Informatica Inc. (INFA) Earnings Call Transcript & Summary

Good day, everybody. Tyler Radke here. Welcome to day 1 of Citi's Tech Conference. I am the Co-head of U.S. Software Equity Research. And for the second session of the afternoon, we have Informatica, Mike McLaughlin, the relatively recently appointed CFO. I think it's been about 1.5 years in the role. Thank you for coming to the conference. We were just -- I'm glad you had smooth travel coming out here after both of our travel issues at the beginning of the month.

But Mike, it would be great for folks in the audience who may not know you, just a quick background of Informatica and then of yourself, 1.5 years into the role, what have kind of been your biggest focus areas.

Sure. So I joined -- Tyler is right, I joined about 1.5 years ago, in January of 2013 (sic) [ 2023 ]. And what I've been doing and focused on since then is to continue the play that started basically in 2015 for Informatica, and maybe roll back the clock a little bit to get everybody on the same page. Informatica is a 30-year-old company, more or less invented the category of independent data management in the '90s. The premier -- the sort of -- the primary product at that time was called PowerCenter. And it was ETL, extract, transfer, load for on-prem data sources moving them into on-prem data warehouses. Great business. Informatica was the leader. But by the time the mid-20 teens rolled around, the world had begun to shift the software world from a perpetual license maintenance model to a subscription term license-based model. And it was becoming more clear that cloud was the future for workloads, and therefore, data. And Informatica needed to begin a pretty deep transformation to be relevant in that kind of a world. So the go-private thesis in 2015 was not about cost cutting, it wasn't about breaking up the company or consolidating the sector, it was about reinventing the company, creating from scratch the industry's leading cloud-native, multi-tenant data management platform for the enterprise. And that's what began shortly after the go-private. And $1 billion of R&D later, that's what we have today, is the industry's best products for data management across all the key categories. That's according to Gartner, Forrester, et cetera. You don't have to take my word for it. Offered on the only built from scratch, modern cloud-native platform for data management that integrates all of the capabilities into a single pane of glass using the latest drag-and-drop tools on a simple, customer-friendly consumption-based pricing model. And doing all that, the best products on the only platform, as the Switzerland of data for the enterprise. We serve and support all of your software vendors that are sources and targets of data, all of the clouds that you use for your compute and your data needs. And our cloud platform supports your on-premise and hybrid deployments and environments as well. So after all that hard work to create really this platform from scratch, while continuing to run the on-prem business, has been a pretty deep and challenging transformation. When I joined at the beginning of 2021 was right at the beginning of the next chapter of that transformation, having -- prior to my arrival, the cloud product be developed, the shift off of perpetual license to subscription being essentially complete, the move to being a cloud-first company. But at the beginning of 2023, we shifted to be a cloud-only company. So we went end of sale or all of our on-prem products, not just the perpetual license products and reshifted all of our energy to the cloud. What did that mean? And therefore, what was I working on and focused on? We had 2 restructurings last year, which were driven by the cloud-only strategy. We were able to streamline and rationalize our R&D efforts because we no longer were developing to 2 different code bases and 2 different delivery styles, on-prem and the cloud. We're able to simplify and streamline our go-to-market efforts because we are only selling cloud. We're not selling cloud and on-prem. So that simplifies your technical sales, your presales, and a similar theme in support and other G&A functions. So 2023 was a big transformation year to make sure that we got off the blocks correctly as a cloud-only company and incentivize the sales force correctly to sell it and make your customers adopted it. And they did. You can see it with our 37% cloud growth in 2023, and get our cost structure in line to reflect the increased simplicity of the business and the ability to earn healthy returns out of the business. And the steps for transformation continues. So for example, the introduction of PowerCenter Cloud Edition in August of last year, which is a new and very customer-friendly way to migrate your existing on-prem Informatica implementations, whether it's PowerCenter or on-prem MDM to our cloud. So executing against that has been a big effort and it's been a big tailwind to the migration part of our business. CLAIRE GPT, our in-product LLM-based interface with our product was in private preview for much of 2023 and was made GA in May of this year. And beyond that, I spend a lot of my time with folks like you, trying to help increase the understanding of what we're doing, what's the trajectory of our business is. As just described, we're a transformation story. Now we're in the later innings of the transformation. But it's not done, and we have 3 different types of revenue in our income statement, and we've got migrations from 2 of those types of revenue into the cloud. We've got COGS rolling in from what was an on-prem model, so understanding how that's going to impact your gross margin. So helping folks sort through that complexity and understand the medium-term model is another big focus of what I spend my time on.

