This is going to be a Kash asks Ash question, when Janesh said, "No, that's too, too tacky," so I'm gonna be part of the mix, and so we just call it me interviewing Elastic discussion. Sounds good. Thank you once again. Yeah, go ahead. Thanks for having us here. Absolutely. I mean, you guys look fantastic. I don't think I've had the chance of having both of you on stage with me. Last time it was you, and I think many years ago in our prior lives, Janesh, it was just you. We've known each other a long time. A very long time. Yeah, a very long time. So welcome to the conference. Thank you. Welcome, everybody, to Goldman Sachs Communacopia and Technology 2024. I won't bore you with the commercials. Of course, the conference is off to a great start. Thanks for investing your time with us. So Ash, let's talk about your vision for the company. I think vision's a somewhat of a fluid thing. Last year, if we asked you something, you had a certain vision. Hopefully, it's not changed a whole lot, but as you have refined your thoughts on where you want this company to go, as you've learned things over the last few years you've been CEO, where do you see the company five years from now? Yeah. So in terms of the way, you know, I would like to see us continuing to grow, you know, we think of ourselves as a search AI company. What that means is, anytime you are dealing with unstructured information, which there's a lot of, we wanna be the runtime platform for our customers to be able to get all the right insights from that data. You know, GenAI is a huge opportunity for us, and I wanna make sure that as we go forward, that we are the runtime platform of choice as customers are building, you know, GenAI, RAG applications. I think that's the biggest opportunity, and that's gonna effectively help all aspects of our business. So it's a platform play that we are absolutely focused on, and that's what I see as the right path for us for many years. Got it. I guess that would be a logical evolution of the origins of the company, right? Yep. How do you connect the past to the future through that search capability? Can you just tell us how you build on what has already been done really well? We've always been. At the heart of it, we've always been a search company. Like even when it comes to observability or security, the solutions that we play in, you know, when we win, we win because of our ability to help customers, ingest, index, make searchable, and get really good outcomes out of really messy, unstructured data at extreme scale, right? So when it comes to security- Messy, unstructured, at massive scale. At massive scale. Yeah. And you think about security, that's the reason why in security we are so focused on SIEM. Mm. Security has lots of subcomponents to it, but when you're dealing with SIEM, the challenge ends up becoming a data problem. It's all about pulling in your application logs, your access management logs, your web logs, your network logs. Try to correlate across all of them in real time, and try and find patterns that allow you to identify potential attacks that might be happening within your environment. When you look at observability, that's the reason why we focus. First, we land with log analytics. Because when you deal with log analytics, that's where the data is the messiest. That's where you need to eventually end up going if you want to do any kind of root cause analysis of what is potentially causing a slowdown or any kind of, you know, negative impact on your application. So that search piece has always been the reason why we win. And in the last several years, we've really doubled down on that, right? That helps our search business, but it also helps our security and observability business. And then, as we've invested for the last, you know, four or five years, in all the things that have allowed us to have our own vector database the ability to not just search across data using traditional textual techniques, but using vector techniques, like, it's allowed us to do a lot more, leveraging large language models to sort of get better security information than we could before. The same with observability. And in just regular search, like, semantic search is the number one thing that we are seeing in terms of interest from our search customers being able to search using context. So it's, you know, you talk about connecting the past to the future. I think that's really why we are excited about what's ahead of us. That's great. And as I was listening to you talk, I couldn't help but wonder, do you ever... And I asked this question of Dev Ittycheria, in fact, on this very stage, I think earlier today, or maybe it was lower level. I forget which level I'm on. Could you become an applications company? Because you have so much capability, and it starts with the infrastructure, which has been built out by NVIDIA, Microsoft, hyperscalers, and you have platforms, which is where you are. Yep. And to make the platform come to life, you need to build applications. So do you envision doing things like that? A lot of people, especially in security, think of what we have built in terms of our SIEM as a security application, right? Sure. So, you know, our SIEM is the primary application, the primary workflow that security analysts use at several large banks, at major telcos, in major government organizations. Our. You know, I don't know if we'll ever get into applications for all kinds of business processes- Yeah, like generic applications is what I- But yes. But specifically around GenAI around security and observability, where we are already playing, like, you will see us doing more and more in