Informatica Inc. (INFA) Earnings Call Transcript & Summary

Okay. Good morning, everyone. I am Mark Murphy, software analyst with JPMorgan. And it is a great pleasure to be here this morning with Eric Brown, who is the CFO of Informatica. Eric, we go back a long way. We've seen quite a few cycles together. And I wanted to thank you for flying across the country to be here at this event at the same time, that Informatica has its conference going on, on the other coast. It means a lot to us, and it's a great way to get this conference started. So thank you very much for being here.

Well, Mark, thanks for hosting us. It's a pleasure to be here with you and your team and the balance of the attendees.

So maybe you could give us just a very brief intro to yourself and the company. Like maybe just kind of a 1-minute version of that for anyone in the audience who might not be familiar.

Sure. So Informatica is a data management company. So in brief, what we do is allow companies to manage, cleanse, secure for quality, et cetera, and usage their data, whether it's on-premise or in the cloud. And we've been around for 25-plus years. We recently IPO-ed. It's a second go-around for us. There are some things that are similar to our business today versus then and some things that have changed quite a bit. But suffice it to say, there's more data and companies now looking to leverage data in the cloud, which is a huge, new long-wave phenomenon.

Yes. So Eric, I've had the pleasure of taking Informatica public twice now, which is quite an interesting experience. When we went back to the first time, the core product was called PowerCenter, right? It was an ETL product. It was basically involved with loading data into a data warehouse or what we call data warehouse creation. So the company grew, it boomed, it went private, it IPO-ed again. Could you walk us through the transformation of Informatica and what you think would be the key differences today?

Sure. So yes, when we went public back about 25 years ago, our run rate was sub-$100 million of revenue. We went public last year in October. Our revenue run rate ARR is $1.3 billion, $1.4 billion. So there's certainly a scale difference. One thing that didn't change between then and now is the concept of us fulfilling the need to be a neutral vendor that would manage data and look at all different sources of data and all different targets of data and treat them in an equivalent manner. And so back kind of generation 1, we're looking to take data from operational systems and putting it into relational databases. It could have been Informix, it could have been Oracle, et cetera, so that the data could be used by analysts and produce some value. So kind of generation 1 analytic applications. Fast forward to today and the addressable market has grown dramatically. We operate in a $40-plus billion TAM. It was measured in single-digit billions back then. So the market's up rough and tough 10x since then. And data has become a part of every business's life. And today, we're seeing digital transformation. So what we did when we went private back in 2015 is we said, yes, we have this great product, PowerCenter, which served the company very well. But we said that's not sufficient going forward. We actually started to build a cloud data management product back in 2015, early 2016. We decided to deprecate our pure perpetual license model, like all of our software revenue back then was perpetual, and build a subscription and a cloud business model. So we basically reinvented the company from within, which is actually pretty rare. So we did this with organic R&D. We built this cloud data management platform to serve the needs of businesses today that are moving now their workloads to public cloud, to the Snowflakes to Azures to GCP, et cetera.

I want to come back and definitely spend a couple of minutes on this newly built cloud product that you have and all the traction there. But just to sort of understand the broader phenomenon that's happening in the world, we have been told that when companies get into a digital transformation, it's going to involve a data transformation, kind of the underpinnings underneath it. So I think we tend to think of -- when people say digital transformation, we tend to think of front office apps and the digital experience. But we -- your customers, have been telling us that they're being advised by the Accentures and the Deloittes of the world that as they're doing that, they need to think about kind of the plumbing underneath that's sort of -- or the foundation underneath that's feeding all those apps. Is that a tangible phenomenon to you in the world? Is that something that you see from your perspective?

Absolutely. What we see more than ever is, for example, the need for customers to stay close to their customers. And so digital outreach, customer lead curation and customer care, all these requires that. You have a really firm grip and definition of what the customer is, their attributes. And so oftentimes, to understand the customer, companies have to pull together data from dozens of different systems. And a lot of the data is resident in a legacy on-premise data warehouse. A lot of the new data is being built by a marketing app in the cloud and then kind of bringing the data together and curating it so that you can actually make intelligent decisions at scale, at speed and at velocity. That's like one great example of what we see in terms of digital transformation, master data management or customer-led initiatives. And so whether it's getting closer to your customer or getting closer to your supplier or getting a better grip on your SKUs, your inventory, how you manage your products through your supply chain, et cetera, all produce and require -- huge amounts of data require data management. And that's what we do. We don't produce the data. We're not on the operational systems side. We don't store the data. We're not a persistence layer. We're the only company of scale that connects all the data. We have 50,000 metadata aware connectors. And so odds are, if there's a legacy data source, 15 years old going to a state of -- state-of-the-art cloud, we know exactly how to get that data from A to B in a cleansed, secure manner.