Yes. And I would say, definitely for the company of your size, the complexity is pretty significant. I think it's one you've done a good job of trying to distill. I know you have a new slide deck that you posted that's I think breaking that down, as well as the Analyst Day. But I guess as we think about the near-term demand environment and the puts and takes on what's driving your business, you recently reported Q2 results a month ago, and very strong cloud ARR growth in the middle of still a pretty uncertain macro environment. I guess talk to us about what you saw, what's giving you the confidence to raise the guide on FY '25 ARR. I mean it's not a -- I'm not sure folks would consider you a beat-and-raise company per se. So I think the fact that you did raise it clearly signals something positive that you saw.

Sure. So let me just step back. There's a lot of debate or different narratives perhaps that you're hearing from some software companies about how tough or not the market is out there for selling an enterprise software these days. We're not really feeling much of a change, talking macro. Deal cycles are long for what we do. They've always been long, but they're not getting longer. Deal scrutiny is high for our deals because it's mission-critical workloads, generally big ASPs, but the scrutiny is not getting higher. And our customers seem to be pretty healthy. The macro, macro, not just the tech macro, is pretty good. The economy is doing well and the need to use data to make your business better continues to be strong. So the macro feels the same to us. I'm not saying that it's great or that you should raise your estimates on us because of the macro, but we're not seeing the headwinds that maybe you're hearing from others, at least not at this time. In terms of confidence for raising the guide, we're in an annual business, very fourth quarter loaded as enterprise software tends to be. First quarter small, second quarter bigger, third quarter, and then sometimes almost 40% of the year is in the fourth quarter. So when you start out the beginning of the year, the visibility is low, and quarter-to-quarter there can be a standard deviation around expectations. And so we try to remind people of that. And this year has been very good in terms of the linearity from quarter-to-quarter being more or less what we expected and very similar to what we saw last year, a little better on a linearity basis. And as we get closer to the end of the year, which is the full year guide that we raised for cloud, we just get more visibility into what is in the pipeline and where it is in the sales process and there's probability of close. So it's seeing actually what we have in the pipeline and in the book of business and how the performance so far this year as compared to past experience, those things combined give us the confidence to modestly raise our cloud guide to the year.

Yes. One of the things that stood out to me last quarter was just the large deal momentum. I think you -- in particular, you talked about a large, I believe, GPU manufacturer that signed a multimillion dollar deal. Talk us through that use case and sort of what products and challenges they have and sort of what you're doing for them.

Yes, sure. And without naming any names, I would add that our business of this type across the semiconductor industry right now is very strong. They're not the only folks that are buying us for the first time. In this case, and in many of the cases, the use case starts with master data management, or MDM. As you may know, data management in our portfolio consists of kind of 3 big pillars. One is data integration and application integration. So ELT and ETL is within that, but it's only a part of it. That's a little bit less than half of our business, is data integration, app integration. The other half is master data management and data governance and catalog, again, rough numbers. Master data management is more of an application than infrastructure technically, but it's so closely aligned and so often, as this use case will demonstrate, sold hand-in-hand with data management because it's mastering data. So managing the data, getting it in, quality controlling is essential. And it's a big part of our business and it's growing really, really well. In the semiconductor case, you're making millions of chips. And those chips have a supply chain, they have a prominence, they have a failure rate. You need to know where they go, and you need to understand the supply chain going into it. And so you need -- so mastering the data for those chips is, because of the volumes involved, had 5 years ago not really been practical, but the technology and the ability to use the cloud to do it on a global seamless basis. And with the Informatica underpinning it, you can now have very granular, very accurate single source of truth data about all of your GPUs or other semiconductors that you're producing and sending out into the market. So that's what they're doing with us. And it's a great example of buying MDM, but also buying IPUs, which is our consumption unit that applies to everything else other than MDM on our platform, which we sell on a per record basis. to do data quality, to do data governance, to do data integration. It's really sort of the circle of life example.

Got it. So that deal in particular was sort of 50-50 MDM and IPUs or...

Something like that.