terms of offering complete solutions, right, out of the box. We've done a nd in my opinion, we've done a good job of that in security. We've seen the benefit from that. We've done a lot in observability. We've seen the benefits of that. And in GenAI now, what we're really focused on is How do we take all the expertise that we have and, you know, make it possible for customers to implement customer support applications very easily, to implement e-discovery applications very easily? So we're gonna do more and more of that connecting of the dots to make it like it's just one single place that you need to go to, to do everything and get that outcome that you're looking for. But, you know, that's not gonna change the fact that we're an infrastructure company. Like, where we sell to- Yeah. It's still a technical buyer. Yeah. Right? It's the CIO, it's the CISO. I don't think we're ever gonna sell to, you know, the Janesh of the world or CMOs. Yeah. Like, I think we are still gonna be on the IT side of the house. Got it. Got it. The search functionality is the pivot around which you've been rebuilding and repositioning the company. And Dev, earlier today, talked about how the document model is putting the company in a very good position to be a development platform of choice because of the unstructured nature of the data, the complexity of the data, the velocity of the data. So... Maybe there's a way to think about how your search capability gives you that, that kind of must-have factor, the differentiation versus others that could come in and say, "Because GenAI could open up the playing field." You could have a lot of people come in and say, "I can do this search too. Yeah. But what, what is it, just like in the case of Mongo, that's their shtick? Yeah. And, you know, I think the fact that when you're dealing with modern applications modern processes, you tend to see a gravitation towards this document model, which I think is something that we've seen now for the last almost two decades. However, if you just go back and say, what is the data that you're bringing in into these, you know, different systems, different data systems? Probably the biggest difference between, say, a MongoDB or a Snowflake and Elastic is the data that, you know, Snowflake or a MongoDB brings into them tends to be such that you can put it into a proper schema. Even though it might be a document-centric model, it still conforms to a nice schema. The type of data that we tend to deal with is often data that does not conform very well to a nice schema. So when you're dealing with Word documents, when you're dealing with PDFs, when you're dealing with, you know, application logs especially they don't have clean schemas. You need to be able to infer things from them. You need to be able to deal with, you know, schema changes that need to be applied literally, like, you know, row by row in a log file often. Like, you need to be resistant to schema changes and changes that happen in the stream that's coming to you. It's that messiness that we do incredibly well at, because we are able to ingest that in and make everything searchable. That's our core strength. So where is this applicable? Well, you take any organization, and you think about how much data sits in their formal databases, and how much data is just sitting on file systems, in SharePoint, and in other places. And you realize that there is way more data, unstructured data, that's sitting in all these other systems, but that's also been the hardest data to monetize. That's also been the data that's, you know, typically used in more manual processes. You're initiating a loan. It's somebody who is reading through all kinds of documents and making decisions. You are responding to a customer support call. It's somebody looking at different sites and reading through things and trying to help you understand how to configure your router or whatever. You are, you know, in a government agency, creating a rule for, you know, what charger goes where or where you're allowed to build a gas station or not. Like, all of these things involve a lot of documents What GenAI is, is offering is the opportunity to start to automate more and more of these processes. But to automate these processes, you need to be able to pull in all that unstructured data, parse it quickly make it searchable, and then using retrieval-augmented generation, allow a large language model to now create responses and automate these workflow elements. That's really the big opportunity for Elastic, and that's where we become the runtime. So we'll work with every large language model out there but we are the runtime on which that data gets brought, stored, vectorized, and searched. And, you know, if you believe that the future is gonna involve more automation of more of these workflows, we have the opportunity to be the runtime. You know, whether you call us a database or not, like, is less relevant. It's the runtime that powers this next era of AI-based applications, and that's what's exciting for me. He's good. He's really good. That was good. That good. That was really good. I like the way you, you explained it. Yeah, we're seeing this from our customers, so that's what is exciting. Janesh, I wanted to ask you, you're COO as well, so you have more purview into deployment of technologies within the company. I'm curious to get your thoughts. Have you deployed GenAI within the company? And also. That's internally. Externally, what are the factors you're looking at before you invest against the opportunity that Ash has laid out? Yeah, on both of those, Kash, we have. We've, we're big believers in sipping our own