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

Good morning, everybody. I'm Fatima Boolani. I jointly run the software research team here. And I am so thrilled to be sharing the stage with Jim Benson, CFO of Dynatrace and we're going to get into the meat and potatoes of a lot of good discussion areas. Thank you for being here.

No. Thank you for having me.

Excellent. Yes, I think a good place to start is, I think there's a lot of familiarity with the Dynatrace story and kind of the key markets that you're playing in, specifically observability. But I think a good place to start for all of us to just level set is just an overview of the Dynatrace story and where you are as a company in terms of your market and product presence and then we can kind of dig into the financial model and pricing model. But just start at 30,000 first?

Sure. All right. We start at the highest level. So for those that don't know the Dynatrace story that we are in the observability space. So sometimes when people say the word observability, they don't know exactly what that means. So think of it as software that can give you insight into how your infrastructure is behaving, monitoring. Think of it as monitoring your IT environment, whether it be applications, infrastructure, network, what have you. So that's the observability space. And it's $50 billion TAM based on the different people's views of it, but it is soft of a very, very large TAM. And the Dynatrace story started back in [ Austria ], probably what, 13 years ago, where it started in the application performance monitoring space. So interestingly enough, there's infrastructure monitoring where you're monitoring someone's network infrastructure and the infrastructure that applications sit in. And then you got application performance, which is a significant level of incremental complexity, Dynatrace has actually started with solving the hard problem. And then pivoted -- interestingly enough, going back to your point about the offerings, and then pivoted into other areas around infrastructure monitoring, digital experience management, which is giving you some insight into your environment so that you can get a better value proposition for customers. And then more recently, we've extended into application security, which for us is vulnerability analytics and protection run time analytics and then most recently, logs -- log management and analytics. So the portfolio continues to broaden. What makes Dynatrace unique in the space is that we have a unified platform at a data layer, not at a kind of user interface layer, but actually literally at the data layer. So our ability to bring data into context, whether it be traces, routes, logs, real user measurements, we're able to do that in real time and we focus on the Global 15,000 customers. We focus on the enterprise customer base. We are not focused on SMB. And these customers have huge complex environments, and these environments have data explosion. And the benefit of Dynatrace is you get analytics and you get AI, which has been built into the platform for a while now and automation. And so others that are in this space had more of a dashboard orientation and maybe sophisticated dashboards. Ours is much more of a analytics, AI and automation. And so that's maybe at 100,000 feet.

Jim, you mentioned this. You've been doing this for the better part of 13 years, right? So clearly, the problem and the concept of application monitoring is not new, right? But it's certainly taken on new significance, new importance, new criticality and I think complexity has to do a little some something with that, right? So can you tell us why Dynatrace has kind of catapulted to the front of the line. Again, this has not really been a new technological concepts. Applications have existed, right? You have had some permutations of competitors in this space or the better part of the last decade. So what has empowered you, enabled you to speed gate past everybody else and be a dominant force in providing clarity in complex infrastructure and application environments?

It's a great question. I think that from the early days of the company, workloads were largely on-prem. Actually, workloads today are still largely on-prem. And so they saw the problem around how is your application performing in a, kind of on-prem environment. Well, if you think about on-prem environments, back then, workloads and changes to software, maybe you had a change monthly. Maybe you had a change, maybe weekly. And so the frequency of software updates was not as significant. Now what's happened as the world has unfolded and you have more and more workloads in on-prem environments. But there are an on-prem environments that ultimately your user is, maybe on a handheld device somewhere. So the complexity that you now have to traverse to identify where problems are has been -- has exploded. Data has exploded. Customers -- now, there's a lot more workloads that are in cloud environments. Now you have hybrid environments. And as you introduce all of this complexity, you've introduced an opportunity for there to be some defect. And you need to be able to detect where the problems are and what has made Dynatrace unique and maybe evolved is as this complexity has unfolded, this platform that we built out was able to expand from an on-prem kind of environment to now solving on-prem and cloud workloads and hybrid workloads and doing it in a way that because we focus on the enterprise customer base, most of these customers are hybrid customers. And so the complexity that they have is the ability to find out where problems are, is heightened. And you can arise on glass chasing alerts, because there's just too much data to do that. So you do need to build a level of proficiency in having analytics in AI and automation drive your solution. The analogy that I would give would be a needle in a haystack. Ultimately, to fix a problem in software or an infrastructure, you need to know precisely where the problem is. We can tell me where the needle is. Others can tell you where the haystack is. But ultimately, view to action something, you need to know precisely where the problem is. And I'd say this has just been years of innovation from our team in Austria largely that has built this out and kind of expanded on it.

