It's day three.
Mic off.
It's day three of the Goldman Sachs Communication and Technology Conference. It's day three, but year four of this rebranded converged conference, and it's been a fantastic success. Rick's been a part of the conference from version 1.0, version 2.0, 3.0, 4.0. Here we are. Thank you for your support. Thank you to our clients. It's really heartwarming to see people from all over the world. I mean, people from Australia, Asia-Pacific, Europe, Canada, all over. Thank you for your continued support of the platform. Real delight to welcome you back. I think we met when you first made CEO of Dynatrace. You joined the company.
Four years ago.
Four years ago, I still remember that we were all wearing our COVID masks, and we had to get certified to test it to be part of a meeting. I still remember vividly that very first meeting, very memorable. Boy, you've come a long way, and the company's come a long way.
We have.
It's been a tremendous story over the last three years.
Yes, sub $1 billion ARR when I joined, now approaching $2 billion. We've made good progress over this period, and we're in a great market.
Great. Can you recap for us kind of the big milestones that you've accomplished in the last three years, and then cast for us a picture of what Dynatrace looks like in the next four to five years?
Is this the one question you're going to ask for the next 30 minutes that I can respond to?
You can explain it.
The whole meeting? That's a pretty open-ended one.
Last year, I'll give you an example. Our CEO interviewed the CEO of Salesforce, and he had a lot of questions. Presumably, he asked one question, and Mark took 20 minutes.
That was it.
That was all about agents.
I'll try not to do that.
You're welcome to do it. It's OK.
Probably the best way to answer this, I would say, is the reasons that I joined Dynatrace four years ago are very consistent with the reasons that exist today to be investing in Dynatrace, and the reason that I'm so excited to be part of it. The first is the observability market is very strong. In a world that is increasingly cloud-based, that is driven by AI workloads and beyond, the need for observability capabilities and observability functionality are increasing, not decreasing. It is becoming more complex to manage the huge amount of data and rapid increase in complexity of that data manually. You just have to have automated tools, automated capabilities that enable you to do that, and ever more in an AI world. The market is clearly a core driver. Number two, incredible customer base that we have. These are the biggest of the big companies around the globe. I've just completed a 2.5 or so month roadshow through seven countries on three continents, and the feedback from our customers in terms of the value that Dynatrace is delivering is just overwhelming. That's an incredible, incredible factor. Our platform is absolutely ready for prime time. It is really, I believe, being constructed as an extraordinary platform to provide an AI lens that hasn't existed for two years, but really more than a decade, into an integrated unified data set. That data set, by being fully integrated in a single data lakehouse, provides exceptional value add in delivering an end-to-end observability experience. Finally, from a financial perspective, it's a very durable financial model. 19% subscription revenue growth last quarter, 33% pre-tax free cash flow. If you add those two numbers together, operating at roughly a rule of 50, we've operated consistently in that sort of range. The combination of those, I think, three or four factors are what we've really leaned into over these past years to really go generate the business that we have. We see all of those four factors as really being drivers for the years to come. If anything, it's accelerating, IT accelerating.
As you look at the road ahead, how do you see the business positioned in light of a market that is not quite consolidated? I have to say that we're starting to see some signs of stabilization in your business and other companies' businesses. The market is still kind of a little fragmented. It's confusing for me, at least maybe it's not confusing for clients as to where does Dynatrace place, where does that company XYZ place. Can you lay out for us how you view your addressable market and how you see the company's opportunities four to five years out? How does this market look like at some point when you're done sorting out APM versus thirds versus observability versus infrastructure monitoring? It's like a lot of things that are going on in there. Who wins?
