Informatica Inc. (INFA) Earnings Call Transcript & Summary

[Presentation] Good morning, and welcome to Informatica and the modern cloud data architecture webinar. We are coming to you live from headquarters in beautiful Redwood City, California. I'm Victoria Hyde-Dunn. I lead Informatica's Investor Relations efforts. Before we get started, a friendly reminder that we will be discussing our technology capabilities. Please see our most recent SEC files for more information. This webcast will be made available for replay, and a copy of the presentation deck will be posted subsequently on the Investor Relations website. We have a full agenda, and we'll be discussing a wide array of topics. We will also have time at the end to take live Q&A. Please send an e-mail to [email protected]. To get things started, it is my pleasure to welcome Informatica's Chief Financial Officer, Mike McLaughlin.

Thank you, Victoria, and welcome to everyone joining us live online and here in our Redwood City office. We're going to do something a little bit different today than what we did a year ago at our Investor Day. Instead of a 30,000-foot comprehensive view of our business, including product, go-to-market and financials, we're just going to focus on tech and product and Informatica's place in the modern cloud data architecture. Amit, Victoria and I have had hundreds of conversations over the last year with investors and analysts. And one thing that's become really clear to us, it's not easy to understand exactly what our software does. And therefore, where do we compete, where do we fit in the stack and why do we win? So we're going to tackle that head on today, and most of the discussion's going to be led by my partners, Pratik and Brett. One quick programming note before we get started. We are not going to give any forward-looking financial information in this presentation, and you should not infer anything about our Q4 financial performance from what we discuss today. So let me get started with a brief history lesson. Informatica is synonymous with data management. We've been at it for 30 years, and we've been a leader in that whole time. So where did data management come from? And how did we fit in that pattern back in the on-prem era? To answer that question, I ask you to cast your mind back to 1993, the year Informatica was founded. Bill Clinton was being inaugurated as our 42nd President, Ford Taurus was the best-selling car in the United States, and agents Mulder and Scully were introducing us to the paranormal with The X-Files. In IT, we were in the midst of a really important architectural revolution. Distributed computing was displacing mainframes, disk storage was becoming faster, cheaper and more reliable, and enterprise software is becoming much more powerful, solving more use cases and generating massive amounts of data. So what did that translate into in terms of a data architecture in the on-prem world? Started, of course, with the data sources, the applications that were generating data, ERP, CRM, et cetera. And you had business users seeking to get value from that data for reporting analytics and business intelligence. The problem was it wasn't easy for those business users to get out the data. The connections themselves were hard. The data in the applications was messy, and you needed data from multiple sources. So they needed to do mergers, joins, deduplications and format fixes. That's where the data warehouse came in. The data warehouse is fundamentally an optimized storage layer combined with an optimized query engine together in one package that allows enterprise users to query vast amounts of data quickly and efficiently. We'll talk a lot more about that as we get into the technicals later on. But to utilize that data warehouse, you had to get the data into it. You had to extract it from the sources, you had to transform it as necessary, and you had to load it reliably into the on-prem data warehouse. That's where Informatica fit in, and that's where we built our business, and that's the business we led for 20 years. One aspect of this architecture that I want to point out, and I will get back to when we talk about the cloud architecture, is that in this on-prem environment, the customer paid for and provided the storage and the compute. It were their storage arrays and their servers. Informatica provided the on-prem software that enabled the data management. This was a great business for Informatica. We grew it from a standing start to over $1 billion in 2015. And we did so amidst strong competition. We weren't the only ones that could do data management, but our innovation and the strength of our product enabled us to be a leader for decades. But by 2015, the world had changed a lot. The Eurozone was in resurgence after their debt crisis. The Toyota Camry was the best-selling car, and Hamilton debuted on Broadway. And in the world of IT, the world was going to the cloud. SaaS applications enabled by cloud service providers were revolutionizing the way we do business, and they were generating exponentially greater amounts of data. Unfortunately, at that time, Informatica was still primarily an on-prem company. So what did we do? We took the company private in 2015, and we began building a cloud-native, true platform for cloud data management using a clean sheet of paper, cutting-edge architecture and building blocks. Born in the cloud, as they say. And we spent over $1 billion in R&D, in the process. The result was Informatica's Intelligent Data Management Cloud, the IDMC. The IDMC as we sell it today consists of the best data management products category by category, delivered on the industry's only cloud data management platform serving the multi-vendor, multi-cloud and hybrid needs of the modern enterprise. Don't just take our word for that, ask Gartner, Forrester, IDC, other analysts who consistently, year after year, rank us as the leader in cloud data management. So let's go back to our simplistic data architecture diagram and compare the cloud data world to the on-prem data world. Well, these days, most of the data is being generated by cloud data sources. And you still have business users that want value from that data, but they're using cloud tools for the same reporting analytics and business intelligence themes, but