Yes. Okay. Got it. And certainly, that sounds like a really compelling use case, I guess, sort of related to AI because it's the chips that are making AI. But out at your conference in May, I think it was, out of Vegas, you had this tagline "Is your data ready for AI?" which I thought was a great tagline. But I'm actually curious like as you -- of the new business that you're doing, how much of those use cases are actually Gen AI or AI use cases? Because while that's really important and there's a lot of attention, there's also just a lot of basic operational non-AI use cases that oftentimes sort of get reprioritized just to kind of get your data state in order. So how do you sort of think about the mix of AI versus non-AI use cases in terms of the new business that you're...

So I'll start by correcting you, that our tag line is "Everybody is ready for Gen AI, except your data." You were close. But it really is resonating. And I'll try to answer the question there that there's a lot of debate about where we are in the cycle and the ebbs and flows and NVIDIA's wave-over or what have you. In terms of waves out there of customer top -- things that are top of mind for customers with respect to Gen AI these days, it feels to us and our customer conversations reflect it, that large enterprises, which are now thinking about GenAI use cases that need to be done at scale and need to be done in a responsible way, not just the simple e-commerce chatbot or the generate marketing copy or summarize a contract, that data, getting your data in order, is a prerequisite. And that wasn't appreciated, I think, to the degree that it should have been 6 months ago or even 3 months ago as it is now. There's a great webinar from Gartner from a few weeks ago, that we had nothing to do with, but our CTO sat down and watched it and immediately sent an e-mail to all of us in the executive team that says, "This is a freaking Informatica commercial." I mean, because they are helping educate everyone and everyone is coming to realize that, to do Gen AI at scale responsibly, you need to invest in metadata, which is what we do. You need to invest in observability, which is what we do. And you need to invest in governance, which is what we do. So we're in a really exciting point from all places in the stack for Gen AI enthusiasm irrespective of what may be going on elsewhere. And we have pilots, proof of concepts for numerous customers, large global customers, that are being run today with Informatica as the data management layer facilitating those POCs and pilots. Not in use case at scale for the most part, but many of them will be. So that's beginning to ramp. And we're not putting material or even any expectation for Gen AI-specific deals in 2024, but when we get to guidance for 2025, we'll see how many of those we think are going to turn into revenue, but it's going to be a non-0 number for sure. With respect to the displacement or substitution part of your question, are the enterprises having to take spend out of 1 pocket to put it in the Gen AI pocket? From our perspective, we're not -- it may be happening, but we're not really seeing it. And maybe it's because other use cases that they're using us for now and Gen AI both require data, and the stuff they're using us for now for the most part is pretty mission-critical, pretty consequential. And so turning that down -- it's not just about -- it's not about optimizing it like you would for your hyperscaler usage. It's kind of an on/off for the stuff that we're already working on. And people just aren't turning our stuff off to fund AI. And I think part of it too maybe is that we just aren't seeing the real spend in a big spend at the enterprise level yet other than this preparatory stuff of these pilots. But all of that spend, we think, is directionally going to be quite favorable for us.

Got it. Got it. Okay. One of the other main topics outside of AI is sort of the discussion around the industry of open table formats, iceberg tables has been very topical in the ecosystems of Snowflake and Databricks. How do you see this technology evolution at Informatica? Is there a potential risk that if more data gets stored in iceberg tables and maybe lives in a Snowflake or Databricks, there's less integration need? Or how do you sort of view that technology?

There's a lot of chatter around that topic. And virtually every meeting we have at a conference like this, somebody asks how bad is iceberg tables going to be for you. I'll start by saying, from a customer perspective and the conversation -- as far as I'm aware and I'm aware pretty much of what goes on in my company. not a single deal has been delayed or put on hold because of iceberg. Not a single architectural decision has changed because of iceberg. We're not taking our pipeline or our forecast or conversion rates down because of iceberg. It's just not a factor. Why is that? Iceberg is a data format at the end of the day. And it doesn't manage your data. It stores your data, it makes it available in a convenient way and a flexible way. But you still need to get it from the source of the data and you need to get it into the iceberg table. You need to synchronize it if the data changes and you need to send it back out. You need to do transformations to normalize it and get it in there. And if you are an enterprise, which is our customer base, your data is in multiple clouds. It's not just in an iceberg table at 1 hyperscaler or 1 cloud data warehouse. I challenge you to name a sophisticated customer that is that reliant and that vendor locked into 1 data warehouse or analytical data store vendor. So to the contrary, we think that it's probably going to be positive for us because complexity is good for the independent sits in the middle data management vendor like us. The more format there are to support, the more places data needs to move in and out of, the more demand there is for connecting it, keeping it under control and governing it.