Okay. So being that kind of connectivity layer and doing the cleansing, somehow that is allowing Informatica to be the #1 leader right now in 4 different Gartner Magic Quadrants. The -- it's data integration tools, integration platform as a service, you've got data quality and then Master Data Management. And so that's something -- we have a lot of companies, they're a leader in one category. Informatica is the #1 leader in 4 of these different categories. So what are the ingredients that you see that are kind of coming together and making that possible?

Well, I think that, again, we've been focused just on data management for the entire history of the company, and I think that's a pretty unique thing. And we've done this with organic R&D as opposed to stitching together a series of large acquisitions. Yes, we've done some M&A, but it's been smaller-scale M&A. And what we've done is we've taken those kind of point products or solutions that you've mentioned, Mark, and we've built them on top of what we call the Intelligent Data Management Cloud platform so that it becomes easier for the customer to leverage one type of product. We were able to, for example, store and curate like all their metadata so they spin out application 1, application 2, 3, 4, et cetera. It's that much easier to deploy. And so it's that intense focus over 25 years on just data management, the organic-led approach and this proprietary library of kind of connectors, which is what's necessary to do this at scale.

Okay. So as we were out there doing the diligence on Informatica, right, which technically I've been doing for a long time. But I mean, the latest round of diligence on Informatica, customers emphasize I think some elements of the company that it is rare, it is unusual, right. It's not what you usually get in the feedback in terms of just what they value, right. So they say Informatica's a long-standing, trusted brand. It has a heritage and history of working with data at scale. And then they'll talk about the neutrality. They'll actually talk about availability of Informatica skill sets in the market, right. It's something that you might not think of all the time. What do you think of that as kind of being the best way to leverage all of that on one end of the spectrum? And then I think at times, we also wonder what is it like for a small startup, I guess, trying to compete against all of that.

I -- think maybe I'll start with the latter question. I think, again, we cover the entire landscape of data management. And so in any of those one categories that you mentioned, Mark, there are like dozens of competitors and they're invariably all small. They have high burn rates and they're probably working really hard to readjust their business model these days. And so they're going to tackle one problem at a time. What's interesting is that we're able to deliver better products, as evidenced by the Gartner Magic Quadrant wins, than any of them. And again, we're the only vendor that can bring all that together in a single platform. And the way that we leverage it, and I think where we benefit our customers the most, is our go-to-market motion, where we have key partners on the cloud hyperscaler side and then the global systems integrator side. So you mentioned Accenture and Deloitte. They're key partners for Informatica. They have thousands upon thousands of people in terms of the organization trained up on Informatica tools and technology. And the benefit to them, to Accenture, for example, they want to go in and pitch not a $1 million or $2 million solution delivery proposal. They want to pitch a comprehensive digital transformation engine, end-to-end $50 million to $100 million covering everything in an enterprise. And by -- almost by default, the Informatica platform is the recommended tool of choice to do that because they know that we deliver, and we cover all the potential use cases. We're not just good for one solution area like a particular startup might be, we can cover the entire landscape. And so there's this really good, natural affinity between us and the global SIs. And then the cloud hyperscaler part, they like working with Informatica because they get paid once a project is live on their platform, and it's consuming storage and compute. And the time from start a project to value realization, i.e., consumption, is generally going to occur much more quickly if the project uses Informatica. We can get the customer from A to B in kind of cloud consumption, which benefits the hyperscaler. So that ecosystem go-to-market is really unique for Informatica relative to the competition.

So you're talking about everything from AWS to Azure to Snowflake, right, when you say hyperscalers.

Correct. I mean our IDMC platform runs on all 3. And so oftentimes, we see most large customers, they're using 2, if not 3 of the 3 major hyperscalers because they're diversifying, they're purchasing the risk, running certain types of apps in 1 cloud versus the other. We come in and say, "Look, we can deliver IDMC on all 3, pick and choose, we'll support them all equally." That comes back to that concept of being neutral and being kind of the Switzerland of data management.