increasingly for data science and AI. And in the middle, it's the cloud data warehouse, data lake or data lakehouse that makes it all possible. But you still have to extract the data, transform it as needed and load it reliably into the cloud data warehouse to make it useful for business users. And like in the on-prem world, that is the value that Informatica provides today. And also like in the on-prem world, it is the customer that provides and pays for the compute and storage for cloud data management for cloud data workloads. Informatica provides the cloud software that enables the data management. Now I want to make one more key point before I turn it over to my technical colleagues. In technology, you will from time to time hear convergence narratives. Analysts or vendors will forecast that a new way of doing things, whether it's at new standard or a new architecture is going to drive a simplification of the complexity that exists today, a convergence on a single vendor or a single way of doing things. And in fact, you're hearing that narrative in some parts of the data world today. In our view, that is not the way enterprise data works in the cloud. It is, in fact, not converging. It is becoming more complex, and enterprise customers are using more clouds, not fewer. We're not the only people who think that. Just yesterday, IDC released their 2024 Office of the CDO survey, and in that survey, among other things, those CDOs reported that they had an average of 18 data repositories in their environment. That's multiple data warehouses, multiple data lakes, multiple operational reporting stores. Forrester in September released their global enterprise cloud decision-maker survey. They asked 700 of these folks, "What impact do you expect an increased public cloud budget to have on the number of cloud providers that your organization uses?" 64% of the responders says is going to increase the number of cloud providers. Only 6%, 1 in 20, said that it's going to decrease the number of providers. This complexity and this growing multi-clouded-ness of the cloud data environment directly drives the demand for and the need for Informatica's IDMC, which is a collection of the best products on the industry's only cloud data management platform, acting as the Switzerland of data. And it is the key part of what has enabled us to grow our cloud subscription ARR to over $748 million as of Q3 2024. So enough of the layperson's history lesson, I'm going to now turn it over to my colleague, Pratik, who's going to talk about data integration and engineering in the cloud. Pratik?

Thanks, Mike.

Here you go.

Thanks, Mike. Hello, everyone. I'm Pratik Parekh, Senior Vice President and General Manager for integration business at Informatica. Been with the company for 12 years. I'm excited to be here. Today, I'm going to cover the technologies fueling the modern data architecture and role Informatica data integration and engineering plays. We'll also look at some of the emerging trends that are and will play a significant role in this architecture. Let's go back in time and understand why a new software stack was needed for analytics in the first place. Transaction systems are highly optimized for transactional workloads. It involves complex schema with data distributed across number of tables, thousands in an enterprise system, such as ERP. This slide shows a representative data model, also called an entity relationship model for orders and deliveries in an ERP system. Here is what a supply chain executive at this manufacturing organization would like to see. He wants to know how they are doing in vehicle shipments, how many vehicles were delivered on time, by type and by region. The top table here shows the raw data as it is in the transactional system. As you can see, it doesn't do much. It has codes and values that do not mean much unless you know what that stands for. The table below is a fully formed for report and visualization. This is the data ready for analytics. The raw data needs to go through preparation to produce the data ready for these analytics. This is called transformation of data. Example transformation includes joins, cleans, conversion, standardization, application of business rules and formulas. Once transformed, data then map to the analytical model. This is just one example. Real-world situations are significantly more complex, requiring specialized skills and technology. Now this use case led to a special class of database that we call data warehouse. This is fine-tuned for analytical workloads also called OLAP or online analytical processing engines. Data Warehouse fundamentally has 3 components: a query interface layer, a data processing layer that executes that query and a data storage layer that organizes the data against the analytical model defined and is highly optimized. Data warehouse differentiated on query processing amount of data it can store and performance and scale. Typically, it existed as one logical unit. Vendors built highly optimized physical appliance building the required compute and storage. Here are some of the vendors who specializes in the market that you will recognize. Analytics vendors overlapped in some functionality, especially the query layer and included processing and storage, but its primary role was to build and serve the reports and visualization to the large user base in the organization. Now the problem that was not solved was data. How do you bring all the changing transactional data under all these lock systems in different formats and different latencies on time for analytics? That gave rise to integration pattern called ETL, which Informatica pioneered. ETL later expanded to data integration that addressed a broad range of data engage and processing patents. ETL involves extraction of data, transformation of data and loading it into the data warehouse. Once loaded, which is typically considered a bronze zone, goes through further refinement and ultimately is ready for analytics. This gold data is supposed to be gold for the enterprise. Enterprise relied on this for critical decisions, and Informatica established itself as the leader of providing this capability. Note Informatica's value was data pipeline management, and did not offer any compute or storage. The software ran on commodity