Dynatrace, Inc. (DT) Earnings Call Transcript & Summary

All right. I guess we can get started. Welcome, everybody. My name is Koji Ikeda. I am one of the software analysts here at BofA. I am super, super, super thrilled to have Rick McConnell, CEO of Dynatrace, here with me. Thanks -- thanks so much for doing this. Super, super appreciate it.

I guess first question, always the level-set question from a very, very high level, for those in the room that are new to Dynatrace and for those on the webcast that are new to Dynatrace, what do you guys do? What's the opportunity you guys are going after? What is observability? In 30 seconds to a minute and tell us a little bit about yourself.

In 1 minute?

Yes, please.

Including myself? Well, that sounds like a good start. So welcome, everybody. Thanks for joining us. Dynatrace exists to help create a world in which software works perfectly. So we do this for enterprise customers. The market that we access to do this is observability, which we define as a $50 billion or so market for observability and application security. And it really gets fueled off of this notion that digital transformation and cloud migration is fueling a huge explosion of data with a massive increase in complexity. And the combination of those elements mandates automated response, automated review of that data so that it becomes more actionable. And that's what we do at Dynatrace that is differentiating relative to others in our space.

Got it. Got it. And I want to ask you a couple of set questions that I've been asking every management team. One about the macro and other about AI. Okay? So this first question on the macro. How does the environment feel for you from your lens out there, June 2023 versus maybe January 2023? And then the year ago period, June 2022? Does it feel different? Does it feel the same? Or is the end market looking at observability tools differently? Just any sort of help of what it feels like out there.

Well, if I play that through chronologically, I would say that January [ 20 ] [Audio Gap] basically, we're still in it. It's really started in earnest for it to become slower, I would say, in the June time frame of 2022. And we still are seeing this in June of '23. The expectation that we have relative to our FY '24 plan is no notable improvement. So we're assuming status quo, parity in the macro environment through our fiscal year.

Got it. Got it. And moving on to the topic of AI, very topical here, been asking every management team, AI, generative AI in 3 different flavors. So the first question is, how does Dynatrace think about leveraging AI within your offerings?

Well, Dynatrace, if I play back to the answer to my first question, Dynatrace uses AI to really provide the precise answers and automation that we use in our platform. So unlike most, we're not new to AI. We've been using it really as part of the core platform of what we deliver and have been doing so for a very long time. Now to clarify the next double-click level or level down, we use what we refer to as or think of as causal AI. This is different from generative AI, which is ChatGPT, large language models, et cetera. The way to think about generative AI is that it is an enhancement to productivity, whether it's writing term papers or film scripts or code, for example. So I can use generative AI to create more code faster, deliver more workloads, more applications, more capabilities, more elements to be observed, if you will. So we do view generative AI as a tailwind to the opportunity generally in observability for those reasons. Having said that, causal AI, what we do is using data that is constructed in real time. These are traces, routes, logs, metrics, a number of data types generated by IT ecosystems that we are constantly creating and evaluating using our AI engine to deliver very precise answers on how your environment is operating. The combination of the 2, generative AI and causal AI, we believe, are very synergistic, highly symbiotic because if you can use generative AI and natural language to query causal AI, then you can actually use generative AI to, through our AI capabilities, deliver very precise answers.

That makes a lot of sense. Now thank you for that.

Really? That's good.

Yes, yes, yes. And I wanted to follow up of the end market that you're addressing and how they are thinking about what AI and generative AI means for their businesses and the pain points that they're trying to address with observability tools. When we look back at January 2023, I don't think -- it would be hard to imagine the kind of awareness that generative AI has today. So presumably the end market, your customers are starting to look at or maybe not, the way that they're structuring their tech stacks or data stacks or observability solutions. So has the conversations changed in the way that your end market is looking at observability?

I was in Canada with -- in customer meetings, round tables last week with many of the largest banks in Canada, retail, train companies, a variety of them. So I got a very real-time update from maybe 20 customers in the last week on their perspective on generative AI because it came up in basically every conversation I was in. And what I would say is the common thread is strong feeling that it will be a significant productivity enhancement; uncertainty as the time line over which that might happen. And so everybody is looking at it, everybody is evaluating it, how to find ways to accelerate productivity. So I said at the outset, generative AI, it's really about productivity gain. And yet, it's accessing a data store that is largely a static data store that the AI results are only as good as that data store. So I think that organizations are really working hard to figure this out in real time. Nobody wants to fall behind because your competitors have found a way to get more productive faster than you. But at the same time, there is reservation to sort through exactly how it is best utilizable by your company.

