Informatica Inc. (INFA) Earnings Call Transcript & Summary

My name is Tom Blakey. I cover infrastructure, software here at Cantor. We're delighted to have the CFO of Informatica here with us; and IR in the front row, Victoria Hyde-Dunn.

So maybe just without further ado, you just reported the fourth quarter earnings. Maybe for those who weren't able to listen to the call, what were you pleased with and what did not unfold as expected?

Yes, sure. First of all, thanks for having us here at your inaugural conference.

Thank you for being so generous of your time.

Hope to be here many years in the future from now. So it was a quarter that did not meet all of our expectations for sure. But that being said, it was one that directionally was consistent with our strategy and gives us increased confidence that we're on the right path and that we have, in this highly dynamic and fast-growing space of data management, that we're well-positioned and have the opportunity to create a lot of value in the future. We delivered cloud ARR growth of 34%. We had called for 34.5%, so it was a little shy of that, and we can talk about it. But that's the largest dollar amount of NARR that we've ever delivered cloud-wise in the history of the company, and represented excellent execution on our -- part of our sales force, and strong receptivity of our products by our customers. We delivered total ARR growth of 6%. Now we had expected it to be closer to 7%, so not where we thought. But reflects the fact that we are managing the decline of our on-prem business while we grow the cloud business, and we're migrating a very healthy portion of those on-prem customers to the cloud. We delivered operating margin improvement, non-GAAP operating margin of over 400 basis points improvement year-over-year versus 2023. And our free cash flow, unlevered free cash flow after tax, exceeded the high end of our guidance. We felt good about all of that. We felt badly about not meeting our expectations. And the places where we fell short started with renewal rates. We did not see the renewal rates we expected in either our cloud business or the self-managed and maintenance part of our business. So self-managed and maintenance are decline businesses. That's legacy on-prem software that we've sold in the past. And then the cloud business, of course, is the growth business of the future. Those decline rates were roughly 2%. Those renewal rates were roughly 2% lower than what we had forecast coming into the quarter. I'll start with the fact that, on the cloud, they're still in the low 90s. So we feel as though the gross renewal rate is nothing to be embarrassed by, and we delivered a net retention rate of over 120%: 124%, to be clear. But it's 2% lower than what we forecast going into the quarter. That had to do, in our analysis, not with a changing competitive intensity, not that we're losing deals to other competitors, but rather more or less one-off, idiosyncratic situations that happened in a handful or a couple of handfuls of customers where they had bought software 2 or 3 years ago and the use case was over-scoped and they ended up not implementing as much as they thought they were going to. So they down-sell that, they bought software 2 or 3 years ago. And the corporate mandate to conduct a global governance project, for example, shifted directions in terms of prioritization, in one case, because the CDO, who sponsored the project, left the firm. Another was an M&A situation where one of our large enterprise customers had a 3-way spin, and our software stayed with 2 of the 3 and the third one decided to go out to RFP. That customer is now part of our pipeline for selling them a new deal in 2025. It was stuff like that.

That's an interesting jump-off point there, Mike. These renewal trends seem to be -- spooked the shares. Obviously, we got some inbound on there from questions from investors. It seems like if it's idiosyncratic, renewal rates are sometimes extrapolated in Excel, right? So talk to us about what you expect in terms of continued trends in the calendar '25 with regard to this metric. You talk about you could go back and sell that customer, or this was an idiosyncratic loss, or -- because renewal rates are sometimes -- maybe even are unfairly extrapolated.

Yes. Well, that's exactly what we've done in terms of our guidance for interesting. So we've taken our actual 2024 experience and pushed that through into 2025, with respect to the cloud. For the on-prem stuff, that renewal rate is declining because it's been end-of-sale for 2 years, and so it's becoming more and more out of date. So we expect that renewal rate to decline. But we've also shifted that renewal rate in '25 down by 2 points versus what we would have guided to prior to the end of Q4. So in a sense, we've extrapolated everything into '25. Now we are hopeful that it's better than that, and we've taken actions to, organizationally, in terms of incentive structures and some systems and processes that will allow us to do better on catching some of those idiosyncratic things earlier and addressing them with the right teams to save and rescue more of those situations. But we haven't assumed any of that in our forecast going forward.

