the guide, we beat by 18. The bottom end of the guide got -- we increased by $24 million and the midpoint by $22 million. That's to signify that we feel much better about how the year is going. And every quarter, we're going to go execute and go revisit that number again.

Right, right. But you would describe the guidance is still pretty prudent in terms of macro assumptions and everything.

That's right. I mean we -- I think I'd always aim to give a prudent guide, and this one is no different. And like I said, we feel good about the year, which is why we raised beyond what we delivered. And every quarter, it's going to be -- we're going to execute and go -- give you a new view of the year.

Got it. Got it. Ash, I'd love to just ask you about the competitive landscape. Obviously, there's a lot of different areas that the Elastic product touches everything from cybersecurity to observability and then search and AI with vector search. But yes, there's certainly been a lot of kind of consolidation pressures, if you will, in both the security and observability market, whether it's budget constraints or OpenTelemetry. How are you seeing that impact the Elastic business either positively or negatively?

So just in terms of where we see our competitive dynamics play in our favor, in security, we've always talked about security as being a data problem. And so anytime you have a situation where customers appreciate the need to bring in all the data and retain it for long periods of time because they understand that the cyber landscape is pretty complex, that plays to our strength. Anytime they need to use, they feel that it's in their benefit to use AI to drive more automation, that plays to our strengths. And lastly, when you look at security, if the customer has a desire to run that security workload on-prem in their own data centers, we are really one of only a couple of choices at scale that works because many security products are only available in the cloud. So one of the reasons why you will continue to see us do very well in our self-managed business is for that reason. There is tremendous advantages that we have, not only as this data-oriented security player, but also because of our AI functionality and our ability to have a really strong offering in a self-managed mode. When it comes to observability, that -- it is a large growing market, but as you know, it's also a very busy market. Our core strength is when the data is messy. So that's why we lead with logs because when it comes to logs, especially application logs, the messier the information, the harder it is to query, the more you need a product like ours, a platform like ours that's inherently designed for unstructured messy data. If the data is structured, if you're dealing with metrics and so on, there are other players that naturally have been in the market for longer, and we see that competition obviously is there. But when it comes to logs, we have a true differentiation. And the work that we're doing in terms of both driving more efficiency into the platform, but also using AI more effectively there is a continued strength for us. And then on the search side, you saw this movement where there were a lot of attempts at trying to define vector databases as a separate market. I think it is becoming more and more clear to people that vector databases are feature as opposed to a category in and of themselves. For us, our focus is always -- our center of gravity is always going to be on unstructured data. You can absolutely put structured data into Elastic, but what we do uniquely well is when it comes to unstructured data, the ability to query that to be able to search and do all kinds of analytics against it, we are very, very good at that. What I expect is going to happen, and we're starting to see that is you're going to start to see centers of gravity continue to strengthen on platforms for different types of data. So if you are structured, you're going to end up with certain kinds of platforms that are really good at analytics. If you are unstructured, you're going to tend to see companies like us. And I would expect our competition to be primarily the hyperscalers in that area. And we know how to compete very well with them. We've been doing it our entire existence.

Right. Right, right. And I guess just as you think about sort of how to drive the durable growth of Elastic and this goal to be the embedded kind of platform within these emerging AI applications, obviously, top of funnel, developer affinity, those are important things for the next generation. You did make the return to open source AGPL, I think, over a year ago, if I'm not mistaken. Give us an update on how that's gone? Like what are some of the metrics you also look at to measure this top-of-funnel momentum?

Yes. So the reason why we adopted the AGPL license -- before that, we had the SSPL license that we supported and the Elastic V2 license. Elastic V2 is very, very permissive, but neither SSPL nor Elastic License were OSI compliant. So AGPL is an OSI compliant, so you get the official stamp of being open source. The reason why that was important for us is in the community of open source developers as people are looking at open source alternatives, they basically look at what's OSI compliant. And where this matters most is in the area of vector databases. That's the evolving space, right? That's the area where you are still -- we are still very, very early in the overall AI market. And given how early we are, that's the land grab right now. So it matters to us on how we are -- where we are seen, how people find our technology, download it, use it. And to me, it's less relevant whether they start by being a paid customer to begin with because these developers are going to just play around with open source technology. And over time, as what they're building becomes more meaningful, they're going to look to a commercial vendor and they're going to stay on that platform. So what we look at in terms of metrics, we look at how we show up in various open source forums. We look at the number of downloads of our product. Those are things that we track. We obviously look at the number of trials that are happening on our products. So there are various metrics. The AGPL license is something that we cared about because we wanted to have that top-of-funnel activity. And everything that we are seeing gives us a sense that it's really been the right decision.

Okay. Okay. Great. And then, Navam, on your end, we have an Analyst Day coming up in a little over a month. Obviously, you're still putting together the plans in terms of what you're going to share. But how are you just thinking about targets, both on growth and profitability? I know it's something the company has guided to in the past. You may have a different philosophy, but how should we be thinking about how you think about long term about the business, obviously, without giving away too much ahead of Investor Day?

Yes, I don't want to give the punchline, then you won't show up. So I encourage you all to attend Analyst Day, it's going to happen in about a month. It's happening in conjunction with one of our user conferences, ElasticON. We have a few of them. This one is in New York. So Analyst Day will be in parallel with that. If you do attend, you'll get to see the keynotes as well as some of the demo booths and our users in full swing. So that's a worthwhile experience. On the Analyst Day side, which is parallel to ElasticON, you get to meet our team, talk about things in a broader context. So the product side, you see Ken and our GMs, one of whom is with us here, Steve, talk more about the broader vision on product along with Ash and the GTM side, Mark will talk about all the changes he's done to GTM and think about that as well and get some information on that as well. In finance, yes, one of the sections we're going to have is the medium-term model for the business and how we balance growth and profitability and what you should be thinking about in terms of our growth algorithm. So I don't want to give it away because then you won't show up. But that would be what I'd expect to see in Analyst Day.

Okay. Okay. Great. And I think I guess just go-to-market, I know we only have a couple of minutes left, but a little over a year ago, you did make some changes there that proved to be more disruptive in the Q1 a year ago, but it seemed like you kind of bounced back from there. So just remind us like what were those changes? And have you started to kind of see the returns on some of those investments or changes that you made?

Yes. So for the last 4 quarters, if you look at our sales-led subscription revenue, right? I mean you just have to go and see Navam's script. We have shown consistent and very strong execution, and that was possible because of those changes that we made. So the changes that we made 5 quarters ago, they did two things. One is they created greater focus on enterprise and mid-market accounts where fewer accounts per rep meant that our reps were able to go deeper and broader into those accounts. As you can imagine, that leads to greater focus that results in higher quality deals, larger deals. We've seen the benefit of that in the last 4 quarters. And in the greenfield territories that we created, we created more of a dedicated hunter motion. And that has also been paying off. So if you look at the number of net new million dollar deals that we've been adding the $100,000 deals, like we are seeing the benefit in all of those metrics. And the best metric to me is sales-led subscription revenue, right? It doesn't matter if it's self-managed or cloud, but that is what we focus on. That is what we drive. And you just need to look at the data to see that it's been really strong execution throughout these last 4 quarters. So we're excited.

Great. Well, let's wrap it up there. I think we're out of time. Thank you both for the discussion, and look forward to seeing you at the Investor Day in November.

Thank you, Tyler. Thank you very much, folks.

Thank you.

Thank you.