Jim, one of the questions I get a lot is you've delivered very steady performance in your franchise over the last couple of years. There hasn't been as much of an elasticity in either direction in a dramatic way as it relates to the bigger picture around cloud adoption, right? And I think there are some of your peers, any competitors in the space, who maybe have more proximity or exposure to some of these cloud native, VC-backed, small but fast-growing customers. I was hoping you could take some time to kind of discuss just from an end market perspective, what your customer base looks like. And the reason I ask is your opening comments were just around, "Hey, you were helping solve really thorny, hairy problems on very complicated applications", which naturally typically tend to reside for some of your larger and maybe [indiscernible] enterprise companies. So I was hoping you could spend a little bit of time on kind of delineating some of those perceptions but potentially misperceptions as well?

Yes. So I'll bring us back to. So we focus on enterprise customer base. So we talk about Global 15,000. So think of this as your largest 15,000 companies and/or governments on the planet. And the reason we focus on them is the environments that they are in, by definition, are complicated. By the way, those Global 15,000, they're not all legacy, "Hey, I have on-prem. I have cloud workloads." In some cases, those are cloud-first oriented customers. So the benefit of Dynatrace is we can do both. We can manage and provide visibility to on-prem workloads. We can provide visibility to cloud native workloads. Your point is not lost on, if you're a cloud native company, are we, maybe not going after them. I'd say we do, in our hunting model, we are identifying those particular customers that have significant opportunity for a total addressable market for us. We will go after them, but we are focused on the customer base that we are because we saw very significant problems. And think of large financial services, organizations, governments, retail. So we are not -- we are evasive in the verticals. I wouldn't say necessarily that maybe there is a misperception that "Well, you go to Dynatrace if you want a legacy workload monitored. You go to someone else if you want a cloud native workload monitored." That's not the case, we can do both. We haven't abandoned that. Having said that, because of what we do, I would say that the -- depending upon who you talk to, you listen to Andy Jassy, he will say 80% to 90% of workloads still out of on-prem. So we're solving what we believe are significant problems, and we can help customers as they move on their journey.

That 15,000, G15K, so Global 15,000 addressable audience, right? You have 4,000 customers today. So that's a pretty significant mathematical runway to kind of go after, right? But your growth on the installed base side has been more measured, right? And I think you've alluded to that in the context of, "Hey, we're solving for quality of logos versus quantity of logos." I wanted to spend a little bit of time on how you're kind of straddling this quality versus quantity, especially as you've also demonstrated that you're doing larger land transactions and if you can -- I'd love for you to put some numbers around that. So what that strategy around installed base and logo growth potential is given 15K is a pretty big number versus your 4K today?

It's a great question. So you're right. We have a little over 4,000 customers. We're targeting 15,000. So this is a significant runway for new logo expansion. No doubt. And just to be clear, you're right, I focus on the quality of new logos. It doesn't mean I completely dismissed the number. Because, to your point, it is important that we penetrate what is a significant addressable market. However, what I want to make sure is that we're measured in that we want to land with the right profile of land size. We have found that if a customer lands at $100,000 or greater, their propensity to expand is significantly greater. Now they get better support and coverage from us when that's the case. And so that hence why we focus on the quality of the logo. But I'd say relative to our growth, probably 40% of our net new ARR is going to come from new logos and 60% will come from expansions as you continue to grow. As you can imagine, your expansion percentage will increase. But we have a motion depending upon the geography that has some dedicated hunting or what we call a strategic acquisition orientation. So people that are just focused on selling to new logos. And then there are environments we have where there are hybrid reps where they're selling new logos, and they're also farming in the installed base. So there's significant runway to expand further with our existing customers and a significant runway to acquire new logos, and that's kind of the general equation. And on the go-to-market side, we go to market, both direct and through partners. And so our ambition on the partner side is to be able to get broader reach from partners that the reality today is that partners are in over 2/3 of our deals. That doesn't mean they're driving 2/3 of the demand. We need to get to a model. They actually drive, call it, less than 1/3 of like literally originated, that they originated the deal. So getting back to your kind of new logo and what the equation is, some of the way you get there is broader reach through partners to be able to reach customers that -- because what we don't want is we don't want to monitor just add new reps, productivity per rep. That model breaks at some point. You need to have an ecosystem that you're leveraging as well.