Maybe double-click on that. That's an interesting point about -- and maybe nothing even specific. Just what can you do? Like if I'm a customer, and you labeled it as idiosyncratic before, but now we're talking about the next customers -- or maybe it's back to that customer, like you said, there's an opportunity at that existing customer. It seems like there's a few -- there's always a myriad of opportunities to go back and try to change renewal rates. What are you doing precisely or -- with the go-to-market changes there?

It starts with awareness and rigor around identifying, catching and acting on the early warning signals. So...

For the customer who has...

For an existing customer, right? So we have really good telemetry from our cloud products that tells us how they're using it, how many IPUs they're consuming, on what services they're being used on. And we have data from our help desk. We provide maintenance -- we provide support and maintenance, and so how many times they're calling in, what they're asking about. And then we have direct relationships with those customers on the field sales side because they're expand opportunities. So we have account executives in there who should know what's going on, and not only with respect to new business, but, "How is it going with the stuff you bought last year or the year before that?" That information is there, but sometimes folks don't focus on it early enough. And so making sure that we are surfacing that information not 2, 3 months before the renewal, but 6, 9, 10 months before the renewal, or more, so that we can see that there's an at-risk renewal coming up, and then deploy the right team against it. So that's the awareness piece and the focus piece. And then to get the right team deployed against it, you need incentives. Because field sales, their first job is to sell new software. Renewals, their job is to renew what's out there. But for a big enterprise renewal at risk where it's under-deployed or the workflow or the workload has been stalled, it often requires field sales and technical sales to get in there and almost resell the deal. And you got to make sure that the field sales team in the right situation is properly incentivized to work hand-in-hand with the renewals team to get the right outcome. And we had a process in place for doing that, but it was sort of ad hoc, one-off, unpredictable. And so we've made that more formalized so that, when renewals team raised their hands and say, "Hey, we've got a yellow flag," and it's 9 months ahead of time, here's how we're going to incentivize field sales to collaborate with us and get to the right outcome.

So that sounds pretty real time, if I'm understanding correctly. And these things are probably galvanizing pretty quickly. Again, just as a remedial spectator here. What are you seeing? I mean you'll give us some anecdotes in terms of what you're seeing, "Hey, I've raised my hand as a flag, you helicopter in with me," and what are you seeing?

It's a little bit too early to declare victory. We're an enterprise software company with typical enterprise software linearity. The fourth quarter is 35% plus of our year and first quarter is less than 20% of the year. So that's a new business basis, which means, since most deals are 1 year or 2-year -- well, very few are 1 year; 2-year or 3-year or 4-year, the renewals fall on that. So there haven't been many renewals come up in the months of January and February. But that being said, I can give you 100% assurance that everybody is laser-focused on this, because it was an impactful mess and we have put all hands on deck to address it. But whether these structural changes that we've made, again, which aren't massive. We're not changing the deck chairs around, we didn't fire anybody, and we're not doing a massive change to compensation system. They're kind of around the edge or marginal things. But whether those work, and we sort of have autopilot focus on this in 2025, remains to be seen.

So let's shift back to something a little bit more fun to talk about, in terms of your cloud business. Very successful. You talked about the fast growth rates, you talked about the solid NRR and the expansion rates therein. I've always been interested in the growth from expansion deals and net new logos. We can talk about that as well. But just along the lines of that kind of color, talk about what you're seeing in terms of cloud migrations. And we can start higher level and talk about what's kind of driving those. But I'd love to kind of double-click in terms of, from a sustainable perspective, how that net expansion of -- that expansion business and new logo business has been growing.

There's 2 or 3 levels to that question, so let me try to answer it from the top. If you think about our cloud growth, it's made up of, big picture, 2 components. One is net new customers or new workloads with existing customers. So it's greenfield. It's not something that has to do with a legacy on-prem implementation that we may have with them. So in the past, that's been about -- well, to be precise, in the last 12 months of our net new ARR in the cloud, over the last 12 months ending 12/31, 68% of that was brand-new logos to Informatica or new data workloads in the cloud.

And that's up.