If you look at the Gartner Magic Quadrant or the GigaOm Radar or Forrester Wave, it just has a few external lenses. The good news is Dynatrace is in the prototypical upper right quadrant of all of them. It is due to the strength of the platform and the capabilities that I described. The real opportunity, I think, as we look ahead, that is really going to separate companies further is really an evolution of observability in, and I would say, three dimensions. These are three dimensions that I've just heard over and over and over again from customers really over the last few months of my ability to meet with CXOs globally. The first evolution is toward end-to-end observability. End-to-end observability really means a consolidation of tools and an integration of these capabilities in a single overall solution. If not a single solution, then certainly a radical reduction from 15 observability tools to two or three or something like that. We at Dynatrace certainly have been a consolidator in this arena. End-to-end observability really includes a few different levels. One level is the data level, integrating logs, traces, metrics, really user data, behavioral analytics, all of the fundamental core observability data types in one data store, one data lakehouse. Only Dynatrace does this. If you look at any of our competitors, they're using fragmented data stores. They're manually tagging data. When data flows increase in size and complexity, it becomes impossible to manage that. The second layer of end-to-end observability is really around domains. It's applications, infrastructure, log management, really users, all using the same lens to the same data. Third and finally is personas. You want not only IT Ops, but developers, SRE, platform engineering. You want all of these groups to have access to the same underlying data store so that they can then provide similar answers, similar insights. If you really do end-to-end observability right, you get two enormous benefits. One, you get better answers, better outcomes. Something goes wrong, you're able to fix it more rapidly based on better insights, more tuned to resolving the problems that you've got. The second thing is usually it comes with a 20% or 30% cost reduction, which large enterprises like as well. End-to-end observability is one pillar. The second pillar is AI observability. This is both using AI to deliver better answers, as well as being used or being able to be utilized for AI workloads, which are expanding quite rapidly as well. We may come back to this if you take us there, Cash. This really gets us into, tell us about agentic observability and agentic AI observability. What does that look like, and why does Dynatrace potentially win in that environment? Third and finally is around business observability, which is that increasingly organizations are looking to us not just to say, is my software working, and is it operational, is it performant, but tell me how my business is operating. Financial services are telling me, I want to know how long it takes to make payments. Is it working better than yesterday or worse than yesterday? Airlines are asking us to help them build control towers that define how long does it take a bag to get from the plane to the carousel, and is that doing better or worse than yesterday? Cruise lines are wanting to know what the customer experience is based on whether you could get into your stateroom consistently, or could I find you on board ship? Retailers, you know, on and on and on. Business observability is becoming core based on business events and business evolution. End-to-end observability, AI observability, business observability, I would say these three categories are really separating the competitive environment. Especially when you look at overall company size and complexity and extraordinary volumes of data, that is where Dynatrace most differentiates from others in our space.
Rick, how far along are we in this sort of end-to-end observability consolidation phase? I mean, you talked about all the tools that are out there. Maybe give us a mark to mark in how early we are in that journey.
I mean, I think we're, it's a great question. I think we're in the third inning, you know, or something like that. We're still in the early third, probably in the first third of evolution. In virtually every customer or company we talk to, there are still many, many observability tools all over the place. Some have been launched on a centralized basis that they're trying to consolidate against, especially in log management. Others have been grown internally, and yet others are deployed departmentally. They're trying to get a handle on it. I was speaking, just to give you an example, with an airline customer of ours, and they used to describe the incident management process before deploying Dynatrace. They would say, we get 30, 40 people in a room when our mobile app went down. If a mobile app for an airline goes down, everybody in this room would panic. Because, you know, I don't know where my gate is, I can't buy a ticket. It would be horrific. I mean, just imagine. What would you do? That's a bad thing. They would describe the incident process as getting all these people in a room. Not my problem, not my problem, not my problem. It's not in logs, it's not in traces, it's not here or there, it's not in infrastructure, et cetera. It wasn't particularly efficient. What they really were trying to do in this example was consolidate the data, the domain, and the personas to be able to get this end-to-end lens and observability to deliver better outcomes. I'm proud to say that with Dynatrace, they've seen an extraordinary evolution in success in the number of incidents, the amount of time it takes to address an incident. British Telecom was an example also where they reduced incidents using Dynatrace by 50%. Of the remaining incidents, reduced the amount of time it took to resolve an incident by 90%. These kinds of metrics you really only get from a consolidated end-to-end environment. Those organizations that have moved to this sort of state are the ones that really, I think, are much more aggressively taking on the opportunity of improved overall software operations.