I think one of the other ways that you've expanded your aperture on driving more adoption, improving velocity of customer acquisition and even expansion is the introduction of the Dynatrace platform subscription. And I think it kind of came into our consciousness early last year. So it's been incubating. So I'd love for you to spend a little bit of time talking about DPS. And then relatedly, what are the milestones and mile markers you've hit? And as the adoption grows, what are some of the things that you're anticipating from a consumption standpoint and KPI, operational metric standpoint?

So you're right. Just to level set for folks that don't exactly know what DPS is, it's the Dynatrace Platform Subscription. And the pain point that it was intending to solve was customers gave us feedback that they love the product. They did not like the contracting process. The contracting process required continued sales effort than to try new products. So Dynatrace Platform Subscription gives you access to the entire platform for a fixed fee or for a finite term, 1 year, 3 year.

Whereas historically, it was you subscribed to one particular product portfolio.

Yes. It was a SKU-based model. So you bought a finite amount of software for application performance monitoring or infrastructure monitoring or application security. So customers were buying product attributes. They were not buying access to the platform. So the point of the Dynatrace platform subscription is you have access to the full platform of the rate card, mid to a dollar amount, whether it be 1 year or 3 years. And based on that, you can leverage any product that you want on the platform. So it's a frictionless model, very similar to -- not exactly -- but very similar to the hyperscalers. That's the way hyperscalers work with. You commit to a committed contract value and then you can access the entire platform. So we announced it generally available last Q1. We did it on a limited available basis manually for some customers and it's ramped. It is -- so you mentioned milestones. So we're kind of a year through it, a little over a year. 20% of our customers are now on the Dynatrace Platform Subscription model and over 40% of our ARR. So they've already made very meaningful traction. Now, as you can imagine, that has spiked. The biggest issue that we have is getting our sales force comfortable selling it because it was different from selling a SKU-based model. So in the thesis is that customers that are on a Dynatrace Platform Subscription will consume faster. And if you consume faster, it means you'll burn through your commitment earlier, burn through your commitment earlier means you'll be able to see an earlier expansion. So your NOR should benefit from that. So the early -- we got early cohorts. Obviously, the only real cohort you have that's kind of burned through this is your Q1 cohort of last year. When we look at all these -- so the customers that are on the Dynatrace platform subscription model consume at a 2x rate to customers that are on the SKU-based model. So they're consuming a lot faster. It hasn't yet manifested itself in all of these early renewals because we're still early days in it. But that's the thesis, and I'd say we've made very good traction to date.

One of the questions I get a lot is introducing DPS to customers just from a financial model standpoint. Doesn't that actually introduce maybe a lot more volatility in the numbers and that, "Hey, you used to be a subscription model. Now, it's commit based." So you've got customers that are maybe burning through their contract in 6 months, right, and so you're having -- that's a great conversation to have versus 1 year, right? So anything you can share in terms of behavioral patterns because I think the other juxtaposition is a lot of CFOs and you're in the seat, right? You're more inclined to under-commit. So there's also a likelihood that you're going to blow through the cap sooner, right? So that's the good side of that. So how are you managing some of those variables? And what are you actually observing in how customers are behaving? Are they habitually and dramatically under committing? Or are they closer to the pin? And then ultimately, as CFO, how are you incorporating some of these volatilities in the way you're thinking about ARR trajectory, ARR linearity?