Yes. Got it. Yes, and I think it's still very early, I mean, even for Snowflake, they haven't really seen the full impact by the customers. But I appreciate the response and good to hear you're not really seeing it either. As I think about your partnerships and relationships with the hyperscalers and a lot of the leading ISVs, I mean, there's a lot of talk about it. I remember out at Informatica World, you have the -- I think you have this -- your conference was at the same time actually as Microsoft. So it was a virtual appearance, you had Scott Guthrie, had appearances from Databricks and Snowflake leadership. Talk to us through beyond just integrating with them. How much of a leverage in terms of go-to-market in terms of referrals and driving new business for you are those partners? And what's sort of the strategy going forward with them?

Partners are a very important part of our business. And there's 2 big categories of partners. One is the hyperscalers and Snowflake Databricks, and the other is GSIs and VARs and systems integrators broadly defined. Both very important, slightly different dynamics with the 2. I'll start with the GSIs, the system integrators, a little more straightforward. We disclosed that on any given quarter, including the last quarter, that 30% of our closed business is sourced by partners. So -- and like sourcing business is the holy grail. How do you find your next customer? How do you get the Glengarry leads, right? And 30% of them come to us and they close. 70% plus, sometimes in some quarters, 90% will be co-sold with a partner at our side at the table, often in conjunction with the hyperscaler partner, the 3 of us at the table from the inception of the -- conception of the workload and what you're going to do with the architecture. And so we're side by side. We don't do any sell-through, it's all sell with. So it's all in our paper, we control pricing, we control renewal, all that sort of stuff. But we are very, very symbiotic out there in the field selling together with the GSIs, and now get to the hyperscalers. So the GSIs love us because it drives a lot of service revenue.

And that's the 30% you're talking about...

That's the 30%. Some of it does come from the hyperscalers, but the bulk of it is from the GSIs. And again, one more comment on the GSIs. We've already certified over 10,000 new individuals at implementation, integration partners around the globe on Informatica this year, and it's going to be close to 15,000 by the end of the year. I mean that is a small army of people out there. Deloitte has thousands and thousands of people that just do Informatica. And so they've got a tremendous incentive to work with us, and they do. I mean I wouldn't trade our position with those people for anybody's. The hyperscalers, they don't do the implementation, they make their money off using their compute and their stores and their services. And they see us as facilitating that as well. So the easier it is to get your data from all your different sources into Snowflake, the better it is for Snowflake, the better it is for Azure, and so forth. Now that being said, those folks have some captive tooling to do some parts of the data management within their own ecosystem. And those are fine tools. They're not as good as ours, if you ask Forrester and Gartner and so forth. They're not as integrated. And they don't integrate with the rest of the world the way ours do because it's proprietary, it's walled garden. But they're good enough for some use cases that are simple and where the customer is happy to keep their data in 1 location, they don't need to do hybrid or multi-cloud. So how do we manage that since we're sort of competing? We don't -- we just don't compete for those use cases. That if the captive good enough within the walled garden products are fine for your use case, good for you, we're not going to spend any time chasing that business. It's for the complex enterprise grade, multi-cloud environments where we compete and we win. And in those, the hyperscalers know that their chances of winning that business are better if we're at the table with them.

Right. Right. Okay. That's very comprehensive. And that 15,000 is -- of new consultants certified just in 2024?

Yes.

That's a big number...

We're not there yet, but we're on 10,000 now. Yes.

Yes. Yes. Got it. Got it. Okay. So sticking on sort of the go-to-market theme, there's been a number of evolutions of, obviously, the pivot to subscription and then the pivot to cloud-only recently. I guess as you step back and you sort of think about when the company made the pivot to cloud-only, which I think was right at the beginning when...

Yes. It was 2 weeks after I joined. No, actually it was 2 days after I joined.