hardware and operating systems. We further expanded our offering beyond data integration to do comprehensive data management, data quality, data catalog and master data management were added to the stack. Now however, we were not alone. It was a competitive landscape. Enterprise data warehouse, database providers, platform providers and analytics vendors all offered some form of data integration. For example, IBM had DataStage then. Oracle had Oracle Data Integrator. SAP had their own version. Even likes of Teradata, who specialize on the data warehouse appliance, had sophisticated SQL-based interface to do data transformation. Informatica was appreciated and acknowledged for the best-of-breed data integration and value it offered. Informatica's value proposition was end-to-end data pipeline management. That included these 5 key capabilities that they did better than anybody else: provided a high productivity environment; had the best workload optimization; provided connectivity to all critical data systems and was portable across platform; future proofing of the business logic; and finally provided the enterprise-grade operationalization. We'll look into each of this later in this presentation. From the economic standpoint, Informatica was not competing for the spend on data warehouse or analytics layer. Customers spend roughly 15% to 20% over and above the analytics infrastructure for data integration. The market was also defined that way, where data integration was a separate market segment and a category and growing rapidly. This led to Informatica's growth to $1 billion on premises by 2015 that Mike talked about earlier. So now let's fast forward to the modern era. We are now looking at years 2010 and beyond. Emergence of cloud was a defining milestone in computing. Conceptually, it enabled infinite compute and infinite storage at a fraction of the cost of the on-premises infrastructure. This led to acceleration of digital transformation. Digital transformation involved optimizing business processes, new digital initiatives, new and modernizing applications, many of which were Software-as-a-Service but also new application development. New data stores and data structures emerged to support the different type of applications, cloud and mobile. Data was generated in more latencies than that needed to be handled for integration and analytics. The end result was organizations were able to bring agility to their business, modernizing business processes and digitizing their assets, but it also created a huge problem of data fragmentation and explosion. This became a massive problem as the enterprise looked at providing analytics and insights to their users on all these newly generated business processes and applications. Reports suggest that enterprise today have over 1,000-plus applications. It is estimated that over 90% of the data is created in the last 2 years and is growing at over 40% year-over-year. With the economics of cloud, analytics stack also went through a big change. Biggest was compute and storage were decoupled. Both can now scale horizontally, independently of the workloads. Optimized storage was split further into what we now call data lakes. Data lake promise to be a single destination for all enterprise data built upon cheap cloud storage. Data processing layers were optimized to take advantage of these changes. On the use case, enterprise now not only wanted to see the historical data, they also wanted to model this data to see the future. That gave rise to data science as a discipline. Again, like in case of previous era, ingesting all this data that is under the transactional system and on-premises was a challenge. The more things change, the more they stay the same. Informatica saw this trend early. We introduced cloud data integration, a cloud-native offering optimized for cloud workloads. We have continued to evolve that rapidly as use cases and architectures have evolved. Gartner believes, and we agree, that data integration is fundamental for a modern data architecture, such as data fabric and data mesh. Organizations need a comprehensive capabilities for operational, analytical and AI use cases. Here are all the capabilities Gartner identifies as critical capabilities for data integration, supporting the core use cases identified here in the blue boxes. Today, Informatica is most comprehensive and complete data integration solution supporting all data integration and engineering patterns, including operational analytics, real-time analytics, data science and AI. Solution is designed for power users as well as line of business, and it is best-in-class with modern cloud-native cloud-first microservices-based architecture. Now just like in the on-premises world, we have heavy competition. Only difference is some names are different. Hyperscalers, cloud data warehouse vendors and point vendors make up the competitive landscape for data integration and data engineering. Informatica has continued to retain the leadership 19 years running. While we are now on cloud-native modern stack and handle the requirements of modern architectures and use cases, the core value proposition that we saw earlier has not changed. Our unique differentiation is still end-to-end data pipeline management that we saw earlier. Let's look at the core capabilities in more details. High productivity development. Informatica, right from on-premises day, has supported a model-driven metadata-aware development paradigm. This involves no-code design experience, with out-of-box prebuilt connectors and integration processes. It offers 10 to 20x advantage over hand coding solution, not to mention hand coding is difficult and error-prone and requires high degrees of skills. Hand coding also binds the core to the underlying platform, to the language, to the engine, to the data warehouse, so it provides low profitability, if any. Next core technology differentiator is the workload optimizer. The graph, also known as the mapping, which was created in that designer, is translated and optimized for the runtime and for the workload. This execution can happen anywhere per customer's choice: on premises, in elastic runtime, in serverless environment or in data warehouse system in itself. Workload optimizer offers unparalleled flexibility as every