And that's 68% of the total. That's actually down. It was 75% of the total at the end of Q3 for the LTM, and we'll talk about that dynamic because it's an important one. But at a high level, think about it in 2025 as being roughly 70-30: 70% new customers, new workloads; 30% modernization of existing cloud workloads -- existing on-prem workloads, sorry, whether maintenance or subscription, into the cloud. So 70-30 one or the other. Second big-picture thing that I would add is that our logo growth in 2024, new logo growth in the cloud was about 8%, to roughly 2,500 customers, 2,467 or 2,468. We feel very good about that. But some who focus on companies that have a different target audience than we do aren't blown away by that number. We're not a new logo dependent company. We serve the large enterprises, roughly speaking, the Fortune 2000. And that's a land-and-expand business. And so we had 8% new logo growth in the cloud last year, but we had 124% net retention rate with our existing logos. So that net retention, that expansion is more important than new logos. Both are important. We focus on both, of course. But it's really important that you understand that when we land in an enterprise customer, that that first use case is never the TAM in that customer. It's a small part of it, and there's lots and lots of expansion room. Once we get them on the platform, once we get them on the IPU pricing model, good things start to happen. Okay. So that's point two. Then the third point is the modernizations themselves, right? So I talked about 30% of that cloud NARR being from modernization. We've been doing that for 3 or 4 years. It's been accelerating, and it accelerated a lot in Q4. It was north of 35% of our total bookings in Q4, and in fact one of the reasons why our cloud ARR coming out of the year was not quite where we forecasted to be. And part of the reason why we guided to 25% for '25 in terms of cloud growth rather than probably several points higher than that is because that mix of migrations versus net new is higher. So more migrations. That's because there's an accounting outcome of a modernization migration deal that, in the near term, in the initial term of the cloud deal, yields less net ARR that we report than a net new deal. And so that mix matters. It's unfortunate, but it's just the way the accounting rules and there's no 2 ways around it.

What kind of -- what difference is that? What kind of relative difference is that?

It's about 20%.

20%. Wow.

On average. So you put those things together, and the dynamics underlying the modernization business, although it's growing really fast and the contribution is valuable and has a really, really great net present value over the long term, masks a little bit the actual underlying growth in the total cloud business.

Is there something that's happening, some dynamic that's happening in the market that's pulling in the modernization?

It's, simply put, the demand to modernize to the cloud by our customers. And the pressure to do that just gets greater every day. Part of it is, if we're ever going to do GenAI at scale, our data has to be in the cloud so it's accessible and governable. But it's more fundamental than that. There's still so much on-prem and hybrid workloads out there that need to be modernized. That's what's driving it. And just as a quick statistic, if you look at the total amount of our on-prem base that's been modernized, it's less than 10%. So we've still got $900 million of on-prem maintenance and on-prem subscription that are candidates to be migrated over time. It's not going to happen all in 2025, it's going to be a multiyear period, because making the decision to migrate is a consequential one that takes a lot and people don't do it overnight. But it's a big opportunity that will remain ahead of us for a long time. One more comment on that though is that most of our cloud growth is net new, winning in the marketplace head-to-head with the competition for a new logo or a new workload. A minority of our growth is coming from modernizing our existing on-prem customers to the cloud. So you should not take away the fact that we are simply mining our base. We're actually, because we have the best product on the industry's only platform, serving as a [ switching limit to ] data, most of what we do is winning in the marketplace against the competition.

That $900 million, a lot of software vendors will give a statistic about moving, modernizing software, you have to get rid of hardware, you have to get rid of labor. Is there a money multiplier that you've offered in the past about the $900 million moving on-prem? And then related to that, I believe, why would you not, as a firm, as a vision, just change incentives there for the customer, make it economic in some way for the firm itself to kind of hasten that migration?