Two days, right. So cloud-first basically since you've been there. Would you say that this transition to cloud, it's happened faster than you would have expected? Or just sort of give us what's the biggest areas that surprised you as part of this pivot to cloud-only selling?

I think it has gone pretty closely to expectations. We expected it to be hard and not without issues. And everybody in the go-to-market organization had to adjust to just selling cloud and figuring out how to not pay attention to this juicy nonstrategic on-prem opportunity that's sitting there, and why can't I spend time going after that and get paid just as much as I would if I land a cloud deal which isn't as far along in the cycle and is arguably going to be harder work. I want to pick the low-hanging fruit, right? So we anticipated that and we set up compensation incentive programs to incent the behavior that we want. So the short-term stuff, we disincentivized. And the things that's strategically important, we prioritized. And so we knew there were going to be those issues, and the way we put in place to deal with has been working really well. The acceptance of the product has been really good. The IPU pricing model, was not new with cloud-only, we had already introduced that, but we cut off everything but IPs. So you can only buy IPs and MDM records. And the reception there has been good. So it's gone as close to plan as one could hope.

Yes. Definitely. And so on the IPU consumption, which certainly the vast majority of your cloud contracts now sort of being under IPUs, is impressive. Obviously, your consumption model is a little bit different than everybody else's, right? It's a little bit more consistent revenue recognition, which I think is good. But help us understand the utilization of those IPU credits, and certainly, we're seeing pretty healthy net retention rates in cloud. But this is still somewhat of a new motion for you. So where are you sort of seeing those -- the utilization of those credits? And is there any risk that you're going to have to go through an optimization cycle like many of the other cloud consumption names have gone through?

So just to get everybody on the same page, the way our consumption model works, it's called the IPU, the Informatica Processing Unit, it's a fungible token of capacity that you can spend on any of the 36 different services that are available to you on the IDMC platform with your IDMC login. You consume your IPUs based upon a scaler depending upon what you're doing. So if it's ingesting data, it's the amount of data you're ingesting. If you're doing pushdown transformation, it's the amount of compute that you're using. If you're cataloging something, it's the number of things you're cataloging. But the IPU you can use to do any of those things is up to you. When the customer comes to us says, "I have this use case that I'm interested in potentially using IDMC to solve the data management," we put a team in there, we analyze what they plan to do, what volumes of particular services they're going to need to do it, and we give them an estimate of here's how many IPUs you're going to need, say it's 100. The way -- what we recommend and what our customers always do is agree with us, but you need more than this expectation. If you remember how cellphone plans used to work, you have the choice of spending the 500 minutes and the 750 minutes, and I think I may use 400 but I'm going to get 750 just so I don't go over. And so I have the flexibility and the peace of mind, right? It's sort of like that. And so we intentionally oversize it, so they've got some extras. It gives them surge capacity and it gives them the ability to use the IPUs to experiment. They don't have to call -- if they want to pilot something or they want to do a POC, they don't have to call us up and say, "Hey, can we get some free usage of IDMC so I can stand up this pilot? And then we call revenue, I promise, if it works." You don't need to do that. You just use it, right? And it's use-or-lose capacity. So you don't -- there's no true-up at the end. You don't get money back. We bill you annually in advance and we recognize it ratably each year of the multiyear contract. So from a CFO's perspective, it's a thing of beauty. We watch the utilization closely. We've got great telemetry that a customer can see and we can see, how much are you using, how much you have left, what are you using it for, where are you running hot, et cetera. And what that allows us to do is to have a very real-time view on customer health and how much value they're getting from the product. And so in our customer success team, which is largely low-cost geography-based, we have rooms full of people that are sitting there watching the telemetry and they've got automated AI-assisted stuff that's saying, "Here's where we should focus today because this customer appears to not be using enough to suggest that they're a healthy customer and they're going to be a renewal risk a year from now or 2 years from now when their contract comes up. So let's get in there quickly. And what are you using it for? Can we help you find value or come up with a different use case." And on the other hand, here's a customer that's running hot. They're at 95% utilization. Why don't you give them a call and see what they're doing and see if they need more. And sure enough, they do. And that part of our business is that interim expansion is growing like crazy. So it all gets to your question about net retention rate, which has been very good, on an end-user basis has been mid -- sort of 125% plus or minus. And that retention, as you know, is a combination of what new stuff you sell and who you sell it to. Is it to an existing customer? Then it's net retention, as opposed to a brand-new customer. As we disclosed, of our NNARR, our new ARR every quarter, about 25% of it is migration, 75% of it is net new, winning in the marketplace. And of that 75% is about 2/3 of it is to existing customers, because we've got 65% of the Fortune 2000, we've been in business for 30 years so we have a lot of customers. But they're standing up new workloads with us. So it's how much of the new sales are to those existing customers, and that feels very healthy. We've got a lot of time and a lot of wallet still to go at that existing customer base. So we're not worried about that. It's about how much are you expanding in turn, and that was the motion that I just talked about, which is going great as customers find new ways to use the platform with the flexibility of the IPU pricing model. And how many are you renewing at maturity? And again, because of the way we can see well in advance, is that customer healthy, is that customer using the product, we do really well on renewals because we've -- if we've done our job right, we've intercepted the churn potential earlier than we might have had we had a different model.