customer and every project has certain criteria of security, cost, performance and scale. We recently published an IEEE paper that demonstrate 7.5x price performance with Informatica optimizer, helping customers to select the best deployment configuration to run their workloads. Next, let's look at data and platform portability. This is enabled by out-of-box connectors and platform-agnostic design time and runtime. Informatica offers over 50,000-plus metadata-aware connections. This gave us the status of Switzerland of data. Cloud data integration can connect to 100 years old system such as mainframe and most latest data structure and formats such as Iceberg, Delta, vector databases and large language models. Connectivity handles all latencies: batch, real-time and streaming; and all data formats: structured, semi-structured and unstructured. Next, I'll cover the key differentiation of future-proofing of business logic. Integrations that we saw earlier has critical business logic embedded, and often this routine remains the same for decades, with minor incremental changes. Future-proofing this means if a customer moves from one cloud platform to another, they do not have to rewrite their integrations. If they move from one data warehouse to other, they do not have to rewrite. If they change one source or target system, example move from Oracle to Salesforce, they do not have to rewrite their business logic. If they change from one technology stack to the other, they can carry forward the business logic. A great example was the big data world and Hadoop. Hadoop World brought a completely new stack with different runtime and technology such as MapReduce and Spark. This tech evolved every 6 months, making it extremely difficult for customers to standardize on. Informatica was able to help customers deploy their business logic on Hadoop platforms without rewriting, and when technology died, well almost, we were able to bring that business logic to the new paradigm, such as cloud data warehouse and now lakehouses. And finally, the enterprise-grade operationalization. This involves scale, availability, reliability, management and security needed for mission-critical use cases in the most demanding enterprise workloads. Today, we process over 101 trillion transactions a month. The underlying architecture is horizontally scalable. Service is available globally and is more secure, providing the highest level of SLA. CDI supports full data life cycle management, with in-built CI/CD, also known as continuous integration and continuous deployment or data operations, financial operations and workload management. This core capability has truly differentiated Informatica data integration offering and why we have maintained our leadership for this long. Let's look at a customer story on what that means. Paycor is one of the hundreds of customers who leverages Informatica cloud data integration, but they did not start with Informatica for their cloud journey. They used a competing solution from one of the hyperscalers. Not only they were successful with the use case, in this case, empowering organizations for faster and better decision-making, they were -- they realize all the benefits that I talked about. They were able to achieve that with reduced cost and better performance. We are talking about 5x faster performance at 50% of the cost. Now our internal benchmarks also showed similar results. We observed 4.5x performance at 42% less cost compared to one of our top competitors. The benchmark was done against all the modern data integration patterns shown on the right of the screen. On the economic front, as we saw earlier, cloud reduced the unit cost of compute and storage. In fact, it continues to fall. That has had impact on the optimized storage component of the data warehouse, and also showing up on the compute tier. However, Informatica is able to maintain our share because of the value we are offering. Also, as we saw earlier, Informatica is not dependent on compute and storage. The slide depicts like for workloads and do not factor the growth of data and new use cases. which counters the decline of the price in compute and storage. That benefits all parties: hyperscalers, CDW vendors and Informatica. Now let's look at some of the emerging technology trends and how it impacts the world of analytics and Informatica. We look at the 3 key trends: open table formats, zero copy, zero ETL and finally, AI. Let's start with the open table formats. Technology is continuously evolving. Right from the early days of analytics, computing world is trying to identify ways to standardize data such that it enables access, reach and improve usage. It got better with every attempt. XML and JSON were largely used for transactional data. There were some analytical use cases there, but it did not become mainstream. Avro and Parquet were a big leap forward on transactional system and for analytical system. It provided ways to describe data, including the metadata, that made it easy to understand. Iceberg and Delta is built on top of this that now allows [indiscernible] operations on data. Basically, in short, it allows data to be retrieved and return like one would do in a database. It is also called [ asset ] properties where transaction controls are maintained. This is great for the industry as it offers a common format to store and manage data that can be understood by different processing engines. However, note that there are multiple of these standards and more evolving, which is given. So let's go back to our modern data warehousing view. We talked about decoupling of compute and storage. What open table does to this view is now you can query directly on top of this open table format, minimizing or eliminating the need for the optimized storage, further reducing the cost for -- cost of the architecture. Given open table is an extension to data lake, you can directly operate on very large datasets, provided query engines can be optimized to take advantage of it. Hence, we see a big push by hyperscaler and pure-play CDW vendors to support this new format. While storage layer is commoditized further, they believe the technology will enable new analytical use cases. In case of Informatica, this is net positive. As customers look to store data in open table, Informatica