Time's running. Okay. Hey, welcome to our next session. Really happy. I didn't realize, Rick, that you were just based here, so that's even better.
Yeah, based in the Bay Area, so this is very proximate.
So, it's a question I'm asking everyone at the moment, and it's, how does it feel out there in terms of end demand? Because, like, yes, there's the stock market, you know, and we can debate that all day long, but, like, it's difficult to find answers at the moment. But how's the real world?
From my perspective, not a huge amount of change in macro. Obviously, incredible amounts of deployment of capital being spent on data centers. We're all very aware of that. I haven't really noticed or observed much change in the overall spend environment for enterprise-oriented software. At the moment.
If you think about it, like, the bigger question, sorry, the big thing going on for you guys is that you're talking more about strategic deals, that you know the customer conversations are changing, etc. Talk a little bit about what's driving that from a customer perspective, you know, and maybe start with, like, where the industry is as a whole and then how you can help the, you know, customers on that.
Sure, well, the evolution of the selling environment for us has evolved immeasurably, even in my four years of tenure at Dynatrace. When I began, it was very siloed. You had the APM vendors, and then you had the infrastructure monitoring vendors, and you had log management vendors and application security vendors. Over the last four years, really, that has consolidated in a material way. The real question is, well, why? Why is it?
Yeah.
Why is it consolidated? And my simple answer to that is outcomes are better. That in a highly complex environment, especially with our target customer set in the Global 15,000 , really even, the Global 2000 or 3,000 at the high end of the pyramid, huge amounts of complexity, hundreds of applications, huge amounts of infrastructure which are growing immeasurably, even more so in an AI world. You have a log management coming into the fray now, which is getting integrated into the overall observability framework. And the result of it is you can't manage it manually. You can't do what you used to do.
When I was at an oil and gas company and the principal walked me into the network operations center and there were hundreds of people staring at hundreds of screens, trying to do manual processing of alerts, that doesn't work anymore. You have to automate that. It has to be.
Yeah.
Oriented around an AI world, and it's gotta get integrated into an overall view of your software environment to deliver the best answers and the best outcomes. And that's why it's consolidating. And that end-to-end observability trend, uniquely, maybe not completely uniquely, but it certainly benefits Dynatrace.
And then, on that note, though, because, like, if we've been talking about full-stack observability for a few years now, and I was like, "Yeah, yeah, I want to do that." I remember, like, I wrote my first big note on observability, like, five years ago, and it was like, "That's, you know, what everyone is doing.
It was brilliant.
Yeah, yeah. Thank you. Yeah. But it was, it's only now coming kind of together. Is that kind of driven by product maturity or, like, back then it was all marketing where you knew you need to get there, but the products weren't there? Like, how do you have to think about that?
It really is an evolution, I think, of capabilities. It is the product capabilities. Like, for example, at Dynatrace, we have, over the last five years, evolved our platform to our third-generation platform. This resulted in, a couple of years ago, the delivery of Grail.
Yeah.
Grail for us is an underlying foundational, massively parallel processing data lakehouse. It stores logs, traces, metrics, behavioral analytics, business events, user data, that sort of thing. That data lakehouse is overseen by a fully integrated, completely AI-oriented or AI-powered platform, which is Davis, which does causal AI, predictive AI, generative AI, headed to agentic AI. We can talk about that if we wish, in a bit, and that has resulted in delivering answers that are trustworthy and credible to then enable automated action, and it is this stack of the data plane, the AI plane for answers and outcomes that are able to deliver automated action that's really come together over these last years. That sort of capability didn't exist for Dynatrace or really anybody else four or five years ago.
Yeah. Okay. Perfect. And then so you have, if you go to your customer, you have it, but now, like, look, take Barclays as an example as well, but, like everyone, we're still on-premise. We're still messy. How do I kind of go from that messy world to your kind of new fancy world?
Hey, this is one of the advantages, not in overselling it, but this is one of the advantages of Dynatrace. In that, our capabilities exist on-prem, in the cloud. They exist in multiple cloud environments, AWS, Azure, GCP, etc., so you can manage your migration from on-prem workloads to cloud workloads to AI-native workloads seamlessly with Dynatrace, and you can oversee all of them using our observability framework at your discretion and at your rate of expansion or migration, and so by enabling that, we really do facilitate a mixed hybrid environment that continues to evolve, and in the largest of enterprises, that really matters because the largest banks on the planet, the largest commerce companies, and the largest pharmaceutical companies, they're gonna have a mix of environments forever.
Yeah.
or at least for a very, very long time to come.