Yes. That vision of sort of the telemetry dashboard is such a big contrast to where Informatica was 10 years ago.

No kidding, and it's been a lot of hard work. And let me just make sure I answer the last part of your question, is there a risk of an optimization? So look, they are multiyear committed contracts. So you do a 2-, 3-, 4-year contract with us and we bill you annually upfront. And there's no true-up if you don't use it. So in term, they're 0 risk. If they don't use anything, they can optimize all they want, but they don't have the opportunity to pay us any less. But if you get to renewal and they're not using it, they can downsize themselves. But that's not any different than any other piece of software that you would sell. And we are not seeing the sort of short-term optimization cycles in our part of the stack that you've seen in some of the other parts of the stack.

Right, right. So contracts are usually 2 to...

New contract duration is in the neighborhood of 2.5 years. We don't sell anything less than 2 years. Some customers want 5.

Right. Got it. And as you think about the -- so I guess a couple of things. Number one, I think we look at your targets that you've laid out, you're assuming at least a 30% cloud CAGR over the next few years. Help us understand sort of the mental model of how you get there. You gave some great stats of, obviously, you got your net retention rate in the mid-120s. You're still adding pretty good new business, too, in terms of new logos. Is it sort of just extrapolating that going forward? Or anything you'd sort of point to that gives you confidence in that 30% being sustainable?

Yes. So there's a couple of different ways that people like to look at it. One would be net retention is going to give you this amount of growth and then you have to get this much net new from the outside to get to that total. That's not the way we forecast and the way we think about it. You're welcome to, if you want, it kind of gets you to the same place. The way we think about it is we think about what the market opportunity is in the net new part of our business. So as I mentioned quickly, you may have missed it, of our NNARR, our net new ARR in the cloud for the last 12 months, 1/4 of that was from migrations. So we're taking existing maintenance or existing self-managed on-prem customers, and they're migrating to our cloud. So that's 25% roughly of our NNARR over the last 12 months. And then 75% is the net new. That ladders up to what we project to be 35.5% total growth for the year. Of that 35.5 points -- roughly 25% of that is going to be migration. So whatever 25% of 35, is like 7.75, is going to come from migration, and the other 27 or 28 is going to be from net new. Okay? So the net new part is, does the market support mid-20s percent growth without having to take any market share or acquire anybody or some sort of heroic tailwind that we feel? And we don't think it does. You look at the IDC TAM, which we have in our investor materials, they think that cloud data management grows at 27% through 2027. If you look at how the cloud data warehouse companies are growing, in the mid-20s, we should be growing at the same rate because we're managing the data that's going in and out there. And you look at the pipeline that we have in the near term, that it ladders up to that net new growth in the mid-20s, and migration on top of that to this year getting us to 35%. And then if you look at our medium-term model going down to 2016, we said a 31% to 33% CAGR over that 3-year period, so you can do some straight-lining if it's going to be 35% this year, maybe it's 32% next year and 29% in 2026. Not guidance, but straight-lining, simple math. The end market seems very comfortable to support that. And then migrations, which we have good visibility on ladders it up to the number.

Yes. And that's a great segue because I did want to talk a bit about the migration opportunity. You obviously have a really nice growing cloud business, but a huge chunk of the business is still sort of in the self-managed and maintenance piece. So how do you sort of think about the -- I guess, first, if we look at that book of business, is all of that -- does all that have the potential to move to cloud? And how are you thinking about the timing of that migration? And also, what levers do you feel comfortable using to sort of manage that migration?