I wanted to dig into that a little bit, and you're saying about customers having a little bit of reservation. Can you dig into what is the reservation? Is it data security privacy? Or anything that they were talking about that is maybe causing a little bit of hesitation for the end customers' perspective?

Well, one element is that generative AI is never going to deliver a perfect result. You're still going to have to provide some sort of manual work on top of that to work through the result. You can't just ask it to develop a code snippet and be sure that, that code snippet is going to work perfectly. It just isn't that simple. [Audio Gap] And so I think that organizations want to make sure that they don't get ahead of themselves and whether it's for security reasons and integrating more viruses due to code libraries that might be called that may not be compliant, for example, or for what other sets of reasons, I think they want to proceed with some degree of sort of explicit view as to what they're getting themselves into and how to use or utilize it most effectively given that it's so early.

Okay, okay. Last question on generative AI or AI is...

I don't believe you.

Is -- understood, understood. Is the monetization of AI for Dynatrace, how do you guys think about it? Is it just embedded in the platform, you get it when you purchase the contract? Does it drive premium SKUs? Could it even drive new products for you over time?

So a little bit of all of the above. As I mentioned at the outset, what I would say is that generative AI enables more workloads, more applications to be created faster. This notion that data is exploding and that it's harder to manually evaluate through those network operation centers where you have an army of people staring at a sea of screens in glass looking at dashboards, they're looking for alerts. Those alerts tell you red, yellow, green. And that is, you see red, and then you begin a triage process. And that is a difficult process. Figure out precisely where something is not working the way that you want. And this problem is exacerbated by generative AI. If you can now take a situation which data is exploding, resources are constrained, I already can't get to the answer quickly enough. And now I'm going to make it worse by adding or accelerating the number of applications? Problematic. So I'd say the biggest way in which generative AI helps Dynatrace, at least in the immediate term, is a catalyst that comes out of what we see anyway, which is data explosion in these other elements. By making that problem worse, it's going to drive automation being at the forefront of objectives of these organizations. And as automation moves in the forefront, that suggests Dynatrace because that's what we do better, we believe, than others in our industry.

Got it, got it.

There's another aspect, Koji, which I would just say, which we also do believe that this notion of connecting generative AI to causal AI really is valuable, and that's something that we will drive as well. That should also be a catalyst to business opportunity.

Got it. Okay. I want to move the conversation over to Grail.

Okay.

Yes. It's a big topic of conversation. Grail. We've definitely heard a lot of positive reviews as we're talking with partners and customers out there. And the one thing that we hear a lot about is this thing is fast. And we just don't know why it's so fast. Maybe could you help us explain or help explain why, from an architectural standpoint, Grail is just so good at what it does.

Well, so to step back a level for those of you who don't know what Grail is, Grail is a massively parallel processing data lake house that we've constructed. We began work on it about 4 years, 4, 4.5 years ago. And we designed it because we couldn't find anything in the industry that we -- that did what we need it to do, that we could leverage and integrate into our platform. So we built it ourselves. And it uses hypergraphing technology which keeps all data types in massive scale in contact with one another. So we believe strongly that ultimately, the right answer in observability is end-to-end observability. That means applications, infrastructure, application security, end user monitoring management, all of these elements will be holistic in providing a single integrated unified platform. And that's what we can deliver of Dynatrace. Well, if you deliver a fully unified platform, that platform needs access to a fully integrated set of data types, meaning logs, traces, routes, metrics, real user data, behavioral analytics, metadata, et cetera. That data gets stored in context in Grail. And Grail was built with an architecture which it doesn't ever do reindex. So the result of it is, it is extraordinarily fast. In fact, for complex queries, what we see is 5 to 100x faster than the capabilities in the market today to do analytics and queries on that data. So that's how we did it.

How difficult is that to replicate when you talk about a massive data lake house and hypergraphing technology and massive scale? Is it just -- if I were to go out there, raise some money and try and build this myself, I mean how long would that take for me to replicate?

Well, I'm pretty sure, Koji, that you can build it yourself, and I'm not sure about anybody else.

Thank you. Thank you.

The answer to your question as well. It took us 4 years to do it, and we started with a very [ constant call ] in mind. The -- there are 2 challenges. One is just building the capability from scratch, which is time consuming. The second is the innovative [ dilemma ] problem of how do you figure out how to shift data repositories from one to another. And we sort of built Grail with that in mind. So I think it's a -- it is a substantial undertaking to replicate.

Okay, okay. What are you most excited about from this Grail platform for the next -- over the next 6 months and then maybe over the next 5 years?