There's no doubt that cloud software is more valuable to the customer than on-prem software, because they don't have to buy the servers, they don't have to buy the storage, they don't have to maintain it, they don't have to patch it, they don't have to worry about the security, et cetera, et cetera, et cetera. The uplift multiple in terms of what that cloud software is worth versus on-prem varies by type of software and it varies by firm. Historically, our uplift multiple has been 2:1. Now we have actually guided that to be lower in 2025. We've guided it to be 1.5 to 1.7. And it's really important to understand why we guided that down and why we expect that uplift multiple to go down. It's not because we're lowering price. We're not lowering price to incentivize people to move. The price per IPU, which is the Informatica processing unit, the fungible consumption token that is all we use to sell our software, is the same IPU per IPU for a new workload or a migration workload. We're not increasing discounting. It's not because we're lowering price. It is because the nature of those workloads that we are now addressing in that $900 million base are the ones that naturally generate the need for fewer IPUs, to reproduce what you're doing. So initially, when we were migrating what was years ago a $1.2 billion base, we wanted the sales force to focus on those customers that were using their software in a way and who are priced based upon historical purchases they made a decade or more ago so that it would generate the highest possible uplift multiple. Because we didn't know how the cloud contracts would behave. Would they drag additional sales? Would they have good utilization? Would they renew at a good rate? So we focused them on those. It went from 1 to 2 or 1 to 2.5, right? And we put guardrails and incentives around to focus them on that part of the installed base. We now have enough experience with cloud migrations that we know that once customers modernize to Informatica on the cloud, they drag a significant amount of expansion sales with it, their utilization during the term of the cloud deal is excellent, and they renew at a higher rate than our average deal. So we're now willing to open the aperture, if you will, for the sales force to go after that portion of the $900 million that will naturally generate the need for less IPUs and, therefore, a lower uplift multiple. So put another way, it's not about where they land in terms of the number of -- the price they're paying for IPUs or the number of IPUs they need. It's where they came from. Does their existing deal today, are they underpaying for value today or overpaying for value? So it's an output of that, not because of price. So we're lowering that to make sure we're allowing customers the opportunity to go. But we're not throwing incentives at them. Because incentives in our business don't work. Moving your cloud data infrastructure from on-prem to the -- moving your data infrastructure from on-prem to cloud is a complicated, time-consuming, expensive process. And we're only a small part of it. A typical enterprise would have hundreds, if not thousands, of applications that are generating data, dozens, if not hundreds, of cloud repositories, analytic data stores, et cetera. You've got to rationalize all those, you've got to get a GSI in the middle to do all the implementation, and you have to do the Informatica software. So customers modernize when they're ready, not when we provide incentives. So we're not doing that.

Very clear answer. Just double-clicking on the 1.5 to 1.7, I don't want to mischaracterize, is there a -- this sounds like pretty in-depth research that you've done at the firm, I got to admit. Is there a -- the only word that keeps coming to mind, Mike, while you were talking was quality. Is there a different -- is this -- you can characterize the word differently, please, for me. But is there a lower quality of workload coming on to the Informatica cloud at this particular juncture? And if that's the case, why?

Yes. That's an interesting way to look at it. The answer is no. It's...

Because I would think -- lower quality defined by me is just used less...

Yes. So yes, because they need less, but not lower quality in terms of what the utilization is going to be, how much expansion they're going to do and what their renewal rate is going to be. All of those are as high quality as anything we've done in the past. It's just that where they started in terms of how they're using our on-prem software today and what they're paying for it, they need fewer IPUs than the folks we've converted in the past.

Maybe stop and pause here for the audience, if there's any questions?

Can I ask a technology question and a finance question? I'll start with the technology first. [indiscernible] Q&A in the past quarter, you mentioned about data integration, there should be more of that. Have you seen that coming in as part of the competition? I know you guys [ can back it up ], but was wondering what you see outside of things. And then for the finance side, I know you touched upon it earlier, that customer will migrate from on-premise to cloud whenever they're ready. Obviously, your cloud business has grown really fast, 24% guidance, 25%, fiscal '25. You have a great balance sheet as well as free cash flow margin here. I was wondering, why don't you just do the rip-the-Band-Aid approach, where you're investing more in the go-to-market team, as well as your customers to, "Hey, can you migrate to cloud this way? [ That will be great for us ]."