will continue to provide the value of pipeline management we discussed earlier. Data still needs to be abstracted and transformed to new data formats. We support all the formats that I showed in the previous slide: XML, JSON, Avro, Parquet, ORC, Iceberg, Delta, and also hundreds of industry-specific formats like EDIFACT, EDI, HL7, that have evolved in the last 2 to 3 decades. From an economic point of view, more formats and more data is good for Informatica. Now let's look at the zero copy, zero ETL. Again, back to the baseline of modern data warehouse. Basically, what zero copy or zero ETL suggests is an analytics directly on transactional data and in theory, you should be able to do analytics directly on the transactional system. As the visual shows, you are querying the data directly on the transactional system. Many vendors have claimed to support this pattern. Underlying architectures are different and in most cases, there is a copy involved, which is a paradox to reclaim. Basically, what is involved is data is made available in open table formats we saw earlier. And because data processing engines can now work with this format, you should be able to directly query it. For a simple datasets, like example here, it does a fairly good job. We call that ad hoc data lookup or query. In this example, what you are seeing is a marketing analyst is collecting a lead from a conference that she just attended. The analyst wants to know how many people attended the conference, how many customers visited the booth, how many opt-ins happened. And those are easy questions that, that particular dataset can easily provide. But what happens when the data involves, how did that conference participation impacted the pipeline in an organization? How many opportunities are created? How is it -- what was the campaign performance? And what is the eventual impact on the sales forecast? And as you can see, these questions are much harder. Remember, transactional data is never ready for analytics. So it needs to go through the same transformation we saw earlier in the presentation, and that results in the deeper analytical queries that I just post. Data integration and engineering is required handling -- handled by the data center of excellence in the organization. So in summary, zero copy, zero ETL is good for ad hoc queries, but it does not impact Informatica, which is used for enterprise analytics workloads. Now let's look at the mega trend that is changing and affecting everything: AI and generative AI. This topic is much broader, needing a separate deep dive, but I'll try to provide a quick overview. At Informatica, we have 2 initiatives: AI for data management and data management for AI. I'll cover the topic in the context of data integration. Let's start with AI for data integration. This is built on years of AI and ML innovation we have done at the metadata intelligence layer of our stack. As you are aware, we call this capabilities CLAIRE. CLAIRE enables us to bring best-in-class generative AI-enabled NLP based services. For data integration, we have 2 primary offerings: DI copilot or data integration copilot that is releasing in Q1 of 2025, and CLAIRE GPT that has been available for over a year. Both allows users to explore the data in the organization, prepare the data, build data integration pipelines and deploy that just with the natural language prompts. We believe this will foster democratization of data and thereby improve decision-making in the organization. For data engineers, it means more productivity and solving more advanced use cases. Underneath this, Informatica has built for-purpose AI and ML models. These are fine-tuned and trained for data integration and data management. Next, we will look into data integration for AI and GenAI. This topic covers how Informatica helps the organization to leverage GenAI and build AI applications. After all, AI is as good as data there is. Data integration for AI enables enterprise to implement their AI pipelines, also known as AI engineering. AI engineering involves 2 critical steps. One, extracting and loading large volumes of data, structured data as well as unstructured data, into a special type of data store, such as vector or graph databases. In the process, you apply special transformations called chunking and embedding that basically generates the numerical tokens or references called vectors that describes that chunk of data. When an AI query prompt is issued directly or via the AI application, it goes through another pipeline that is called RAG or retrieval augmentation and generation. Simply put, this generates the AI response by interacting with large or small language models and blending that data with the enterprise data or vector tokens that were created in the first step. While there are more steps involved in any AI implementation, these 2 are fundamental to AI implementation in an enterprise. Informatica brings the same value proposition we discussed earlier: enable data engineers to now do AI engineering. It is no surprise that data leaders consider supporting AI initiatives in the organization as the top strategic priority. And next on the list is good data for analytics. This is from just-published CDO survey results from IDC. We are adding new capabilities rapidly across all of IDMC to handle unstructured data, support the major vector databases and large language models and enable integration of GenAI applications, GenAI capabilities into the applications. The investment customers [ are made ] to build modern data architecture can seamlessly be extended to support AI initiatives. Today, I only covered cloud data integration and how it supports the modern data architecture. Modern data architecture also needs to have strong governance foundation and support broader data management use cases, including master data management, data marketplaces, application integration and data and process orchestration. Informatica offers a full integrated platform that all of the -- with all of these capabilities to enable next-generation architecture and support the full life cycle of data. I covered data integration and engineering along with related components, including connectivity and CLAIRE on this slide. I'll now hand the mic over to Brett Roscoe to cover data catalog. Thank you.