Well, the answer is probably similar in both cases, which is I get really excited about what Grail can deliver by way of delivering an observability unified platform, an end-to-end observability platform, covers all data types and covers all of the use cases that I described earlier, be it AppSec, be it log management, be it end-to-end observability and tool consolidation. These are all use cases that our customers come to us and say, I want that. They're not necessarily coming to us saying, I want and need Grail. They're asking us for us, I want a unified observability framework. Too many tools, too much dispersion in access to those tools, and it's too complicated to manage. Moreover, I need scale, performance, et cetera. So this is what we're focused on delivering, an answer to those requests for purchase behavior and use cases. Grail is an enabler to that.

Got it. I want to ask you one more question and then open it up to the audience to see if there's any Q&A out there. There is a microphone. Please wait for the microphone to make sure that we get your question for the webcast. But my last question for you before Q&A is I think about Grail, I think about large data ingestion, I think about the fast speeds that it has, but then I also think about margins. And we've talked about this before, but I wanted to dig into it a little bit more. Just if you're ingesting all this data, it feels like it could be a potential impact to gross margins as you're ingesting all that data. Walk me through what Grail means for gross margins as it continues to scale.

The net of it is we don't expect it to have any material impact one way or the other. To be on Grail, you need to be on our SaaS platform. If you're already there, we expect the margins are already reflected. If you're not, you're in a managed environment, you move to a SaaS environment, there's going to be a modest uptick in pricing from managed to SaaS, which should make up for any cost differential. So net of it is we're not projecting any material gross margin impact from Grail.

Got it. Thanks, Rick. Any questions from the audience, please raise your hand, and we'll get a microphone over to you. We got one question over here, please.

What's the state of competition in the industry right now? I guess, for a long time, there's thought to be a relatively crowded space, too many players. So I guess is there a feeling that there's going to be M&A consolidation in your industry in that subsector?

So I'm not necessarily projecting any major consolidation of major players. I do think we're already seeing separation of organizations that are competitors in our space. And I think that our view of that is that the end-to-end observability and unification of an approach is the way that we have been competing, the way we will compete, and I think that's differentiable as we look at.

As we think about the competitive landscape in the observability category, it's easy to sometimes bucket certain companies and competitors within certain aspects of observability. Dynatrace, very, very good at APM from the onset. Other competitors may be log management, infrastructure monitoring, et cetera. But all these competitors, when you go to the observability space, when you look at their websites, it feels like they all have platforms. So walk me through, when you're talking with your target end market, the top 15,000 out there, what are they looking for? How are they thinking about consolidation? Is the platform value proposition becoming more meaningful for them?

So the -- in my view, the platform proposition is absolutely becoming more meaningful. I'm talking to customer after customer where they say, I've got open source here, internally designed observability monitoring [ and ] monitoring here. I'm using one particular tool here, another one here, and it is unmanageable. I mean in a network operation center, you can't have 15 different tools, which is what we see sometimes. BT, a great example, British Telecom. They onboarded within the last year or so, tens of thousands of host units. They had 16 tools. They -- for observability, they told us. They wanted to drastically narrow that down. As a result of doing that to Dynatrace to get a much more comprehensive platform, they put out metrics to us that said they reduced incidents by 50%, 5-0. They reduced MTTR, mean time to repair respond, by 90%, and they projected saving GBP 28 million over a 3-year span. These are the kinds of metrics that our customers want. They want meaningful observability capabilities that are going to make material inroads into the metrics that they're concerned about, like number of incidents, MTTR and the like. So they want to pull this down. The other fact I'd share with you all is that last quarter, we had more than 20 competitive takeout deals of $1 million TCV, total contract value, or greater that occurred in that quarter. And these are competitive takeout deals from others in our industry. So we asked our sales team, why did this happen? What was the driver? And typically, the answer to it was automation, automation, automation. This notion of manual processing of red, yellow, green alerts and dashboards replaced by an automated capability that can be delivered by Dynatrace, using the combination of our end-to-end platform and our AIOps engine really does have appreciable and differentiable impact.

Is that the first thing that customers see when they adopt the Dynatrace platform coming from, call it, a basket of best-of-breed point solutions, it's that immediate meantime to fixing a problem, or what are the ROI that they're looking at? Is it -- what is it? Just you're seeing having higher visibility first? Or incident response is better? I mean what do the customers really feel first as they go to the platform versus a basket of point solutions?