Yes. So let me try to answer those in order. So competitive intensity from folks like Snowflake and others that are -- could be perceived creeping, trying to creep into our space in data integration. The hyperscalers, the cloud data warehouse providers have their own captive native tools to do some of what we do. Those tools are good enough in many cases: if you're not multi-cloud, if you have no hybrid sources and you don't need to connect to the broadest array of data sources possible. And if you're comfortable as a customer with the vendor lock-in that using their data warehouse and their data integration tools provides. That's a fact of life, it is a competitive environment -- it is a competitive environment out there. And in those kind of situations, we generally don't even compete. Because they'll give their tools, their good-enough captive tools, away for free. And if you have the appetite to sort of roll your own with that and live with the vendor lock-in, good for you. We'll see you in a couple of years when you realize the downside to that. So it's important though to understand that some of the comments that folks in that category make about data integration in terms of its importance to their business create a misunderstanding about what is competitive with us and what is not. So I have heard -- I hadn't heard them personally, but I've been told that those companies cite some pretty big numbers in terms of the percentage of their revenue that they say comes from data integration or data management. Like 20%, 25% of their total revenue comes from data integration. That is not what -- that is not our TAM. When they say that, that is the storage and compute consumed by the data integration workloads, the transformations, the data quality, the governance and catalog that is being pushed down. And their compute resources are being consumed to do that. They don't have a data integration SKU. They don't have a data integration SKU. They have storage SKUs and they have compute SKUs. What is frequently the case is that that workload that is consuming compute and storage is being written and managed by us, using Informatica IDMC, and pushed down. We don't do the compute; we don't do the storage. We design the pipeline. We design the transformation. We use our drag-and-drop tools and our recipes to -- in a super-efficient way to create the SQL code that's then pushed down to those platforms, where it runs and they generate revenue from that. So don't be confused by the big numbers they cite about how much data integration. That's not taking food off of our table. Now with respect to your incentive question to get people to migrate, it is a fair question, and it's a net present value question, frankly, that we think about regularly, is should we do something unnatural to accelerate this...

Because a couple of companies like [indiscernible] they have the same on-premise thing, but they did the rip-the-Band-Aid approach where, "Hey, we'll incentivize our sales force, customers. Let's get it done." This is like apples-to-apples comparison than just cloud-to-cloud comparison.

Yes. Look, it's a reasonable thing to consider. We're a profitable company, over 30% gross margin, and our goal is to deliver high single digits and ultimately double-digit total growth in combination of our cloud growth and our declining on-prem business. And that's the financial model we're working to. But look, you can disagree that maybe you should sacrifice your profitability and take it down to 10% or 15% and go through 18, 24 months of the valley of death there, but that's not our strategy.

Any more questions from the audience? Maybe we'll just shift to AI. I mean Informatica benefits from AI, you've been at the epicenter of data in general. But maybe just double-click on that kind of higher-level statement. What are -- you've given a lot of examples here in terms of the forensics, in terms of the usage and the yellow flags and the renewals. I'm just very impressed with what you guys are doing to look inside the usage of your product. Help us understand, take -- use this as an opportunity to help us understand what Informatica is being used for currently with AI and what the current trends you're seeing maybe heading into '25 here.

So for us, AI means 2 things. One is what we call GenAI from Informatica, emphasis on the word from. So that's our CLAIRE GPT. It's a service on the IDMC that you consume with IPUs, which is our consumption token unit. And it's a GPT-based interface that allows the user of our platform to be better, smarter, faster, through all the benefits of natural language GPT, not only interpreting what you want when you type in a sentence and then figuring out what the next question to ask you is and refining and all that cool stuff that GPT does. But increasingly, it has capabilities to use GenAI technologies to look at your data estate and, in an automated way, build data pipelines for you and find data quality issues and so forth. That GenAI from Informatica is a -- it's a service. It's like a Copilot, but supercharged. We're the only ones that have it. We're likely to be the only ones that are going to have it for a good long time because you've got to have a platform that has horizontal visibility of your data for it to be of any use. And it's an adoption, it's a user experience, it's a productivity enhancer. So we're going to win more business. We're going to win more new deals because we have CLAIRE GPT. Because if you buy us versus the competition, you're going to be able to get more out of it and get more out of...

They're connectors to look at other platforms.