All right. Thank you, Pratik.

Thank you.

All right. Well, hello, everybody. My name is Brett Roscoe. I have the honor here at Informatica to be the general manager of our data governance and cloud operations here, and I have been here for about 4 years. And I have the honor today of really sharing some insights into the Informatica data governance and privacy portfolio and how that solution is having a great impact and value into this market space. So let's set some foundations, because one of the things I do want to make sure we talk about is the fact that organization's data environments continue to become more complex. We see data coming from on-prem, from the cloud, hybrid, from SaaS applications, more and more diverse data, more and more volume of data coming at customers. And a great example of this is one of our large telco customers has over 20,000 databases across on-prem and cloud. They have 3 data lakes across AWS, Azure and Google, and they have petabytes worth of data that they're managing. Now you might say, well, that's one example. Well, we recently did a survey with our customers, and it's reinforced, right? The survey illustrates that over 40% had an average of over 1,000 data sources, and nearly 80% of said those data sources were continuing to increase and the complexity was continuing. Mike alluded to this in some of his opening presentation. Now the data warehouse and data lake vendors, including the cloud service providers, continue to build out their offerings. They continue to attract data scientists and data analysts to their platforms. And several of these vendors have started ventures and building things, even open-source projects, to build catalogs, to build governance capabilities and other complementary tools to simplify the experience of using their data warehouse and data lake environments. And this is great value. I mean it's always good to have contextual information of the data in your environment, helps you manage that, helps your customers manage that. But we know that they're very much focused on the use cases within their platforms. And so Informatica is no stranger to partnering with these ecosystem partners, Pratik talked a lot about that, with the cloud service providers. This value of us being vendor-neutral, of being the Switzerland of data, as Pratik mentioned, brings great value to them and to our customers and obviously to our business. And I want to talk about 3 areas where we add tremendous value on top of this, and I'll elaborate on each of them. The first one is really comprehensive coverage, covering the data state from end to end. The second one is about how we drive multiple personas to use our products together. How do we have projects with different people in the company working together on the same data? And the third one is the ability to provide augmented data management tools driven by AI, and accelerating business outcomes for our customers. So let's dive into each of those. The first one is about comprehensive coverage. Again, this vendor-independent solution, this Switzerland of data which allows us to work across vendors and with our partners. Now if we go back to the local catalog discussion for a second, they have obviously been working on building some interoperability among them, right? Because it's difficult for somebody like a Snowflake or a Databricks to offer catalog but not offer connectivity to something like S3 or an ADLS where much and much of their customers' primary data is. They want to connect to that to make sure they can pull data into their data warehouses and lakes from those sources. But again, the primary goal is to simplify the experience for their focused use cases. Now their business model is to win in the data warehouse and data lake market, which by definition, kind of puts them at odds with each other, making it very difficult to provide connectivity or end-to-end, across the entire ecosystem, support for all of those data sources and data targets. Now to complicate the issue, there continues to be thousands of production data sources on-prem as well as SaaS-based applications that are continuing to produce different kinds of data. There's even unique and kind of tailored databases for specific verticals, like think health care or maybe some financial services. And now most of these born on-prem data sources can potentially be hosted in the cloud, although the cloud service providers may or may not be aware of them because they're probably just an infrastructure play for them. So minus an enterprise global catalog, this leads to multiple tools to cover this environment from end to end. Mark Beyer is one of the leading experts at Gartner who spends much of his time discussing and analyzing the importance of metadata. Mark talks about this issue of multiple metadata stacks, the utilizing multiple stacks perpetuate silos. It perpetuates the isolation of data management within the enterprise. And you can imagine the kind of problems that can cause. Informatica, with our data governance and privacy portfolio within the Intelligent Data Management Cloud, provides the most comprehensive view and control of the entire data state. Informatica covers a broad set of cloud, on-prem, hybrid data sources and the ability to extend those capabilities. We can leverage SDKs with our partners and with our customers to even build custom scanners to get rich metadata from unique databases. And we work with all ecosystems, all cloud service providers, to provide the most federated data environments a centralized way to control, govern and provide privacy and security. Even IDC agrees, right? They believe a comprehensive multi-cloud and hybrid approach to data management is the right answer for most enterprises. So our mission is to bring comprehensive data coverage across the entire customer landscape. We support, as I said, all major ecosystems. We even connect and support to a lot of these third-party or local catalogs, allows us to pull, extract metadata through partnership into our catalog to provide a catalog of catalog, if