Okay. So now let's go back and talk about the cloud product, which is the very, very high growth vector of the business. You've been very clear when you've described the product. You've said it's not a Band-Aid lift and shift. It's an organic product-led transformation. The customers saw it that way when we were doing the diligence. So can you give us a little look under the hood at what it was that you built there? Maybe just how big of an engineering project was that?

Well, shortly after the privatization, starting in the first half of 2016, we started out with essentially kind of a clean-sheet vision of what a cloud data management product should be. And so we took the PowerCenter product, the on-premise product, basically put that to the side in more of a kind of a sustaining engineering mode and then rapidly moved the vast majority of our resources towards developing IDMC Cloud. It was developed kind of the tenets that you would expect, microservices-based, highly secure, and we also infused AI as a design feature into the platform very early on. And AI is important where -- when you think about a digital transformation initiative, oftentimes, you're looking at 30, 40, 50 different data sets. You're trying to get a handle on it. You're trying to figure out what type of data you have. Is it a customer record? Is it sensitive data that has to be masked, et cetera? And so we built AI into the platform so that it could become smarter. So those were the design tenets. We were able to do this because we had a very profitable business. The PowerCenter product, its installed maintenance base renews it 95% year after year. So you have that kind of cash engine, which is able to finance what turned out to be over $1 billion in organic R&D spend over that 5-year period of time dating back to 2016 up until our IPO.

So $1 billion went largely into the cloud product.

Correct.

Okay. So you have this cloud product. It has been growing right around 40%. What do you see when you look out a little farther? Because I think we've got that guidance basically through the end of this year. Did you have any feel for how long could this cloud product grow over a certain threshold, say, over 30%, over 25%?

Well, our objective -- I wouldn't get into 2023 at this point. But our objective is to drive cloud to $1 billion as quickly as possible and then go past that. What we're doing literally every quarter, we look at our net new bookings or our NRR shift. And more and more of it is going net new to cloud versus self-managed or on-premise licenses. And so there's a very clear preference shift that we're seeing in the customers, and that's combined with our cloud road map, and we've just basically completed that. Every single product that we have is now available on multi-tenant cloud. And our sales team, in connection with our ecosystem partners, is leading with cloud first. And so I would expect that we're -- we'll hit our 40% this year. Beyond that, we'll have to wait and see. It could taper. We start to build on some pretty significant numbers. I mean 40% growth mark gets us to a little under $0.5 billion in pure cloud ARR as of the end of this year. But we firmly believe we can get to $1 billion in cloud and beyond.

Okay. Great to hear that target. Now part of what's happening as you're doing this is you have this new pricing model. You've called it IPUs. The -- one of the comments was that when you look at the new cloud bookings, which I think would be something like 1/3 to 1/2 of the total bookings, that if you drilled in there, more than 40% are coming in on this IPU model. So can you walk us through that? Just at a high level, why was it done? How is it working? What do you think that this would mean for the financial model?

Yes. In Q4 last year, big bookings quarter, about 45%, 46% of our new cloud bookings were denominated in the IPU. And it's an important innovation because rather than having the customer decide, I'm going to use a specific capacity for the next 2 to 3 years of a certain cloud product, we sell them the IPU, which is a -- it's like a fungible scaler. They buy X IPUs per month, commit it over a 2- to 3-year term, and they can draw down any type of admixture of cloud services. If they have to spin up a big project, they need to do bulk load, a lot of DI, that's fine. And then they can kind of taper and remix that. And we allow a little bit of burst capacity there as well for seasonality. And so it's been very, very well received by the marketplace. And since now that we have all the products available in cloud, my expectation is we'll very quickly have more than 50% of our net new bookings in cloud-denominated IPUs. And what we like about them is that they allow for a natural expansion. So we're able to look at the usage on a daily basis since we can kind of predict in advance when the customer is going to exceed their IPU threshold. 3 months in advance, we can give them a heads up. We do an upsell motion. And it's always easier to sell more to an existing customer, particularly when the consumption is kind of scaling naturally with the business. So the IPU, while it's a relatively new innovation for us, about 1 year, 1.5 years old, it's really ramping up in terms of traction. We effectively went from 0% to, call it, 45% of our cloud bookings as of last quarter.

Okay. I like that term fungible scaler. I'll have to remember that one. Is there a fungible scaler on the net dollar retention metric? The -- so I wanted to go into this. You've been trending, I think, 113% to 114% on the retention, but there's this long-term goal of 120%.

Correct.

And so I think one of the questions we'll get from investors is, what is the motion that's kind of getting you there? Like how do you think you would build the bridge from the 113% to the 120%?