Very interesting. I want to touch on the log management opportunity. Grail has been a very exciting part of the Dynatrace story for a while. At the same time, it seems like you have security vendors going after this opportunity. You have observability competitors going after the opportunity. What's Dynatrace's right to win in this category?
What a great question. I would say that logs is one of our biggest opportunities as a company. It's certainly one of our biggest growth drivers. It will be our fastest business, fastest new business, to $100 million in consumption, not quite there. We'll report to all of you when we get there. We're well in the way to achieving that goal, and the opportunity is immense. What do companies want from Dynatrace vis-à-vis logs? The answer is twofold. One, I come back to the end-to-end observability comments I made, which is it is all about outcomes. It is about answers that can be trusted and acted upon. The more data you have, the more data types you have, the better your insights. I felt this four years ago when I joined Dynatrace. I feel it today. The way the market grew up to have logs separate from other observability data types just by way of company deployments didn't make any sense to me, because having logs as part of traces, metrics, really user data, other analytic data types gives you more data to be able to ascertain what the real root cause issue is. If you have an automated AI engine that is inspecting all those data types, you're going to get better outcomes. If, on the other hand, you're using a separate log vendor from an observability vendor or core observability vendor, then you are doing all of that manual correlation yourself, and you're probably not getting meaningful contextual outcomes. First and foremost, I think organizations are looking to simplify their overall environment of vendor and vendor management. They're also very much looking for better outcomes. You're going to get better outcomes if you have all the data types integrated into one overall data lakehouse, and that's why moving logs in with your observability data types is a significant advantage. The second piece of it is cost, which is, without getting into details, there are log management vendors out there that have had, if not a monopoly, a near oligopoly for a long period of time. I was down in Australia a couple of weeks ago, and I had a large Australian bank say their cost on one of those particular vendors was meteoric, with limited value, with limited incremental value that is not commensurate with the increases in cost of the overall log management system. We can provide a pretty material reduction in cost while producing better outcomes for observability logs, and this becomes a significant advantage. This is why we are seeing such immense growth. Log consumption for us is growing more than 100% year -over -year. It grew 36% quarter -over -quarter for us on increasingly material numbers that we're very excited to see. We see that evolution becoming durable.
If I could just interject very quickly there, what is the technological breakthrough that is allowing you to displace the traditional vendors? What is it that they have not done that Dynatrace has been able to do that is allowing you to get that price performance advantage?
Yes. The breakthrough that we really delivered was Grail. Our underlying data lakehouse, which stores all observability data types, inclusive of logs, in context in one location, is what enables us to then apply our AI engine to that data analysis. In doing that, that's what gets you better outcomes. That was quite fundamental. It was the maturity of Grail to the point where it was able to handle logs at extraordinary scale, which really for us happened in about October of last year. In some ways, we're still in sort of the first year following our delivery of that capability. It is since then that we've seen the really marked increase in log capability on the platform. That was a primary driver. I would say there are other elements just in terms of cost and pricing models. We didn't have an innovator's dilemma problem in pricing. We didn't have to say, we're going to reduce prices by 30% and take a 30% reduction in revenue. We could come out with pricing models that enabled us to grow and grow with the market. For example, one of our pricing models is called a queries included or included queries model, where you can specify a number of days, and we'll give you unlimited queries on the log data set during that period of time. It could be 15 days, 30 days, whatever you want. That just eliminates the uncertainty regarding, oh my god, you know, I'm having to throttle usage of logs within a company of enormous scale, which is very difficult to do. Instead, say, have at it. Third and finally, our architecture with Grail and the overall platform makes no distinction between cold storage, warm storage, hot storage. You see other vendors in this space sort of like, if it's more than 30 days, it's cold storage. I can't access it. Others are warm storage, which maybe I can access it, but it's delayed. For us, storage is storage, and we're going to access and provide essentially hot storage access to logs at all times. The performance is exceptional relative to others on the market.