So we're still in the early innings of migration. We still have roughly $450 million of maintenance revenue in ARR and approximately the same amount of self-managed subscription or term license on-prem ARR. The total we've contracted to migrate to date is something like $70 million or $80 million. So less -- that's like 6.5% of that number, and we update that every quarter. So it's early innings. We have enjoyed life to date at 2:1 uplift multiple. So if you're paying us $100 for your maintenance or self-managed today, on average, you're paying us $200 when you contract to move to the IDMC and do your data management from the cloud. And there's a lot of dynamics in terms of which -- there's a pretty wide standard deviation around it because we sell it on value, not on foot multiple, but anyway good uplift and healthy uplift. The pace of that migration is accelerating because of PowerCenter Cloud Edition, which we can talk about because we made it easier for customers to move. But it's really a customer-driven decision. It's not like a simple application that you can say, "I'm going to move this application and just do it in cloud instead of this." We're infrastructure software, we sit in the middle of a stack of a use case or an IT environment. And the primary driver, the motivation for a customer to move is digital transformation. We need to move the whole shoot match to the cloud or we have a road map to do it and we're going to start doing it in pieces. It's not because your friendly Informatica salesperson showed up and say, "Hey, how do you like a good deal on some cloud data management software today?" And they say, "wow, that sounds good. I'm going to buy now before the price goes up next year." It's not how it works. They've made a high-level strategic decision to transform their legacy on-prem stuff to modern cloud and they do it when they're ready. Us pushing them or jack-in-up price on the existing stuff, it might have some effect on the margin, but it's our judgment that it's going to do more harm by pissing off our loyal customers than it is to accelerate the move to the cloud. We do have some carrots. We offer some implementation services credits. If you do it, we give you credit for the on-prem that you're using now so that you're not double paying during the implementation time. But it is a customer pull, not an Informatica push. And just by the nature of where we sit in the stack, it's a transformational project that customers don't take lightly and they do it when they're ready. Now the forces of that, the digital transformation, moving to cloud in general, are as strong as they have ever been. And that's why our migration business is growing somewhat faster than the rest of our business. We're happy about that. And GenAI should and seems to be reinforcing that and making digital transformation even more urgent for our existing customers.

Great. Well, maybe in the last minute, I'll open it up to you if there's anything you wanted to cover, just any message you wanted to leave with the audience before we wrap it up.

I think that what I'd love you to leave with is to have a very clear understanding of the building blocks of our financial model over the next few years. We have 3 categories of revenue in ARR. We have maintenance on those old perpetual licenses that we don't sell any more, constant decline at a rate we've got a pretty good handle on how that's going to behave. We have self-managed or on-prem term licenses that we don't sell anymore and are going to be in perpetual decline, and we've got a good sense on how those are going to behave. And then we have a cloud business that's almost 50% of our ARR revenue. Now that's growing, 35% this year, 31% to 33% CAGR through 2016. If you just do a simple 3-line Excel model and put a decline rate against the maintenance, a decline rate against the self-managed, and a growth rate on the cloud, and you can complicate it by migration if you want, it becomes pretty obvious that, as the cloud keeps growing, it becomes a bigger and bigger portion of the total. So the growing part gets bigger, and the shrinking part gets smaller as a contribution. And so the overall growth rate goes up. And we've committed in our medium-term guidance, which we offered in December, which we're still comfortable with, that we're going to exit 2026 with 11% or better total ARR growth. Total ARR growth. And that's because by that time, cloud ARR is going to be roughly 70% of the total, and it's still going to be growing with a 2 handle. And the shrinking bit is only going to be 30%. And it's still going to be declining but it has less impact on the total. And so it's not complicated math. And in our view, it's not a heroic leap of faith to see our revenue growth going from 6-and-change revenues, 7-and-change ARR this year, to high single digits next year, faster than that in '26 and faster than that in '27 and beyond, putting us in a position of being a sustainable teams grower while we're making money at the same time. Our operating margin this year is guided to be 34%, and it's going to get better and better as we scale and we have more economies as a well-run software company.

Great. I think it's a great place to leave it. Mike, thanks so much for coming. Thanks, everyone, for joining the session, and have a great conference.