Yes. I think in about 6 months, we'll see like the first real kind of 2-year cohort of the IPO. So generally, we do 2-year to your committed deals. And so at that time, we do renewal, upsell. So it will be interesting to see what happens with the IPUs at that point in time. The other thing, though, that drives the NRR down is if in a given quarter, a higher percentage of our net new bookings is from brand new logos that does nothing to our NRR calc. And again, the way we measure ARR is we land a logo, you start the clock in the cohort, like nothing happens in terms of NRR, it's just new bookings. And then we look at that contract 1 year later, and we'd say, "All right, we initially booked it at $100 per year of ACV. It's now at $115." That'd be 150% NRR. If in a given quarter, like I said, you're landing a little more of that new logo business, it does nothing to the NRR metric. And so we've had a slightly higher-than-expected mix of new logo business. And so that's one factor. And then the other is, like I said, about in 6 months or so, we'll see the first kind of big anniversary of that IPU cohort.

Okay. So one of the other, I think, subtle drivers you have in the background, I don't -- we don't think it's a main driver, but it's somewhere kind of on the back burner, is this migration trend, right. Because you do have the more mature base of customers that are running the PowerCenter product on-premise. And I think investors are usually interested in the math, okay. So if they're going to move from there to your modern cloud platform, what happens kind of economically, right, to your revenue stream? And I think part of the question as well, is how -- what kind of builds the confidence for you that they're going to, in fact, stay with you and do that rather than churning off? What is it that -- because that is -- you are seeing quite a bit of success converting them.

It kind of points back to -- this is really the PowerCenter installed base that we're talking about. So these would have been perpetual licenses sold 10 years ago. And those are workloads that customers are running on-prem, using their own data center resources, servers, database administrators, et cetera. So there's an ongoing cost of ownership even though they "paid for the perpetual license." And it's -- customers come to us and say, "Look, we want to modernize, move to the cloud, move that workload off of our equipment systems and personnel and move it to an AWS or Azure, for example. The transforms that have been built up in the on-premise app have taken place, again, over the course of a decade. And these are -- there's thousands of mappings. And so do we worry about a customer migrating from PowerCenters, some other company other than Informatica? The answer is absolutely not because they would have to fully reimplement the project. We have proprietary tools that can go in and look at those mappings on our on-premise product and convert it to our cloud product. We have like one customer that went out and had it bid separately. It was like 3 years, $7 million, no guarantee of completion, and we're able to do it at roughly 1/10 of the time and the cost there. So it is an order-of-magnitude-type difference if someone else comes in and tries to do it. Now as to the economics, what drives the customers to consider this is cost of ownership. It's again, the thing that works really well with cloud in general. You take that expensive, on-premise app, you offloaded, say, to AWS. You can jettison the servers, the equipment, the personnel, save it or repurpose the dollars, et cetera. When we went in -- and as part of the kind of the IPO modeling and road show, we weren't sure a year ago at what rate this would ultimately convert. We assumed the conversion ratio would be 1.3:1, meaning that if a customer tendered $100,000 worth of maintenance, they would pay us $130,000 of cloud. We deliberately chose that number. Again, we weren't sure. It's kind of the breakeven. It's the gross profit margin kind of breakeven point. In other words, $100,000 of maintenance versus $130,000 cloud produces roughly the same net margin to us because you've got an extra 20% plus cloud COGS. And so we weren't sure if we're going to get high or low. We didn't want to have a lot of volatility in terms of the net margin assumptions. So we rolled into the program. Now we're several quarters into it. Life to date, the conversion ratio turns out to be about 2.0, meaning a customer will tender us $100,000 of maintenance and steady state, they'll pay us $200,000 worth of cloud. Now in between, there's work that has to be done. It's not as simple as turning on a switch. We'll provide some credits while we do the work, some pro services credits and some maintenance credits. But going in, if the example is $100,000 maintenance steady state when we're done, we'll end up getting about $200,000 of cloud, which is margin accretive to us.

Okay. So your assumption at the time of IPO was kind of dramatically too conservative, essentially, based on what you've seen so far.

In hindsight, yes. And back then, there was a lot of uncertainty. And so we wanted to choose a number where we weren't going to be too far off in terms of the margin impact either way.

Okay. Is this a pretty safe assumption then? You've got customers. These workloads are running. They're running properly. They're running with good consistency. The mappings don't exist elsewhere. It's just going to be sort of a very default motion that they're going to move to the cloud with you.