Absolutely. So we don't charge very much for it. In fact, we've announced that we're going to continue to allow people to use it for free through the rest of this calendar year, because we want people to use it to build more pipelines and to create more quality rules and to establish more governance on their data, which is what consumes IPUs in bulk, not just using our GPT agent. So that's GenAI from Informatica. Super important, differentiates us from competition. But we're not going to monetize that or a monetization of that is not going to be noticeable. What is going to be noticeable in the future in terms of financial impact is Informatica for GenAI. From Informatica for GenAI, right? So this is using the IDMC platform, which doesn't require CLAIRE GPT, to manage your data for a GenAI workload. That's where the real money is going to be. And you're right, we are in the middle of that based upon the importance of data for GenAI. We're seeing it in terms of proof of concept and trials. We've got over 100 customers who have stood up orgs, which is a word for an environment in our cloud platform, that we can see, from our telemetry, are using GenAI functionality, like connecting to LLMs, doing unstructured data transformations. And so we know that those are GenAI workloads. And in some cases, we have direct relationships and we're tracking them directly, and we have case studies that we talked about publicly. But none of those have materially gone into production and started to consume a meaningful amount of IPUs. The test environment doesn't generate much. Furthermore, in our pipeline tracking, we see the increase year-over-year in the number of pipeline opportunities that are tagged as GenAI-related is multiples more than it was at this time last year.

That's not the CLAIRE side, that's the...

That's the Informatica for GenAI, that's using Informatica for a GenAI workload at the customer. That's where the money is going to be. Now we haven't baked anything really materially into '25 forecast for that because it's so unclear when those are going to turn into big, production-level activities. But the signs are there.

And not to be pedantic, you're saying GenAI, but this is also related to agentic AI. Is that accurate?

For sure. Yes.

Because you have connectors and -- okay.

Yes. I don't want to overstate, I mean we don't do the agentic part per se. So if you're using Salesforce's agentic model, right, we can be the data backbone to get all of the data from all of your stuff that is all over the place, that that GenAI model -- that GenAI agent needs, into the right place, in their place, case in data cloud. That if you're using somebody else's agentic platform or tooling, we can dump it all into Snowflake or BigQuery or whatever that agent needs to get out the data, and we can control its quality and we can provide the access rights so we can make sure there's no PII and that sort of thing.

Okay. Because from a strategic perspective, we get a lot of questions from investors, and we think about it ourselves, like what does the AI data stack look like in the future? Going to be something of the hyperscalers and all being still S3 and blob? I mean, and there'll be some sort of layers? We've been through 20 to 30 years of software iterations here, and middleware, we used a term that isn't really used anymore. So where -- back to that forensics, where are you kind of seeing -- you said you haven't baked any of this into '25, so that's a positive. But what's the likelihood or where are you kind of seeing initial trials where Informatica has given you the confidence that they will be part of that future AI-related data stack?

That confidence is based upon 2 things: complexity and multi-cloud. For those customers that are -- yes, that are able to -- able and willing to stand up an agent that uses a limited and constrained set of data sources, and the data, once it's cleansed and merged and deduplicated and governed, is in one place with a big, high ring fence around it, that's not our business. But that's not the way the enterprise works. Enterprise has data all over the place. It's in all kinds of formats. Those formats are changing every day. Enterprises are becoming more multi-cloud, not less multi-cloud. And for that customer, it's just by definition, there's nobody else that can handle that the way we do. So we feel very good about the durability of our relevance in that even as formats change and stacks may look different, that -- somebody needs to manage that complexity. And on the data side, that's Informatica.

That's a great place to end. Oh, you might have to take -- okay.

Back to that expansion metric, that net new metric, 8%. Can you discuss like that 8% growth of net new customer, who do you guys face in terms of new logo, and then [indiscernible]?

So we don't share that specifically. And win rates are hard because is it versus top-of-funnel, is it mid-funnel, is it when there's an RFP? There's very...

But what are the type of customers?

The type of customers, they're large enterprise and they're dispersed across all the usual categories: financial services, healthcare, pharma, manufacturing, retail, federal government, state and local government. We're not excessively concentrated in any part of the economy; it's really quite broad. The common theme is larger customers with more complexity that values having the best-in-class products on a platform where they don't have to stitch together point products, it's future-proofed against all the conceivable changes in formats, and it works across all your clouds and your on-prem data as a [ switching limit to ] data rather than a captive tool. That's where we win. Great. Thank you.

Michael, thank you very much for your time.

My pleasure. Thanks for having me.