Yeah, yeah. So then is, so if I'm like a big enterprise on-premise cloud, so I can start consolidating at home towards Dynatrace, as I move to the cloud, I can use Grail, or can I use Grail on-premises as well? Or is that more cloud-core? And so as I move to the cloud, I just do more and more Grail and such?
One important distinction, Raimo, is that.
Yeah.
Is that even if you have an on-prem workload, you can observe that in the cloud?
Oh, I see.
That enables Grail.
Yeah, yeah.
Yes, the observability needs to be happening in the cloud, but that can apply to all sorts of different workloads.
Okay. Perfect. Okay. Makes sense, and then you mentioned log as a thing that came up a lot. That was probably the last leg to bring all the stuff together.
Yes.
You start talking more about logs and, you know, like, where are you on that journey? And I'm asking because it does look like there used to be a log company that was very big, that might where a lot of deals might come up for renewal. Like, what are you seeing there?
Logs has been the fastest-growing business that we've seen. We really got to the point of enabling logs, I'd say, that were ready for prime time at enormous scale beginning in about October of last year, so October of 2024. At that point, logs was a very small sort of single-digit millions-of-dollar business for us. That has grown to the point where in our earnings release for our second quarter earnings that we just completed a month or so ago, we talked about that number being almost $100 million in consumption.
Wow.
So it is, logs is essentially a business that has grown from zero to near $100 million in one year. And it's growing at well more than 100% per year. Well, 100% growth on a $20 million number is a lot smaller than 100% growth on a $100 million business. And so that really is fueling a good amount of growth. And I would say it is driven by two factors. One is priced relative to incumbents in the market. But it's not quite the way that you would think about it. And without overly belaboring it, I would simply say that it isn't that we're just coming in and saying, "Well, we're a lot cheaper." It is that if you incorporate logs, traces, metrics, other analytics, user data, you actually don't need to store as many logs in order to get better outcomes, better answers.
Oh, yeah, yeah.
So the result of it is that for enterprise customers, you can actually get more efficient in your overall cost without us coming in as Dynatrace and just say, "Well, we're gonna give you a 30% discount over what you're doing today. Just continue it the same way." So the two drivers of logs business for us are, number one, overall cost of managing a logs environment, as I just discussed. And then the second is delivering better outcomes. And that better outcomes piece is, I would say, much more relevant, especially as you head into an agentic world, for autonomous operations, because this is where you want all data types to be delivering the answers that you need to be able to take that action. So you can think about it, end-to-end observability really consists of three different layers: integrated data layer, integrated domain layer.
This is application performance monitoring, infrastructure monitoring, log management, etc., and then the third layer is the persona layer. You want developers, IT Ops , SRE, platform engineering, even executives, all to be able to access that same data. That is what end-to-end observability is. That is what enables the best outcomes, and that is ultimately what will enable automation of that environment.
And can I, because I'm the financial guy, so I look at numbers.
I don't believe that.
Yeah, yeah.
I think you've got way more to.
The.
A lot of it in there.
But if I just look at numbers, the big log guy who is like massively bigger than.
Yes.
Thank all of you guys.
Yes.
In this new world, if I'm doing less logs, like, is the market the same size up for grabs? Is it gonna be a bigger market? Is it gonna be a slightly smaller market? How do you think about that?
Gosh, you know, Raimo, I look at this as sort of like the mainframe market that was always going to disappear.
Yeah, yeah.
That never did.
Yeah.
You know, it was supplemented with client-server , and then it was supplemented with cloud, and then it was.
Yeah.
And I think this is the same. We can get more efficient with logs. But gosh, you know, as AI workloads grow, almost in an unbounded way, what's gonna happen? The overall workflows to be observed is gonna increase immeasurably.
So, okay. Yeah, yeah. Okay. That answers that. And then, since we talked about players in the market, we just saw, obviously, it's an interesting market, and, you know, if something's interesting, then other people wanna share of that. So we just saw, like, Palo Alto do, like, a kind of a step into the with Chronosphere into your market. How do you think about kind of how this will play out? Because, like, at some to some point, they look at the same data set for some of it. On the other hand, not fully. I think your data volume is much bigger. Like, how do you think about how that's kind of going to play out?
I would say, several comments on this. First, it provides absolute validation of the observability market.
Right.
Being ready for prime time.
Yeah, yeah.
You know, again, four years ago when I began at Dynatrace, it was, you go talk to customers, and it was difficult for me to get a meeting with a CIO.
Yeah.
You know, it's sort of observability. How do you spell that?