Yes. Just to make sure that before I answer that specifically, just to make sure we level set that the model still is a ratable subscription model. So your customers committing to a dollar amount in a term. We recognize revenue ratably over that term. You are right that there's -- it's a consumption drawdown. So customers could burn through their commitment earlier. If they burn through their commitment earlier, using an example of a customer commits to a 12-month contract, $1.2 million, you generate $100,000 a month in revenue. It burns through the commitment in month 10 and you sell $200,000 worth of revenue, you haven't recognized. They did an early renewal and maybe they extended it for another 12 months. You then recognize that $200,000 over that extended period with an expansion. So it's still a pretty -- it's not a model that's going to be lumpy. It is a model that is relatively consistent. So it isn't like a consumption model where you can see volatility from period to period. Having said that, one of the things we are looking at, Fatima, is one of the things I think we've begun to introduce to investors is that consumption will become a more important attribute for us to be talking about. And when I say consumption, I mean P times Q, price times quantity, think of it that way, that -- because ultimately, that is what is going to lead to an earlier expansion. So early days with DPS is customers are consuming at a greater rate and pace. Now, I'm also intellectually honest to say that many of those customers probably would have expanded or consumed more anyway just because customers that already like you tended to use that vehicle. But I'd say going forward, I think the good news is the data points around -- we have over 70% of our new logos now land with DPS. Now 40% of our ARR. Now, we're in the pace of -- and just so you know how it works, that you land and then we have a CSM, customer success team model. That drives the adoption. We've telemetry built into the platform that customers can see. So the other thing is that this is not just what's good for Dynatrace, it's that what's good for the customer that we provide visibility to how they're consuming. So they can know I'm actually burning at a rapid clip current [ closing ] speed. I'm going to burn through my commitment earlier. Now, we don't want -- if a customer is getting value, you hope that he would renew it earlier. Could they dial something back because they don't want to get into maybe an overage situation? Yes. But we're monitoring this. And I'd say early days, we feel really good about it that it has solved the impediment, which was getting an ability to access more of the platform. And now it's about customer value. And if they're getting more value, then they'll do an earlier expansion.

Is the overage piece becoming more prominent or is that just sort of a tail dynamic?

It's immaterial. Having said that, that I was telling an investor earlier that -- and we don't call it an overage, we call it an on-demand consumption. The reason we do is that overage implies that you're charging a premium. We do not charge a premium if you burn through your commitments. So your rate card doesn't change. However, you're incented to do an expansion because if you do an expansion and you increment in your commitment, you'll get a better unit price. So they have an economic benefit for them to consume earlier. But you will see it modestly show up in revenue, which is most of our revenue is subscription revenue ratable. You'll have a very immaterial amount of on-demand consumption for customers that maybe are in the, "Hey, I'm 2 months from my renewal. We'll just say on-demand consumption until I renew." So -- but it's immaterial, but there is something.

The other comment you mentioned was, "Hey, consumption is tracking very nicely ahead", almost 2x what you're monetizing?

So the 2x comment is on customers that are on the DPS contract versus a SKU-based contract.

Got it.

Consumption of the platform is growing significantly faster. Like all-in DPS customers, SKU-based customers look at it in aggregate, is growing significantly faster than ARR growth. Again, what it tells you is that if you're growing consumption faster than ARR growth, at some point...

It should catch up.

Exactly. As you can imagine, you have -- do the math of it, depends upon where you are in your life cycle. If you're -- if you just early renew and you're consuming really fast, you might not have a renewal if you're later in your contract cycle. But the whole premise is the longer you -- we're trying to drive consumption growth and hopefully, customers are getting value from it. Consumption growth should lead to ARR growth.

And what is that. So it's a very natural outcome that ARR and net retention rate is going to have to mechanically catch up to the amount that is being consumed by the installed base. And so what are some of the evidence points you're seeing? And is there a kind of a time frame? Is this a 6-month lag? Is this a 12-month lag? And I can appreciate only 20% of the base is on DPS, right? So how should we internalize the downstream financial impact and when that should actually provide upward -- positive upward pressure on net retention rates in ARR?

So we've -- at current course and speed, I would say, if what we've seen continues, you should see stabilization in NRR, maybe in the very back half of the year. And then if it continues, you should potentially see modest NRR accretion for fiscal '26. But a lot of variables that -- I said at current course and speed, if...

Cruising altitude.

What is that?

Current cruising altitude.

Exactly. The current cruising altitude, if that continues, and you continue to have consumption grow at the rate that it is above the level of ARR growth, it should lead to a bottoming of NRR and then an acceleration of NRR. The timing of what you say, is it 6 months? Is it 9 months? It's tough to call, to be honest.

I think DPS, clearly, the shift towards this very attractive contracting vehicle, has been a pretty significant change in the way the sales organization has gone to market, but there's also been some other more express changes to how you are going to market, right? You've had a new CRO come into the role. Can you walk us through some of the intrinsic sales organizational changes that you've instituted over the course of this year and the end of last year? And how do you expect some of those outcomes to play out?