I want to switch to kind of the go-to-market changes that you've made over the last kind of year and a half. Can you give us an update on the progress that you've observed and where we stand with that evolution?
Yes. We've added a number of people in the Salesforce over the second half of last year. This is our fiscal year, so that would have ended March 31. Those people, we still have roughly a third of our Salesforce within their first year of tenure. We did this on purpose because we felt that the market opportunity justified the expansion in the Salesforce. Those individuals are still in their first year, and it usually takes 9 - 12 months to get up to speed. To become fully productive, those individuals, as we get into the second half of our fiscal year, we really expect to start showing growth and productivity.
Right on down, right?
Yeah, once we get into the December quarter, you'll really be in the second half of this fiscal year for us. We expect to see that.
Imagine telling me, Cash, your models are wrong. We need to take our numbers up for that.
Yeah, no, I'm not going to comment on that, Cash. What I would say is to give you a very tangible example of that. It used to be the case that we had reps, a strategic rep that would have eight customers. These are mega customers. Yet when we looked at it, they were making their number on four of them. This is just using averages, obviously. The other four would be highly productive accounts, and they just didn't have the time to get to them. We felt that there was more territory capacity to expand to that. That's where we've added the reps, and that's where we expect to see the productivity enhancement as we get into the second half. That's part of it. Another part of it is selling logs. We talked about that. It is an expansion of the portfolio to really sell across the board. We're seeing, as I mentioned earlier, good traction on that. Thirdly is partner evolution. We're seeing like one of our global system integrators has a pipeline that is 2x or 3x what we expected it to be. We're seeing more and more capability out of the partner channels, in particular in GSIs as well as hyperscalers.
Maybe talk about the kind of push up market into these strategic IT 500 accounts in context of potentially longer deal cycles. I think over the last two quarters, you've disclosed some pretty punchy statistics around pipeline growth, you know, 40%, 50%. Maybe just talk about that dynamic.
It is, as I mentioned earlier, Dynatrace wins at the largest organizations because we have the most differentiation there. It's not to say we lose at the lower organizations, but where we most win, where we have the biggest differentiation, is at large organizations because of data, magnitude, volume, and complexity. That's where organizations that are using other vendors, other partners for observability move to us. As we look at the opportunity to come, it really is of significance for us to be leaned in in the partner community, to be leaned in on the evolution of this particular set of cohorts to drive it. What happens as a result of that is deal sizes grow. As deal sizes grow, you have more variability because if you are delivering $50 million, $60 million, $70 million of net new ARR in a quarter, depending on the quarter, and you have $3 million, $4 million, $5 million deals. Last quarter, we had 12 seven-figure ACV deals. The loss of one or two of them has pretty significant downside risk, and the gain of one or two extra has pretty significant upside risk, upside opportunity, I should say. There's more variability. As a result of it, we are conservative in our guidance because of that variability. That does cause or drive some of that conservatism in our portrayal of guidance so that we're assured that we can hit the numbers that we provide to the market.
Yeah. I want to shift to DPS with the last 11 minutes.
love DPS
We spend a lot of time here.
OK. All right. Good.
Forty-five percent of customers, 65% of ARR. What has been the main driver of DPS adoption, key learnings to date that you can implement moving forward?