Yes. We just don't see there being risk them going to someone else's cloud DI solution.

Okay. So I want to come back and ask you about the AI and machine learning that you've built into the product. You had mentioned that a moment ago. We, of course, made sure to blanket your customers in our diligence process and ask them if they were familiar with it, if they were using it and kind of what it meant to them. And it was something that came up. So one of the customers said, this was the quote. "The biggest differentiator is the ability to have some sort of AI and machine learning that will help understand what metadata is coming through and automatically curate that and put that into a glossary. They've nicknamed that CLAIRE." And so -- and I think that was pretty representative of what we were hearing. Could you explain the basics here? I don't know -- because we have a bit of a nontechnical audience, that's not doing data transformations. Could you explain what is actually happening? Because it felt like this was an actual practical use case rather than, I think, there's a lot of pie in the sky, every company kind of saying they have some machine learning.

Yes. So we start with the definition of metadata. So metadata is just information or data about a raw data set. And so to use like a simple -- let's say, you export your data from your old mainframe system and you'll put it to a -- in a flat file. Think of it as like a giant Excel file. It's got whatever, tens of thousands of rows and hundreds and hundreds of columns, and you see lots of values, but you don't know what the column labels are. You're kind of like guessing it's like, "Hey, I've had this whole data set. I want to move to the cloud. I want to use it to manage my business." And so you're doing this at volume, at speed. And so you can either have like a human analyst who like maybe coded that mainframe system like 15 years ago, look at the data, going and like label it all and like interpret it. Or you could have a metadata aware connector that's familiar with kind of the mainframe, will look at the output, will recognize like a date, a currency field, a character field, et cetera, and not only recognize the fields, but looking at in total, realize what type of data that it's looking at. So that's a metadata aware connector transforming the data. And the way that CLAIRE works is because it's looking not just at one customer's data set. It's looking at multiple customers data set, all anonymized, of course. And so it learns, it becomes smarter, more intelligent about predicting or identifying what that -- what those column labels should be. So as you're taking not just this one mainframe data source as an example, but like dozens and hundreds and melding the data together, you can figure out where you have duplicates. And you can start out with a really good guess of like what that data really is as opposed to humans spending hundreds and thousands of hours interpreting all that data. And then the fact that it's a metadata aware connector, if you swap in a different source or the column order should change in one of the underlying systems, it's going to know that and react to make sure the data that comes out, if a field changes, you're going to end up with a predictable result in your final output set. So that's how you handle data -- interpret metadata at scale, at speed and use kind of AI techniques, machine learning techniques.

Okay. So Eric, I'd be remiss if I didn't sort of ask you to kind of review Q1 a bit. And just what you're thinking about the macro environment. It was a solid quarter. You had 43% cloud ARR growth. What we -- the fact of the matter is we do have -- we have plenty of software companies now talking about deal slippage in Europe, right, that either they've seen some or that the headwinds are mounting. How are you looking at the demand environment? And how are you feeling about maybe the business confidence that customers are projecting toward you?

I think that we had, in terms of Q1, we didn't really see any impact down the stretch. In fact, Europe actually had a pretty good close, slightly better than average, which is kind of counterintuitive, given what was going on there the last several weeks of the quarter. Everything else basically kind of performed as expected. We have done one thing that's pretty, I think, important here is not only have we transformed to a ratable business model. But as we've moved to cloud, we have a larger number of net new transactions and lower ASPs. And so it used to be really dependent upon $4 million, $5 million, $6 million perpetual license deals. That's no longer the case. We make it up in volume. And I think in a time like this, what tends to happen as things tighten up a bit, it's the large -- it's 7-digit plus deals. So you're going to get more scrutiny, more layers of signature, it's going to go to like executive committee, et cetera. And so it gets tougher for like really, really big deals. We don't do really big deals anymore. We're not dependent on those $5 million transactions. So I think that allowed us to get through kind of Q1 without any significant surprises. I wouldn't make any interim comments on this quarter, but I would just state that we're still on that ratable, smaller ASP model here, more of the mix over time continues to move to cloud. And so I think that helps mitigate what I've just mentioned.

Okay. So it's reasonable, if you're me, maybe thinking about this, the large deal environment, there could be a reason to think it will tighten up in the back half or towards the end of the year given the cycle. But you've got a way to kind of manage through that given you don't have as much exposure. Is that a fair way to think about it?