It's a great question. So we talked a little bit about it, I think, in our fourth quarter call. So I wouldn't say he's new anymore. I guess he's relatively new. He's been here a year. But our Chief Revenue Officer, one of the things he looked at in his first 9 months was he looked at the coverage model. And obviously, we focus on the Global 15,000. That hasn't changed. He kind of checked the box. He agreed with that kind of as a your target market based on what we provide. But he looked at where resources were deployed. And he said, "If I literally look at the Global 15,000, and then I say, let me look at the TAM of the Global 15,000." He said, "If you look at the TAM, the top 500 customers represent 50-plus percent of the TAM." So the 50% of the TAM is in the Global 500. And he said, but when I look at my resources, I don't have the right resource attention there. And that's the biggest sweet spot for Dynatrace. That's the -- those are the problems that we solve best. So we had on average in the Global 500, 8 to 10 accounts per rep. So one of the changes that we made was we reweighted and we took the resource allocation from 8 to 10 accounts per rep to 4 to 5. So we cut it in half. And the premise is when you're selling to those large enterprises, you have -- in order for you to go deep into them, you need to understand them. You can't do that with that many accounts. So we now -- we did it effective April this year, we reassigned accounts. So we added more resources in the Global 500 from a capacity perspective. So that's one change. Good news is we've, first 90 days, we were a little bit worried about whenever you make sales force changes, there might be a little bit of disruption. We did not see any. That was one change. The other change was that we had a very established play of land and expand with go find an application owner, get to a POC, proof of concept, land with the customer and then expand workloads from there. Tried and true model. We now have a presales place. So that still is one. There is now an end-to-end absorbability sales play, which is customers that are looking for some level of tool consolidation or vendor consolidation. So that's a specific sales play. And then there's a third sales player on cloud modernization and you hear it in terms of shift left, shift right. But it's basically the developer in the IT operations' communities working more together around being responsible for workloads and ensuring workloads work correctly. So you have the -- we've changed some customer segmentation, more weighted to higher propensity to spend customers. We have introduced more sophisticated sales plays. We have adjusted on the partner side. So partners are critical because we can't just add reps. We also need expansion from the partner community. And even though over 2/3 of our business is influenced by a partner, the reality is we want more deals originated from a partner. We want more pipeline. And so there's been work done on the partner enablement side where we're segmenting how you work with GSIs versus hyperscalers versus regional channel partners. So there's -- and then there's a whole sales enablement that goes around that. So I would say we've significantly enhanced the sophistication and enablement on the go-to-market side. So sometimes, I think people think that, "Oh, they're just whale hunting now. They've changed all their resources." That's -- there's a bunch of things we've done. We've not ceded below the Global 500 to other competitors. We're still focused on that. We just needed a little bit better resource weighting and then we needed a little bit more sophisticated sales model to go after that.

I think the other piece on the chessboard as it relates to go-to-market has been you implementing the biannual quota system. Talk to us a little bit about that, the impetus there. I think maybe some of it is obvious, but what was the motivating force to kind of do the biannual quota. And I think you've rightly been conservative in the way you're thinking about guidance. But why wouldn't we actually see almost immediate productivity improvements because you either make quota in 6 months or you don't?

Yes. No, it's a great question. I personally have not seen in my career two 6-month quotas. Our COO has. So he's done this before. So as you can imagine, the obvious benefit from two 6-month quotas is linearity is something sales cares about them. When you don't have -- when you have the annual plans, not surprisingly, they have a year to hit their number. So it shouldn't be a big surprise that what is Dynatrace's linearity. Well, historically, our linearity is 60% of our net new ARR comes in the second half of the year, and it used to be more in Q3. Now it's actually more in Q4. And so two 6-month plans changes now the psyche of the sales organization where they now need to hit a first half number to second half number. Now for them, they can make more money. So the premise of it is the -- we've actually did the math of it, and it actually is a little bit more costly as a result of doing that. But the premise is you will get, hopefully, some improvement in linearity to your point, like, that it's obvious. Whether or not that proves to be true, we'll see. I mean part of it is with 6 to 9-month sales cycles you may not necessarily see a huge impact in the first half. But the other benefit you can get from two 6-month plans is the -- when you have annual plans and if you want to make even small adjustments midyear, you can't, because it's wildly disruptive. You can make midyear adjustments, not that we're planning on making any, but you can if you wanted to. And so maybe you get an improvement in linearity and maybe you get an ability to make some modest sales changes.

And none of that is really factored into the way you're thinking about ARR?