champagne, chewing our own dog food, pick your favorite expression there. Yes. But we have, and we've used it as in a couple of different ways. One is around customer support, which is a very popular use case that we see with our own customers, and we're doing the same. We've also enabled it for our sales organization, as you think about sharing knowledge and information and making a salesperson eventually more productive as we leverage that capability in terms of things that they can bring to a customer, and how to represent the technology, and assets that they can leverage from within the company, and so forth. Still in the early stages of deploying those and scaling them, but the early results have been very encouraging for us, and we're looking at a number of other use cases as well in terms of some of the other parts of the business. But in terms of how we think about the financial metrics or the investment metrics for GenAI, you know, to start with, I think it's just... I think about the size of the opportunity ahead of us, and it's big. And I can't tell you with precision exactly how many billions of dollars of TAM there is, but we know it's big. It's a big shift that you see these kinds of shifts, you know, once every generation. And, more often than not, with these kinds of large market shifts, what you'll see is that they take longer to play out than you initially expected, but ultimately they turn out being much bigger than what you initially expected they would be. And so with that, with that view in mind, recognizing that it can take a long time to play out, and Ash gave some really powerful examples of all the different ways in which it, it can play out, and all of those directly relate to our business. The way we think about that then is, what are the near-term milestones we should be looking at, and what are the near-term indicators of customer success and traction? If you think about how our customers are deploying the technology, it'll start with first selecting Elastic as their technology partner of choice on which to build their application. They need to then build the application, test it, eventually start to build integrations to it, deploy it, and that's when a lot of the data starts to come in and starts to scale in terms of deployment and volumes. Our revenue model with customers is one where we're successful when they're successful, and as their consumption eventually ramps, that's when we start to see revenue at scale. At the initial stages, the more important metrics for us are around how many customers are actually selecting us to build their applications, and what do those actually look like? And we've talked in our most recent earnings call, and for the past several calls, we've talked about adding a few hundred customers every quarter that are actually using us for our vector database functionality. And that's, by the way, all in the cloud. That doesn't even count self-managed. And, now there's more than 1,300 of those. More than 200 of those customers are in our cohort, which is more than 100K in ACV. That also doesn't count all the ways in which people are using us for GenAI, and observability, and security with the AI assistance and some of the additional functionality from that we've launched there. So I look for those signs of success in the initial stages, and those have been really encouraging. In the most recent quarter, we saw a slight mix shift towards search in terms of the composition of our overall business that the field team closed. So that was really an encouraging indicator as well, and I think all of these eventually will translate into revenue, but those are some of the things I look for in the near term. Yeah. Just, if I might just add Please to what Janesh said, like, just to make it more concrete in terms of the internal use cases, so semantic search, we've implemented semantic search on our website. Again, like, you know, we see that if you are coming to our website and looking for information, we want that experience. To be really good. The one, you know, apart from support, the one use case that I personally thought was great is, our... One of the things that happens when, as a company, we ship new capabilities in the cloud every three weeks and in self-managed, like every two months, it's, you know, keeping customers abreast of new developments is super important, right? That, that's what lets them know there's a new capability that they might want to leverage, that then motivates potentially more expansion, and so on. And one of the challenges has always been: How do we become better at account-based marketing, if you will? Like, how do we get the right customer, the right person, to know more about the things that are relevant to them? So our team internally built something that they call Sales GPT. Like, I forget the corny name, but basically what it does is, you know, if I have a set of customers, like we've slurped the data from our own Salesforce instance into an Elasticsearch instance internally, and we've built a RAG application on top of that. So as a rep, if I want to send an email to my customer, letting them know, "Oh, eight-fifteen, eight dot fifteen just came out," like, that Sales GPT application looks at what the customer has used in the past, based on information that's in Salesforce, looks at the release notes, and takes those two things and effectively crafts an email for the rep to send, to say, "Hey, here's something that just came out in eight-fifteen. There are these new capabilities that are really relevant to what you've been using. You might wanna check it out," and that just improves the whole