That's a fair way to think about it. And also, we have a good kind of about 50-50 mix of selling to net new logos versus upselling existing installed base.

Okay. Let me ask you one more quick question, then I want to see if we have anything from the audience. Which is how to manage costs in an inflationary environment? Because you've been -- you have CFO -- you've been CFO of Electronic Arts and Tanium and McAfee. You've seen a lot of companies. We haven't had a scaled software industry ever, right, in a period of high inflation, because I think you have to kind of go back to the '80s. And so we get this question a lot, that for a software company, headcount is the major cost, right. And you have wage inflation right now. So how do you think companies maybe will tackle managing through that without taking a margin hit?

Yes. I mean for software companies, your #1 variable expenditure is headcount. Like the first thing I did when I got to Informatica was get a daily headcount report worldwide. Like every day, I know the number of FTEs that we have by function. I think it's really important because you're going to have to adjust. Demand pulls in a little bit, and you'll have to slow your rate of expansion and vice versa. So you have to have like control and telemetry to run your business properly, and it starts with headcount. And then also like really good metrics to understand where your returns are. The second thing is that you have to be -- and this is something you can't just like wake up and say, "Oh, I'm going to like -- I'm going to address this now that things are tightening up." Like how you contract in your business is really important. And so like if your legacy business has been on a lot of all you can eat, installed base, enterprise license agreements, you don't have a chance to really upsell, right. You've kind of saturated installed bases. And so historically, we've been pretty careful to make sure that we can put price increases, CPI, CPI+ in our subscription or maintenance contracts. And so having the ability to increase price a bit is pretty important. And secondly, having a consumption-based pricing model is important as well. I mean the cloud hyperscalers have clearly figured that out. And we're now mimicking that with our IPUs. And so I think it's a combination of those things of basic kind of cost management, focus on headcount returns and having some pricing increase ability that kind of -- you can trace back to your contracts with individual customers.

Okay. Great answer. I want to do a quick check for anything in the audience. Just raise your hand, and we can get you a microphone. Going once, twice. Okay. Let's finish on a quick thought about the customer segmentation. We -- I think one of the critiques or questions that we will hear from investors is they will say, "Well, Informatica, cloud is growing fast, ARR is growing fast. We're not sure that the customer count is growing as rapidly." And so we look at it, it's a who's who, right, of pretty much every big company in the world. But that piece of it is not growing at light speed. How would you respond to that? And I think one of the questions we get is, well, would Informatica ever consider doing what MuleSoft did, for example. I mean, there's a long list of companies that have done this where you have -- they have a studio product, right. It's kind of aimed at being a little more mass market. And then you can kind of potentially open up more of the mid-market. Is that anything that's ever crossed your mind?

The short answer is yes. And we've referred to it a couple of times before. But we're working on more of a freemium, low-cost basic data ingestion product that we're going to be bringing to market here in the second half. So the idea is that there's a lot of department-level use cases that start with a very simple scenario. I want to move one data set off of this on-premise system to a cloud data warehouse. So it's department level, it's relatively simple. It's nothing near any of the other scenarios that I described. And so for us, like we clearly have the product capability. We have to engineer down our product to account for this. And we'll have to put in place a slightly different digital-led selling motion to address it. But in terms of customer count, just to kind of -- so the group understands where we are. We have -- we're very G2000-centric, but we have, as of the end of Q1, 3,678 active subscription customers. So just under 3,700, with an average ASP or average subscription ARR of $231,000. That grew about 19%, 20% year-over-year. And so we have this really large, broad, diversified installed base. So it's across every geo, every vertical and growing at a really healthy clip on that large base year-over-year. So we already have breadth of participation. So when people talk about customer adds, again, unlike a lot of IPO -- recently IPO-ed companies who had 300, 400 customers and they're growing 90% year-over-year, well we added about 400 customers year-over-year. It's just that we're starting with much larger numbers given the history there. So I just want to make sure everyone understands we are adding hundreds and hundreds of net new sub customers year-over-year. It's just that the denominator is a lot larger, so that percent doesn't seem quite as high compared to others.

Yes. Okay. We have a question now. I think we're basically out of time. Do you want to try to do -- you want to do it in 1 sentence? Or you want to just kind of go off stage here? Let's wrap there. You guys can chat in the hallway. Eric, I can't thank you enough, and we're looking forward to hearing about all the announcements that come out of Informatica World this week. Eric, thank you for being with us.

Thank you, Mark. Thanks, everyone.