Well, I mean, as you can imagine, let's be clear, that our guide is prudent. The guide suggests that the performance in Q1, growth moderates between now and the end of the year. as you can imagine, our internal plan is more ambitious than that. But we tried to build some what we think is transparency for investors around. We don't know where these sales changes, even though they haven't had any impact in the first quarter, the benefit of them has not materialized yet. So we're probably building a level of caution into that. And then the macro environment is still uneven, not better or worse. And then what was an emerging trend in, maybe our third and fourth quarter of last year of customers considering some level of tool consolidation, we're actually seeing that become -- which is why it's a sales play -- we're seeing more interest in that. And by definition, those sales cycles are a little bit longer. So we try to factor that in to try not to get ahead of ourselves.

So you read up macro. So I'm going to ask you the obligatory macro question. You do have a sizable international business. So hate it or love it, you are exposed more to what's happening in EMEA and APAC. So I was hoping you could spend a little bit of time on what some of the behavioral patterns, technology adoption patterns are that are maybe distinct from geography to geography. And I know maybe EMEA gets a little bit of a bad rap that, "Hey, they were kind of behind the curve on cloud adoption. So they may be a little bit slower on the top curve." And I don't think you've experienced that. So I'd love to get your perspective on it. But just the business performance from an international and regional theater perspective? And any observations on if customer maturity or customer procurement behavior is markedly different?

It's a good question. I would say that -- so we do roughly 40% of our business outside the Americas. And I'd say from a technology adoption perspective, you're right that they probably do vary a little bit. You probably have EMEA that is probably less cloud native in their cloud native journey. I think you have Asia-Pac, maybe a bit more on the accelerated path on the Cloud native side. And then you're seeing it's continuing to grow in the Americas, that you see some notable distinctions by geography. I do think in EMEA that there's more macro behavior like the caution on the macro side is more evident in EMEA. So I'd say customer behavior is more cautious. Now the contrast to that is that's actually been our best performing geography. So it is ironic that I do think just from a customer behavior perspective, they are more price sensitive, and it takes longer to get deals done. But I think that we've executed well. Our sales team has executed very well in EMEA.

And by the way, has that manifested more on a slower net expansion. So if you were to decompose net retention rates from a geographic perspective and layering on some of the comments, are you seeing that EMEA behavior manifest slower and slower logo growth or slower expansion?

I would say the more on the expansion side than anything. And then the other geographies that I would comment on, I think we have a huge runway on Asia Pacific and Japan that -- we have a new sales leader there. And so I think the growth opportunities there, I think, are significant. I think we have significant growth opportunities in Latin America as well. And Americas, let's face it. We are making 60% of your business, I hate to say, so goes the Americas, so goes the performance of the company. It's just the nature of the beast. But I'd say relative to your point about -- I feel good about where we're at, that the new CRO, his leadership team is established, he's made some changes. We have a new Asian Pacific leader. We have a new Americas leader. We have a new Latin America leader. We have a new channel leader. So...

And this is all in the span of the last year?

In the last year. In some cases, it was an individual retired. And in some cases, it was just a change that was made. So I'd say we have a very seasoned leadership team that has kind of seen the movie of scale. And so we're evolving, as you can imagine. We are -- it was always a very scrappy sales force, and don't want to lose that nature, but we also want to build some sophistication into -- when you sell into large customers, you need to have a level of enablement that I'd say the new CRO brings.

Jim, I wouldn't be a card-carrying software analyst if I didn't ask you about Generative AI. I know you have a couple of ideas on how Generative AI is going to be operationalized for yourselves, but for customers, but just kind of in broad strokes, what's the strategy here? Are you immediately monetizing this? Is this just going to be a vehicle to make Davis that much more of a sticky user experience?

So it's probably more that. So obviously, GenAI just at the highest level, GenAI -- the promise of GenAI means you're going to have more developer productivity. More developer productivity means more workloads. More workloads means more opportunity for observability. So just that general theme, we have nothing to do with. It's just literally, if they drive productivity and you can drive more -- I would even argue that potentially you introduce more problematic workloads they get introduced because that you're now leveraging GenAI that maybe has not been tested as well. But our direct monetization, we have -- we call it Power of 3. So we've always had Causal and Predictive AI built into David AI, with the Davis engine. So we've always been able to use AI to identify root causes on the causal side. And then looking at patterns that existed before to predict when maybe an outage was going to occur. What we now have with GenAI is an ability to leverage Causal and Predictive AI. The worry that people have with AI is hallucinations. And leveraging our data engine, which already has Causal on Predictive AI, you kind of know that the hallucinations are going to be limited because you're using a trusted source of data. And the way it will get monetized is that as customers query more and leverage it more, that means there's more activity. More activity means more P times Q relative to consumption. And so that's -- we're not going to -- we're not charged extra for us. So it's not like there's a SKU that you buy or a capability that you buy. Might you want to do something different in the future. But right now, it's more -- we want more leverage of it and the more -- because to us, the more you can get GenAI, you don't need a sophisticated power user. Now you can expand the number of people that can do queries because you can use large language models, natural language models to get more people leveraging the platform, which -- more people leveraging the platform means more consumption, which means more dollars.