process of how we communicate with our customers. So it's like those kinds of automation of tasks that, you know, are meaningful at the end of the day. So, you know, we love to talk about. I love to talk about these things because, it's like, this is the knowledge sharing that is helpful for our customers. So this is not a hype cycle, it's real? We are seeing value. We are seeing our customers actually get value out of it. You know, the level to which different workflows are getting automated is gonna vary, but, like, I'm seeing automation happen everywhere. That's great. So, Janesh, growth versus profitability. Where are you in that profile? It looked like we had indexed towards more profits, but then GenAI is opening up opportunity sets. So how are you thinking about the balance between the two? Yeah, Ash, so Ash, Kash, it's very similar to the way we've always thought about it, which is that it is a balance between revenue growth and profitability. Software business models, as you know from all of your work, have enormous leverage inherent in them, and as the business continues to scale and grow, there's a lot of that ultimately comes into the kind of falls to the bottom line and gets reflected in a higher operating margin. Is it just a GenAI scale or better retention or better upsell? What is it that, I think it's all of those, and sales and marketing efficiency is also part of that equation. Which factors. And so I think it's all of those factors at play, and we've been the beneficiary of that, every year as well. And this year, our fiscal twenty-five, we said we were gonna lean in and invest a little bit more heavily towards GenAI, because we see such a significant investment opportunity. But we also, at the end of our fiscal Q1, when we released our earnings, we also dialed back our guidance assumptions, for the year a little bit on the top line. As a result of that, we also stepped back and thought about: how do we wanna balance growth and profitability this year? Fundamentally, nothing has changed in terms of our longer term thesis and our core prospects and the way we think we can capture market in the long term, and so we didn't wanna do anything that cut back on our ability to prosecute growth in the long term. So we're staying the course on GenAI investments, particularly in the R&D side. We're gonna continue to build sales capacity, in places where we think we see the opportunity to drive that growth. And we did a little bit of belt-tightening in other areas where that are not immediately focused on driving growth in the near term. So for us, it is always a balance, and, as we continue to deliver growth, we believe that it's a privilege to keep investing in the business, and that level of investment is dictated by both the opportunity as we see it, as well as our ability to execute against that opportunity. And, how far are we in the process of analysis of the things that surprised you in the last few weeks of the quarter, and what are your takeaways? I guess with every week passing, we could get even more clarity, get more insights. Yeah, but what are your latest insights on the action plan? Well, I can talk about it, and then I'll obviously ask Janesh to maybe add on to it. But you know, fundamentally, there were three changes that we had done at the beginning of the quarter. The first was, so our sales segmentation is such that we have a strategic segment, which is the very top of the pyramid. Those are the, you know, typically, it's Fortune 500 or very large tech companies. Mm, Mm, that fall in there. Those are accounts where we already have a presence, and we see the biggest opportunity to grow. That's our strategic segment. Below that, we have our enterprise segment, and then our commercial segment. So the first change that we did was we moved more accounts and sellers into the strategic segment. In that segment, you know, like I said, we've had a lot of success in displacing incumbents and really growing our footprint through consolidation of other workloads onto Elastic, and, you know, we wanted to put more energy there. The second change that we did was in the enterprise segment, historically, we have had. You know, reps have had a lot of accounts, way more than industry norm. You know, when you do that, what tends to happen is most reps, even our data showed that most reps would work maybe five to six accounts, because realistically, that's many accounts they can have strong relationships in, and go deep in. The remaining accounts just don't really get worked. You have the situation then when nobody else is working on those accounts because they're assigned to a rep, but you know, there's no progress being made. That's not a great answer for how you scale in the long run. The second change that we did was we reduced the number of accounts that each enterprise sales rep was assigned, brought it closer to industry norm. And then we took all the accounts that had no ARRs, so they were greenfield accounts, and we turned them into distinct greenfield territories, and we assigned reps that have more of a hunting, you know, profile to those territories. So those were the three changes. Historically, every year, there have been some changes that happen in any territory, like there are accounts that change hands. This time it was more than that. It was significantly more than that. All these changes were done only in the Americas, outside of public sector. And you know, the thing that we got wrong was how we executed the account transitions. The changes themselves, even as I look at them, I reflect on them, on everything that we've done, you know, I feel very strongly that