For those of you not as familiar with the story, DPS is our Dynatrace Platform Subscription. This is a new mechanism that we launched two years ago or so in our pricing model. It was in reaction to customer feedback that our prior licensing model was cumbersome and arduous to expand because we did it based on host units. You might buy 10,000 host units, and you came back and you needed 2,000 more, and it was a new contract or a contract addendum. It was not very fluid. What the Dynatrace Platform Subscription does is it gives you access as a customer to the entire platform for a specific amount of consumption commit over an annual period. Sometimes it could be a multi-year contract. What it has enabled is that it has enabled a customer, if they, in my example, if they needed 12,000 host units, immediately they had it. It was there. They just start using it as opposed to contracting and coming back and iterating through it. It also gave them access to the full platform. Maybe they were only using application performance monitoring, but now they have access to infrastructure monitoring, log management, application security, and beyond. The result of all this is that DPS customers are showing consumption growth that is 2x the consumption growth of our non-DPS customers. Moreover, DPS has much more rapidly accelerated and penetrated the installed base than we saw previously. Now, as you say, 65% of overall ARR is already on DPS. We believe that the result of this is consumption being a huge driver of future opportunity and one of the primary growth drivers. In fact, consumption for us, which we're now starting to communicate more broadly, is so compelling that we're starting to share those numbers. Consumption for us growth is growing in excess of ARR growth and subscription revenue growth. As I mentioned, subscription revenue growth is 19% per quarter, 19% in Q1, year -over- year. That would indicate to you that consumption growth is in the 20s. That's where we see it growing. That is a huge leading indicator to future opportunities to then provide upgrades to customers because their consumption is growing into their DPS contract, sometimes beyond it. You have to repackage it or renew or expand your DPS contract. Consumption is really an important growth driver and precursor. If you take together the log opportunity, the consumption opportunity, the pipeline numbers, and DPS, and you put those four elements together, it gives you a pretty good sense of the growth drivers in the business and the opportunity we believe we have as a company.
How are you incentivizing the Salesforce to help ramp consumption with their customer base after having landed a DPS contract?
We have strike teams now that we have deployed that are in our combination sales and customer success organization that are compensated on consumption exclusively. In fact, our vast majority of our services organization, Customer Success Managers, Customer Success Engineers are all paid on a variable basis based on consumption. That is a change that we made this year.
We're just two quarters in.
We are on a major pivot to really driving consumption. It is the combination of this consumption plus DPS model that we believe really has an opportunity to unlock future revenue potential in the company, both in the installed base as well as new customers.
Can you talk about, there's a bit of a wrinkle with the on-demand consumption piece versus renewals? Where are we in that journey?
Yep. How much time you got to answer this question?
Seven minutes.
Yeah, seven minutes. I'll do it in one. You know what is, again, for those of you familiar with the story, in a DPS contract, you get basically annual chunks added. Let's say it's a $1.2 million annual contract. We would recognize revenue on that in our model rateably, $100,000 a month over a 12-month span. If you achieve that $1.2 million in consumption as of month 10, then you have two months where you need to make a decision. Either I'm going to pay you two months on actual consumption until the DPS contract renews at the $1.2 million beginning in year two, which you can do, or you can renew the contract. The incentive to renew the contract is you usually up the commit. If you up the commit, you get a lower unit price. It benefits you by renewing and then getting a lower price point at a higher volume. We would have said that, you know, everybody's going to renew. We wouldn't have said everybody's going to renew. We thought that the vast majority of customers would renew that to get the lower price point. When you're dealing with customers the size of our customers, you know, the mega customers around the planet, or the mega companies around the planet, they put in a three-year DPS agreement. They're like, I don't want to have to go through another contracting process when you're in on a three-year contract. They're more than happy to just pay on a consumption basis for those couple of months and then allow the DPS contract to renew. It is a combination of subscription revenue, really a combination of ARR commitment plus the sort of stub period consumption elements that exist in a DPS contract.
Yeah, that's helpful. I want to flip back to the AI piece with the last five minutes that we have here. You talked about the explosion of data and complexity driving demand for observability in the age of AI. Can you help frame for us a bit, either qualitatively, quantitatively, how much of a tailwind that can prove to be for Dynatrace in sort of the near to medium term? Because you've got the product angle in terms of AI observability, but then you also have the knock-on effects of more data complexity driving increased demand for IT infrastructure monitoring, APM, so on and so forth.