Yeah, yeah.
that is no longer the case, I assure you. In the last 30 days, I have done dozens of meetings with CIOs, CTOs, and others. Observability is ready for prime time.
Yeah.
It is broadly understood that you cannot do it manually. You cannot engage in the same way that you used to be engaged. So that was one element. The other element is just the evolution of AI is exploding these workloads and requiring more and more observability as we go. And so as we look at these components, I think all of these are factors in how the market is.
Yeah.
Is evolving. And it's evolving very, very rapidly. And relative to the Palo deal specifically, I think it also validates the merging of application security with observability. We've been talking about this for some time.
Yeah.
I think it is inevitable that we see that convergence, and we've looked at that as well. The one thing I would say is that I do think it's not without its challenges. The reason for Palo is because the buyer profile is a little bit different. Selling to the CISO or the CISO organization is very different than selling to the CIO organization. I believe that the buyer for observability is different than the traditional buyer for Palo. So cross-sell is not something that's as easy to get done there.
Yeah. Okay. Perfect. Hey, guys, could you in the back of the room? All right. Could you just kind of quieten it? Thank you. The next thing is, like, you mentioned GenAI already a little bit. Can you just, like, you guys have been doing AI, like, for a while, and there was more kind of anomaly detection, etc. Like, how's GenAI kind of play into this internally?
Yeah. Let me answer it this way. What I would say is that we sort of view the observability market having evolved through several phases. In a way back, way back when, a decade or more ago, it was really reactive. It was, "Wow, something broke. What happened? And then you start doing research.
Yeah.
It moved from there to maybe what I would call phase two, which was proactive, which was automated root cause analysis. Something went wrong. You had immediate root cause. You took a look at it. You knew what was gonna happen, and then you resolved it as quickly as you could. The third, which is sort of where we are now, is predictive operations, which was anticipating issues but still resolving them manually. So you were using causal AI, predictive AI. Predictive AI was adding in machine learning, anomaly detection, those sorts of elements to anticipate problems so that you could get them resolved before end users would see them. Where we are going next is to a world of autonomous operations, and this is where it really gets exciting, which is you take predictive operations to the next level.
Yeah.
That is to not just see it, anticipate it, cut it off before end users see the impact, do so manually, but rather through an agentic world, an agentic ecosystem. You could actually handle those sorts of elements on an automated way through MCP servers, auctioning out through various different other agents. It could be a ServiceNow agent. It could be an Atlassian agent. It could be a GitHub agent, a hyperscaler agent.
Yeah.
And then resolving issues really in real time before they become an issue. What's amazing is of the companies I've talked to recently, they get the fact that you don't need autonomous operations on 100% of predictable incidents in order to matter. If you get to the ability where you can autonomously react to, respond to, and resolve 20%, 30% of the potential incidents for the largest organizations on the planet, that can save them tens of millions of dollars in annual cost savings.
Yeah.
Not to mention the user experience benefit that they get by not having the challenges that they might otherwise see.
So that's one aspect. The other aspect I wanted to speak to is, like, in theory with GenAI, like, the whole idea is, like, there's gonna be a lot more software, there's more applications, etc., which basically means there's more observability that is needed. Is that a fair statement?
Absolutely. I think that's exactly right. You're gonna, to the extent that you can write code so much faster, deliver so much more, in the way of applications, you end up requiring or mandating a lot more infrastructure. The result of all of that is an explosion of workloads.
Yeah.
The more workloads you have, the more you need to then be able to observe those. You know, by the way, there is an added element. You take a look at what we would refer to as AI observability.
Mm-hmm.
And AI observability is what we refer to as observing AI workloads. You have to do all of what you do today with regard to all of the other workloads. You have to.
Yeah.
You have to observe applications. You have to observe infrastructure. You have to do log management. You have to do all of what you do today. But you also have an added mandate, an added ask from customers, and that is, by the way, could you please make sure that the content that is delivered is accurate? And this is where you get into hallucinations. You get into guardrails. You get into these sorts of elements of, please tell me also, can you validate the data that is coming out of the answers that are being delivered? And that makes it even in some sense more complicated, but also more mandatory.
Yeah.
That observability becomes a requirement.
That sounds all really interesting. Like, at the moment, if I talk to the industry players, there's a little bit of a bragging competition going on of like, "Oh, yeah, I work with this." What are you? What's your kind of kind of.
Differentiation or position?
Yeah. Differentiating position. Do you talk about, like, AI with, you know, customers or whatever? What's the situation here for you?
The pithy phrase I've been using to denote the answer to this question is really, "Answers, not guesses.
Yeah, yeah, yeah.