these were the right changes. These are not changes that are, you know, groundbreaking by any means. Other companies in enterprise software have done them in the past, and they're all kind of necessary at some point to get beyond a certain scale. The thing that we could have done differently was maybe stagger them a little bit, but the account transitions are the things that we got wrong. The level of inspection that we had on those handovers between the outgoing rep and the incoming rep was an area where we didn't do as well, and we caught it very late in month three, because our quarters tend to be back-end loaded, so you know, by the time this became obvious that this was a meaningful situation, it was in the second half of month three. As soon as we realized it, we started taking actions. Ever since, we've been really doing very diligent inspection of the opportunity progression and new opportunity creation in all the accounts that have changed hands. A lot of those deals that slipped have closed since, and you know, it's been across the board, so it's not just in cloud or self-managed or one use case or another. It was across the board, and we've seen a lot of those deals closed, including you know, displacements, where we've displaced incumbents, but we lost time because of that. There was a period where we didn't have the same kind of productivity as we've been used to. So After you reported the quarter, these, deals No, we saw the issue Right after the close like, you know, in the, In late July last part of July. Oh, okay. So in the last No, what I meant is, the slipped deals that ended up closing- Yeah, they Was it after you reported or were they? No, no, we even talked about, like, this Okay. Many of them closed even before Yeah We reported. Okay. And more have closed since. Yeah. So it's been, it's been ongoing. More have closed since that? Yes. Okay. I mean, it's like, the teams are working hard. Yeah. Let's not the deals that closed, you know. The things that I look for are: Did the dynamics change in the deals? Did we have to do anything different in terms of additional discounting? Did we have to do anything. Did the opportunities just go away, or did the customer reduce their engagement with us? Like, was there anything that tells us that the dynamics in the market have changed? None of that is the case. Like, the deals literally slowed down because of the mistake that we made in terms of, like, having that smooth progression. And, you know, that's gonna have an impact. That's what's causing. That's what caused us to bring down the guide. But we fully expect that every quarter from here on out, you should see us improving back t o get back to the level of productivity that we've been used to in the past. Very encouraging. Janesh, you've been in the finance domain for a few years. I say that because of your youthful appearance, although I know it's decades. Right? You've been through rate cuts, rate increases, whatnot. What do you make of... and I think you're a finance major in college, if I'm not mistaken. I think we've talked about that, and MBA as well. Not that you need to be, but you are. What is your assessment with rate cuts around the corner? What are the things that would be beneficial for your customers? Could there be some kind of unlock? Yeah, you know, on rate cuts, Kash, you're more the expert than I am, and your crystal ball is probably even better than mine. No, no, yeah, Jan Hatzius is saying that we're gonna get a rate cut. So yeah, he's. It's not even like a thing. It's like it's 25%. Yeah. So, I mean, if I think about what that means in terms of customers, right? And, how customers think about it, you know, for us, what matters is ultimately what happens with IT budgets. Yeah. Maybe if I think about what happened in 2022, when we started to see macroeconomic impacts more broadly in the business, and in the industry as a whole, that's when customers started to dial back on their spending. And as things get better in the macro and as rate cuts take effect, if that all ultimately happens to translate to increased IT spending, we will be the beneficiaries of that as well. But I think one of the important pieces to keep in mind is that as part of our core value proposition to customers, as part of everything around platform consolidation, a big part of the reason why customers ultimately look to Elastic is because they see that we have the ability to save them dollars. That's, you know, it's incredibly powerful in any kind of economy, whether it's a good economy or a bad economy. We've seen them move away from legacy vendors onto Elastic because of the ability to get not just great business outcomes, but to do that at a cost that is quite compelling. We feel really good about that motion, and we hope that that continues for some time to come. Great. Thank you. Pulse check. Anybody with questions, just raise your hand. We have two LLMs that are waiting to be prompted with some good questions. Okay, I'll continue to prompt the discussion here. I mean, prompt- You can, you can be the RAG right now. Prompt engineer. No, I don't know if I'm the RAG, but the prompt engineer, and the RAG is sitting here somewhere. I mean, with LLMs, et cetera. You be the RAG, and he'll be the LLM, and I'll be the prompt engineer. How, how does that sound? Sounds great to me. Okay. All right, cool. With respect to the platformization ahead, as you start to entertain ideas that this could be a multiple of current size, right? The things that help a company go from zero to hundred are not the things that help them from hundred to five hundred, five hundred to billion. You get