you will, a way to have a localized catalog supporting a federated environment with centralized views. Now the second area I want to talk about is the richness of metadata, right? How covering breadth and depth of sources is not enough, but the amount of metadata, the richness of your metadata is equally important. Now let me start with a little bit of how our customers use our products. Our modern data governance portfolio is made up of 4 tightly integrated but persona-focused solutions. Our data governance and catalog service provides simplified and consolidated views of our catalog and governance systems. It enables data profiling and lineage. It provides both technical and business views of that metadata, proactive alerts, integrated tools to discover, to provide glossaries and definitions, classifications, and it's all to derive this intelligent metadata. Our data marketplace is our storefront. It's where data consumers, think somebody in a line of business, maybe somebody in marketing or HR, is able to go search, discover, explore and even procure data products or data assets for their analytics and AI projects. Also, centralized teams are implementing more DevOps and FinOps, right? We have evolved to provide more visibility and proactive quality remediation across the data state. So our data quality and observability service allows customers to know when and where their data is being used. It provides tools to optimize around cost or performance, the ability to proactively improve quality or identify anomalies across the entire organization. And lastly, our data access management lets you secure sensitive information while still building trust and drive data sharing across the organization. It's a global data access management solution that leverages metadata to set policies, and those policies can be set all the way down to a data asset or even a row or column level within datasets. And it allows you to set access by location, by role, by data attribute or any attribute, you can imagine, set in a policy with metadata. All right. With that, now how do you power all that, right? Powering a scaled, complex multiple set of products requires rich metadata. And we leverage a single common multi-tenant cloud-based repository collecting metadata from all of our customers and all of their data sources and all of their operations. This allows our customers to utilize this portfolio of tools that have -- that allows each of these personas basically to have a different lens into the same metadata, now all of that with security and privacy and controls built in. But to support a single metadata repository, you need rich, high-quality metadata to better understand the data within that environment and expand the use cases. Now we've all heard that metadata is data about data. But I'll tell you not all metadata is the same. The richness of metadata can help enterprises better understand how to manage and utilize our data assets, right? It provides amazing amount of intelligence on the data, its history, its sensitivity, it's quality, it's purpose, any classifications it needs. It's locations. It's sovereignty. All of those things can be extracted to make sure rich metadata can fuel projects across the organization. Now because we want to have multiple personas and teams using metadata, we know that using disparate, separate multiple stacks can lead to multiple metadata depositories. Now let's talk about the problem of that. Because as soon as you have multiple metadata repositories, by definition, they're going to get out of sync. And at that point, you start to wonder, well, what's the one single source of truth? Again, we rely on a cloud-native microservice multi-tenant-based single metadata repository that serves all of our tools and all of our services on the Intelligent Data Management Cloud. Now a great example of how we create value from that is auto-cataloging. So every time -- Pratik talked about the integration business. So every time one of our customers does an operation in the integration side, we automatically collect metadata to provide almost a real-time view for our customers on the changes to that data state. So great value in terms of having all those people work together. Now let me provide a quick example. This is one example of lineage on a single data element that's being used to populate a BI report, in this case, customer ID. As you can see, capturing the right level of metadata is a comprehensive task spanning many sources, many data transformations and operations. Now any lineage that's not captured through this scanning, through this deep scanning process has to be filled in with manual labor. And even then it's probably going to have gaps, which is just going to lead to errors down the line, right, as that data moves into these analytics and AI projects. So data analyst for us, data analysts, data stewards, can catalog and classify data. governance and security people can drive access and compliance policies. Data consumers can search, discover and procure data through the data marketplace. Additionally, we provide collaboration. We have communication tools. We have rating systems for consumers to rate data assets. We have the ability to chat or request additional assets if they don't have access to that already. And data owners or data stewards can, through our data access management, can automatically grant access. And that takes the time of what typically took manual labor of e-mail or even going and talking to somebody, which would slow the process down, this provides a very fast way to get data into the right hands of the right people very quickly. Now let's talk about this third element. To enable this level of scale, this level of metadata to be used across the organization, we believe an AI approach must be taken, one that provides AI to simplify, automate, recommend task, provide insights that only in a non-AI environment required an enormous amount of human capital. Now we believe that an AI architecture that collects this extensive metadata will enable more teams on more