Yeah, it's important to sort of bifurcate when we talk about AI observability into two pieces because they're different. The first is the ability to observe AI workloads. We already have hundreds of customers, existing customers that are observing AI workloads with Dynatrace today. Observing those workloads to some extent is just like observing any other workload. Plus, and the plus is you also want to know things like, is it hallucinating? What are the guardrails? What's happening in inference? There are added elements to observing AI workloads beyond just the normal workload that you would observe that we put into the model. That is a significant opportunity as we see explosions of AI workloads at even our existing customers, let alone AI-native companies. The second evolution of AI observability for us is we really do believe that the vision, the long-range vision for Dynatrace, and what we believe quite fervently is the ultimate outcome that organizations want out of observability is a true autonomous AI observability platform that basically can take corrective action before an end user ever sees an issue. It shouldn't be about incident reduction and MTTR or meantime to resolve issues. It should be, wow, look at that. We eliminated incidents altogether in this category of problems because we fixed it before anybody ever saw it. The way to do that is through deep insights and an understanding of the appropriate answers that are trustworthy. For that, we believe you need Dynatrace. It is because of that whole integrated platform structure that I talked about that delivers answers, not just dashboards, based on causation, not correlation, so that you can determine with certainty what the issue is. Every one of our competitors is going to be talking about agentic AI. I would submit to you, as I tell our customers, as humbly as I can do it, that you cannot take agentic AI action if you can't trust the answer. In order to take action, you have to be sure that you're fixing the problem that actually was created. You have to start with a trustworthy answer. If you start with a trustworthy answer, then you can take action. Moreover, we believe it will be an autonomous ecosystem, not just Dynatrace. We're certainly not saying we're going to take every action. In fact, quite the contrary. We see a problem in the system. We can push that through an MCP server. That can be picked up by Jira to fix a line of code, by GitHub, by ServiceNow, by a hyperscaler. You may need to provision more storage. You may need to do an application rollback. You may need to fix a line of code. There could be a variety of outcomes that would be taken by other agents. Our agents can determine what the issue is, what action should be taken, to then submit that to allow other agents to take action on those trustworthy insights. It is that that is so exciting for us, I think, at Dynatrace, because it is our foundation that really enables that to be a closed autonomous ecosystem that really could be quite compelling.
I'm curious, as you guys move from simplistically like root cause identification to like autonomous remediation, how does Dynatrace make sure that they're capturing that incremental value from a pricing perspective?
It's a great question. Ultimately, we need to make sure that we're delivering the metrics and the value proof points to ensure that, wow, it isn't just my software is working so much better, but what did Dynatrace do to contribute to that? This is where we see it to some extent in consumption across the board.
DPS should help, right?
DPS should help. By the way, those consumption numbers are really not, they're not just in DPS. We're seeing consumption expansion across the DPS and the foundational elements, so really across the platform. That is probably the best indicator to the answer to your question.
Final few seconds, anybody has any question?
I have one question. AI natives, are they trying to build something on their own because they're so smart and they've got all the technical expertise?
I mean.
To the exclusion of entertaining Dynatrace in their IT org.
Sure, presumably they could. I think the number of customers at this point that are going to build their own observability system with the level of complexity that we've designed into Dynatrace with integration at data layer, domain layer, personas, autonomous AI observability engines and capabilities, that is just unrealistic. I would never say never. It's tough. Even the hyperscalers I get asked about all the time, would the hyperscalers do it? The problem is Dynatrace delivers in a hybrid cloud environment. It's on-prem, off-prem. It delivers in a multi-cloud environment. Across hyperscalers, is one of the hyperscalers really going to build a hybrid cloud, multi-cloud environment for observability? I don't see that. I think it's going to be in the vast majority of companies willing to take that sort of action.
On that note, thank you so much for your support of the conference. Thank you to our clients, and enjoy the rest. Another day and a half left behind.
Thank you to everybody for joining. Appreciate it.