the picture, right? As you're looking ahead, what are the kinds of things you're looking to bring into the company? What kind of talent are you looking to bring into the company that can help you for the next chapter of growth? Very often companies react, and, "Oh, my God," so looking backwards is not the right thing, so... Yeah, no, there are some things that we've been incredibly focused on from a product perspective that I feel we will continue to stay loyal to, right? The first is this notion of having an integrated platform. You know, our ability to get into security was really insignificant for us five years ago. Today, it's over 25% of our business, and that's happened- Was it Endgame, right, was it? There was the Endgame acquisition but it was not a massive acquisition, by any means. I'm just trying to impress you with my throwing names like Endgame. No, no, it's great. That's, that's all there is to it. I really appreciate that you know, that you have all those details. It wasn't really a massive acquisition, but what we got right in that acquisition was we took the effort to integrate the technology deeply into the Elastic platform. Yeah. Because of which, now when we go and talk to a customer, you know, if the customer is using us for observability and is bringing in data, like they already have Elastic data collectors deployed everywhere, those same agents that are used to pull in data for observability also can be used for security purposes. Because there's endpoint security built into them, and those same collectors can be used to bring in data for security analytics or SIEM, so that kind of product integration has been very valuable for us and we've gotten that right. Our ability to bring in really talented people through these kinds of acquisitions, and it's not like we do acqui-hires, but when we've done them, like, our focus is less on acquiring revenue and more on acquiring technology and talent. You know, the people who have since run meaningful parts of our business have all come through these acquisitions, which has been great. We've been also able to then grow them internally. You know, the next focus for us is serverless, which we've talked about. Like, as I look ahead, cloud is gonna be very important for us. We are doing a lot to make sure that our focus on cloud stays front and center. Yeah, serverless is all about that, right? It's all about creating a very easy experience and a growth path to make that happen. And the second thing that we are doing is refining and continuing to improve our motion to sell higher. The changes that we made on the go-to-market side in Q1, when you peel the onion back and you think about why were those changes even necessary? They are necessary to improve, over a period of time, the productivity of the organization to make us able to do larger deals without, you know, to become more and more efficient on the go-to-market side. So effectively, it's about improving efficiency on the go-to-market side, improving efficiency on the product side, and when you do those things correctly, in a market that has a large TAM associated with it, like what we have with GenAI, like, that just sets you on a really nice path. That's what we are focused on. Like, again, like, you know, like I said, we could have done them a little, you know, differently, and implemented them with, you know, more care? Absolutely. But fundamentally, these are the right changes, and this is what sets us up very nicely for the coming years. Got it. We've got a minute and thirty seconds. Anybody with a quick prompt? Yes, go ahead. Ash. Hey. Who do you see as your competition for Splunk? Is it CrowdStrike, Cisco Splunk? Cisco Splunk is typically the one that we see the most. For our scale, the two competitors that we run into most are Splunk and Microsoft. Got it. Any other final question? That was actually gonna be my question. I was gonna ask you about the competition. Any notable changes after the Cisco acquisition of Splunk? No, they've. I mean, anecdotally, we've heard that they have increased their pricing and so on, but. Okay you know, apart from that, I know that they've announced a lot of changes. Like, I think it's been public news about the org changes and so on that they've announced. So, you know, from a competitive standpoint, we continue to be very aggressive. Like, that's where we see a huge opportunity. We are seeing a lot of success in all verticals, including public sector. Last week, I was at the Billington Cybersecurity Summit, which is probably one of the largest conferences in DC for public sector, and it's, you know, very senior level government officials, CISO, CIOs. There is so much incumbent technology that they are trying to displace, and a lot of those environments tend to be in their own data centers. So for them, you know, they're looking for somebody that can operate at the scale of Splunk, provide better economics, provide all the capabilities that they're used to, and do it on-prem. They also need hybrid at times, but, like, not having on-prem is not an option. So Microsoft Sentinel becomes a very difficult choice for them because it's all in Azure. Yeah. You know, we are seeing a lot of good opportunities. Great. On that note, thank you very much for joining us- Thank you, Kash. you, Kash. Appreciate it. The first day of the conference, and thank you for your involvement as well. We hope you have tremendous insights to gain over the next three days. We've done a lot of hard work to put this kind of content and programming for you. Thank you once again. Thanks, everybody. Let's give a round of applause for Ash and Janesh. Thank you.