What Dynatrace delivers is answers, not guesses. We always have. This is why the biggest of the big, the most complex of the organizations around the planet tend to buy Dynatrace. And they do so because in Grail, we can handle billions of interconnected data points in context, manage them across all data types, across all domains, and give you an answer that is highly contextual, and that is causal, not based on correlation, but causation instead. The result of that is, with a very high probability, we can tell you precisely what the answer to the problem is. Why did it occur? And we will tell you precisely what that answer is. We don't guess based on correlations. We know based on the environment that we've structured.
Yeah.
And for those of you that have used generative AI and some of the overall LLMs, and you've sent out data, you've asked questions, and you've gotten wrong answers back, you know that there are issues with that underlying trustworthiness of data. In our case, in order to take autonomous action, you must trust the answers to why the problem is occurring because otherwise you're solving the wrong problem.
Yeah, yeah.
And the last thing that you can do is have some sort of autonomous agent or agent ecosystem take action on solving the wrong problem. You have to trust the answer. And if you trust the answer, you can then enable autonomous action. And what Dynatrace does uniquely, it is our superpower, is to be able to tell you precisely what the action is. Not to leave you a dashboard that's red, yellow, green. Not to tell you, "These are the things that are happening in your environment that you may wanna look at based on correlation," but rather, "This is the answer. This is the cause.
Yeah, yeah.
And that enables you to have the trustworthiness of that answer to take action. And that is what is enabled in an autonomous world. And it is obvious that every vendor in observability, you know, you take a look at a Gartner Magic Quadrant, there are 20 of them. Why is Dynatrace in the far upper right of that? It is because of our ability to deliver answers. Foundationally, that's a great starting point. And as you move into an agentic world, well, every one of those vendors is gonna be talking, "Well, I've got an agentic system," because you have to say that.
Yeah, yeah.
The ability of those agentic ecosystems to actually take the right action is predicated on the ability to deliver the right answers. And we would submit that Dynatrace does that better than anybody else.
Okay. Okay. Perfect. And then shifting gear the next few minutes, so there's the product, and the product has evolved. Now let's talk about the organization because the organization needed to be evolved as well. And there's two aspects here. There's the platform and platform pricing, and there's the go-to-market.
Yeah.
Those are like about a year to two years both.
Yeah.
Both of these initiatives. Can you speak to kind of how this is coming together for you on both kind of vectors?
So go-to-market was one, and the other one was.
Pricing. Pricing. Yeah.
Pricing. Let me attack go-to-market first. On the go-to-market side, we simply realized that it is the largest of the largest organizations, the most strategic organizations that needed observability from Dynatrace the most. That is where we had the greatest differentiation. Those are the ones that are moving to end-to-end observability. Those are the ones that are driving AI observability. Those are the ones driving business observability. This is where Dynatrace wins.
Yeah.
We restructured at the beginning of 2024.
2024.
Our overall go-to-market to really focus and attack that. And that's resulted in substantial benefit. For example, our pipeline we reported this last quarter in strategic accounts, these large accounts, was up 45% year over year. The number of seven-figure ACV deals that we closed last quarter was up 53% year over year. So we are seeing enormous traction in the biggest of the big because that's where the complexity is, and those are the ones coming for end-to-end observability from Dynatrace.
Mm-hmm.
And so I'd say that has been a major evolution in a positive way for the overall environment. Pricing has been part of that. So covering the pricing component, it was funny, but when I first began and I would ask customers, "Jeez, what do we need to do better?" one of the elements that came back was, "You need to do pricing and licensing better.
Yeah.
And it was because we had SKU-based pricing. And if you bought application performance management from us and you wanted to buy infrastructure management, we'd have to come back and add a contract and sell more. And that was painful. And God forbid you want log management because that was another contract, more painful. So what we did was we put in place our Dynatrace Platform Subscription, DPS, as we refer to it. This is basically an annualized commit. It could be a three-year contract, but essentially an annualized commit that covers the entire platform. We've seen enormous traction in it, as you know, Raimo. And it has resulted now in 70% of our ARR just over the last three years becoming DPS in orientation, more than 50% of our customers, 70% of ARR.
And it has substantially broadened the access to the platform. So, for example, one of the reasons log management I would submit to you has grown so rapidly for us is we've eliminated the contract overhead of that. If you have a DPS contract, you have access to log management. Just that simple. Turn it on, enable it, begin to use it, and watch it grow. And it makes it seamless for customers to then replace other third-party solutions.