What they want is they want near-instant response of a better ability to get to the heart of any issues. Moreover, what they want is more proactive analytics looking ahead so that they don't have problems in the first place. And our vision is, as an indicator have indicated, to make the world software work perfectly for our enterprise customers. That means a good break in the first place. It wasn't about rapid break fix. It was about avoidance of a problem. And using the analytical framework within the Dynatrace platform, we believe that we can drive that set of capabilities that you're getting more predictive capabilities to avoid the problem in the first place. So that's what they're after. But the short answer to your question is they want reduction of incidents, more rapid ability to isolate and identify issues and then resolve them much more rapidly when they occur.

Got it. I wanted to ask you one quick question on cloud optimizations. Very topical for a long time, still topical today. Have you been seeing any change in pace in cloud optimizations out there? And maybe why or why not would Dynatrace be more shielded than others out there?

Well, we've been saying for 6 to 9 months now that while we haven't looked at cloud optimization necessarily as a huge tailwind, so I don't want to overstate it, we would still like the cloud to [Audio Gap] precisely what we do. We enable much more efficiency out of your cloud deployments. We want your cloud deployments to work better. We even call -- our tagline, we even talk about our tagline is, where our mantra is cloud done right because we offer more sophisticated and automated data-driven analytics to enable that to happen. So I wouldn't say we're immune from cloud optimization, but what I would say is that it is an objective which we aspire to deliver against.

Got it. I can't believe I forgot to ask you the question. I was going to ask this right at the beginning. You had an announcement, new Chief Revenue Officer.

Yes.

Yes, yes, yes. I guess the big picture question, why now? And what does the new Chief Revenue Officer maybe bring to the table?

Sure. Well, Steve Pace has been our Chief Revenue Officer for the last 7.5 years. He's done an exceptional job growing, I think, a great company at Dynatrace at the rate of growth that we've delivered. That enabled us to surpass the $1-billion-and-then-some ARR mark. So huge kudos to Steve. He's at an age when he wants to retire. And this is a well-planned transition, highly thought through. So he and I have been working on this now for some time. The time has come when this is a good time to complete the transition. So we opened a search a little bit ago. And we sought to add a CRO with certain capabilities, one of them being scale, global capabilities, partner capabilities, good cultural fit, a great leader and proven leadership talent. And we found all of those in Dan Zugelder, comes from VMware. So super excited to have Dan joining in early July. I can't wait to have him be part of the team. Steve will stay on board through a transition period through our fiscal Q2, through the end of September, early October, to assist Dan in getting up to speed. So we'll make this a seamless transition as we possibly can.

Got it. Now thank you for that. Before I ask a couple more questions, I just wanted to open it up to the audience. Any questions from the audience before we -- yes, we got one up here. Please wait for the microphone.

Do you guys have any plans of introducing Copilot-like solutions to interact with the data that you already have, leveraging that generative AI capabilities? Or is that something that just doesn't make sense in the industry for whatever reason?

The capabilities -- well, we would rely upon partners or third parties mainly to do that, I think, at this juncture. We would benefit from more code capabilities coming out of those types of initiatives, but probably we wouldn't be generating those [indiscernible].

Rick, I wanted to ask you a question on security. Security, you recently talked about kind of a milestone goal of $100 million in ARR by fiscal '25. So are you seeing any certain patterns or trends within the customers that are adopting security today? How are you thinking about go-to-market for security over the next, call it, 12 to 24 months? And what gets you most excited about security?

Well, security is an area where I believe that application security or certain aspects of application security and observability are converging at a rapid rate. We are not focused strategically on being all things to all people in security. We will leave that to the mega security vendors and others. What we are focused on are areas of application security where observability data adds differentiable value and [indiscernible] management being the first one, but areas like runtime application protection, maybe, over time, even SIM capabilities, we believe that bringing not just logs, but logs traces metrics, all the data types together can, in many ways, deliver a better SIM than SIMs in the market today. So these are elements in which we'll be investing going forward. I've -- the good news is -- the good news is I've had lots of personal experience in application security from my background in growing more than $1 billion business in security. So I've seen the movie before. That doesn't mean this movie plays out the same way. So we'll see how it adjusts. But I do believe that it is possible. And I think that possible in the sense of growing a notable business, I can't say to what level. But at least at the moment, that $100 million plan, over a 3-year span, is on track. We're 1 year in, 2 more years to go against that plan. And we feel good about it. We are, today, about 10% penetrated at AppSec into our customer base. No reason that can't continue to grow to 20%, 30%, 40% as we look at.

Awesome. Rick, we're all out of time. Thank you so much for doing this. This has been a great conversation.

All right. Great.

Appreciate it. Thank you so much.

Thank you. Thanks all.