Got it. I'd be remiss if I didn't ask you about how you're thinking about the investment envelope. You've obviously demonstrated you can walk and chew them at the same time. So you can grow, but you can also show margin expansion and free cash flow conversion. So when you're thinking about these competing priorities, at the top line, what influence is that having on how you're thinking about sales capacity or driving more efficiency from distribution from partners? Anything you can share with us on the kind of the philosophy from an OpEx standpoint?

So the benefit that investors have had with Dynatrace really from the get-go is you get the benefit of strong growth and profitability, and we're committed to both. And we've always been committed to both. So even when that wasn't necessarily in vogue, you always got...

You did it before, it was cool.

Yes. Exactly. And so we now have 28% operating margins, call it, 30% pretax free cash flow margin. We're the rare company that pays cash taxes. So our philosophy is, we're not going to be telling you, "Hey, I want to invest more, I'm going to erode margins, and let me give you the promise of growth thereafter." We think there's enough that we can drive for continued efficiencies. It's just -- it's a muscle that we have within the company to look for areas of efficiency to be able to reinvest and whether it be R&D or sales. And so we drove 300 bps of margin accretion in fiscal '24. We will not likely do that again. The guide doesn't imply that we will not erode margins. So we kind of think of it as we might not accrete margins in a significant way this year, but they're not going to erode margins. And probably, if we execute well, we'll probably able to have some modest accretion this year. And that's -- we think we're big enough that we'll be able to continue to both, which is make the investments we need for the top line while continuing to be mindful about margins.

How are you thinking about maybe the uptake of some of these COGS or margin hungry products like Grail, for instance? There's just more ingestion, like there is incremental COGS associated with that. So to the extent that scales because DPS just makes it that much easier. Is that something that the model can absorb? And would that maybe be a contributing factor to having more muted expansion because it's a high-quality problem to have?

It's a great question. And your thesis is right, the more consumption of some of these new product areas, make your COGS go up. But just understand that we have -- there's R&D that is focused on, "Hey, here's a new product feature. Here's a new capability. We have a lot of R&D also oriented around ensuring that we continue to drive efficiency in things of that nature. So there are initiatives underway to ensure that we continue to drive the needed cloud hosting efficiencies that we need. So we do believe it can be funded within the model. And that's what's great that we -- it's an R&D organization that focuses as much on innovation as it does continuing to drive improvements in the efficiency and the way things are done. And by the way, Grail by definition and the way Grail was architected, is very efficient. And so more people leveraging it, yes, you will increase your cost, but the more we can drive around innovation to help with that, the better unit cost we can drive out.

My very last question for you, Jim, is distilling a lot of what we talked about, how you get the business from, what is 15% ARR growth right now back up to the 20 ZIP code, the 2 things that need to go right and 2 things that are keeping a little bit paranoid that could maybe have you take a little bit of a detour to get back to 20% plus on ARR growth?

I really believe we are not market opportunity constrained. So the market opportunity exists for us to get growth back with the 2 handle, that we need to get this go-to-market pivot right?

That's a good thing.

Perfect, but we need to get it right. And so that's probably the #1 thing that we need to get right. We certainly have the products to be able to fuel that there's more in the pipeline that we will be announcing. So I don't think we're opportunity constrained. I don't think that, of course there's always things you can do on the product side. But I do think on the go-to-market side, we need to get these changes settled in, in turning into productivity, and I'm optimistic that we will. I think that it -- we are still early days. I can't tell you, "Hey, do I have a proof point that suggests we're there." But that's probably the #1 thing. And then I'd say some of the newer areas like logs, I think that there's a lot of room, not that application security doesn't have room for growth. But logs just by definition, is just a very, very large TAM, and I think that we have to continue to get better traction. So I wouldn't say it keeps me up at night, but I would say consistent focus on are we getting what we are looking for on the go-to-market side.

Terrific. It's always a great conversation, Jim. I appreciate the time.

Thank you.

Thank you.