Can I, as we go through this now, like, there's one thing that I'm talking with investors about that's kind of always raises the question. It's like your product story is coming together. Your go-to-market is coming together. The pricing is coming together. But the overall growth numbers haven't changed.
Yep.
What's the disconnect?
The good news is we look at leading indicators. Consumption is one of those.
Yeah.
We see consumption in the low 20s, and while we don't wanna make the overall metrics more complicated than they need to be.
Yeah.
which is some of the feedback that we've clearly gotten is, "Jeez, you know, don't provide too many metrics because we wanna make sure that we're tracking the ones that are most important." Make no mistake about it, the ones that are most important is ARR growth.
Mm-hmm.
The fuel to ARR growth is net new ARR. Net new ARR comes from a number of metrics, but inclusive of new logo ACV and expansion and RR and those sorts of elements. One metric that we did think was relevant to share was this consumption metric because as consumption grows, that provides the fuel to then lead to future upgrades and future expansion.
Yeah, yeah.
So, we see consumption growing over 20% as a relevant leading indicator of where ARR growth should converge to. Certainly, I'm not suggesting that ARR growth immediately moves to north of 20%. But we believe that especially logs and application security and some other elements, growing well faster than that, provides the fuel to future ARR growth. And make no mistake about it, our objective is to re-accelerate ARR growth.
Yeah.
As we head into FY 2027. We're certainly not at the point of providing FY 2027 guidance, but from a.
That's the idea.
Overall strategic and conceptual perspective, that's what we're.
Yeah, yeah, yeah.
That's what we're looking to achieve.
Yeah. Because like at some point, it needs to show up. Like these and full disclosure.
Absolutely.
Yeah. I see it also in my client conversations on the Dynatrace clients. Like, they love it. Yeah. So yeah. Okay. The one thing that you probably got feedback as well in terms of complications or where you're different is if a client overuses.
Yeah.
On the platform side, some of your peers would say, "Well, then I'll give you like the full, you know, or like a higher price. And then that way I encourage early renewals, have bigger discussions." You don't punish clients. And on the one hand, it's good, you know, like you because you want adoption. On the other hand, it kind of maybe takes away a little bit one of, you know, the upsell opportunity.
Takes away the stick.
Let's stick a little bit. Like, talk a little bit about your thinking there.
In the largest of enterprises, it has been our observation that you don't hammer customers with a stick. It just doesn't work very well, so we obviously are going to charge you overage if you go over, but we haven't been punitive in that regard, so we don't charge 120% of your cost to force you to do an upgrade. We have customers that have three-year contracts, and if they're one year in or 10 months into their first year and they're running at overage levels, the last thing that an IT Ops team do wants to do or a CIO wants to do is go back to the legal department and say, "I did a three-year contract.
Could you please renew this contract after 10 months?" So we have our account execs obviously have all the incentive in the world to come back and sell you an expansion, so they can do that. And in many cases, that will happen. And in some cases, they'll just pay us overage for a month or two. And then they'll.
And then you just rotate into the next year.
Yeah. Either way, we're largely indifferent as to how that goes. Ultimately, that is gonna show up in ARR.
Yeah, yeah.
And that's. We play the long game. And the long game is ARR growth. And that's what we wanna achieve.
Okay. And then what drove the decision to kind of try to smooth that overage out a little bit, you know, because it was.
Accounting.
Accounting.
Is the short form.
Yeah.
We, for those of you not tracking the short form, it is we were showing ODC revenue, on-demand consumption revenue on an as-incurred basis. And it showed you precisely what the overage was in that quarter. Because we do Rev Rec ratable on an overall ratable basis, given that as opposed to a direct consumption basis, that would suggest that we would have to do ODC in the same way. So we had to move it to an accrual basis, which flattened the ODC. So now you can't see the variability in ODC as you look.
Yeah.
As you look at it, it takes away one of those metrics. But on the other hand, it smooths the curve.
Yeah. And then last question from me before I let you go. Like, how do I think about that, that margin versus growth envelope now for you guys going forward?
We've committed to maintaining the margin level. But our number one focus is accelerating growth. So we wanna achieve both. You're not gonna see, or at least our current plans don't suggest a lot of margin expansion, accretion. We wanna be able to use that powder to accelerate growth. We believe that the market is ready for prime time. AI is driving it. We're in the midst of this sort of tornado. We wanna take advantage of that. And we wanna lead in observability as we look at.
Perfect. That's a great closing statement. Thank you.
Yeah.
Perfect.
Thank you.
Thank you. Thank you. Good to have you again.
Thank you. Yep. Good to see you.
Thank you.
Thanks, everybody. Appreciate you coming.
Thank you.