projects with more capabilities. After all, data governance and catalog has evolved far beyond its original roots of really providing regulatory and corporate compliance. All these things of data sharing and observability continue to fuel the additional use cases for governance and catalog. Now our data catalog and metadata repository is the fuel of the engine behind CLAIRE. So Pratik talked about CLAIRE. He talked about the different things we do on CLAIRE. He talked about some new things coming on copilot and integration. But today, CLAIRE copilot is our end product experience. It provides AI and generative AI to drive automation, recommendations, all kinds of tasks in auto glossary reassignments, auto classifications, things that typically take hours of manual labor. With CLAIRE GPT, new sets of customers with less technical skills can drive data management tasks through a natural language query interface. By simply asking questions, you can now search, explore, discover, build pipelines, improve quality through CLAIRE GPT, a great way to bring in more users into the platform. Now our customers value each of these. They use Informatica solution for an increasing number of use cases and projects across their enterprise. In fact, I will tell you the evolution of governance and catalog continue to evolve our investments. It continues to keep us on our toes to drive new capabilities so that we can expand their use cases and the things that they're doing with data across a broader set of users in their organization. Now Pratik showed a slide from this IDC research just a little bit earlier, but there's another data point that I thought was worth talking about. I didn't talk much about the impact of AI and generative AI on governance, but it's been big, right? Top of mind challenge for enterprise is making sure they have governance and privacy around the data being used for these projects. Because we've seen projects go awry. We've seen in social media, we've seen in mainstream media, some of the outcomes when you run too fast with a generative AI project that's not using unbiased, relevant and highly understood data. Informatica is uniquely positioned to work with our customers and provide end-to-end solutions spanning the enterprise for these AI and GenAI projects. Let me wrap up by just saying, look, all of this is delivered in the Intelligent Data Management Cloud. So imagine all of the integration in our governance portfolio that I just discussed, integrated even further with our MDM, our 360 applications, our integration business. The goal is to design a solution that provides the broadest end-to-end data management experience across a multitude of sources, applications across on-prem, cloud, multi-cloud, multi-hybrid architectures. And the goal is to deliver trusted, secure and high-quality data from all of those data sources to all of the data consumers: your data scientists, your data analytics team, your data consumer. And with that, I'm going to hand it back over to Mike, who's going to wrap us up. Thank you for your time, and I'm sure I'll be back for the Q&A.

Thanks, Brett.

Thank you.

We've covered a lot of ground today, and I want to thank Pratik and Brett for shedding some light on what our software does, where we compete and why we win. And I'd remind everybody that we're only able to cover, in the interest of time, two of the big areas that we participate in. We didn't cover master data management. We didn't cover API and application integration. We didn't cover data marketplace. Perhaps in a future webinar, we'll have a similar session. As we wrap up, I want to make a few key points. Cloud computing has driven massive changes in data management. Complexity has increased, vendors and competitors have changed and multi-cloud is the norm, and that is increasing, not decreasing. In that environment, enterprise-grade data management from an independent best-in-class vendor is every bit as essential in the modern cloud data world as it was in the on-premises era. Informatica's born in the cloud IDMC platform is uniquely able to serve the needs of enterprise customers in the cloud today and in the future. Informatica's value proposition in the cloud is unique, and it cannot be fully replicated by the competitors we face today. Not the point providers, not the open source and roll your own solutions, and not the cloud data warehouse and hyperscalers themselves with their proprietary tools. The IDMC, which is a collection of the best cloud data management products on the industry's only cloud data management platform, acting as the Switzerland of data, serving our customers' multi-vendor, multi-cloud and hybrid needs, allows us to deliver the differentiated value proposition that we've been discussing today in cloud data integration, in cloud data catalog and governance and in MDM and the other categories where we participate. Now I'm going to end with a little bit of data that I think validates this assertion. It's our own usage data from our IPU customers on the IDMC and their usage of 4 of our leading cloud data management services on the platform. As you can see from this data, we are the leading cloud data management solution provider today. So I'm going to end it there and let's turn it over to Q&A, and I'd like to invite my colleagues up here to kick that off.

All righty. We're going to jump into Q&A. Thanks for sending me your questions. Keep them coming. The first one, I'm going to direct to Pratik. Zero ETL. Seems like zero ETL is mainly for ad hoc queries. Where is the processing of the query happening? Is it happening in the CDW? Could it be possible for the CDW to provide certain ability to join data in the transactional systems directly to answer complex questions, or would that be processing, latency or cost prohibitive?

Yes. Very good question. So as we covered, zero ETL does, in fact, work really well for ad hoc queries, and it's a very good question in terms of where the processing happens. There are 2 places that the processing can happen. What it allows us to do, and as I covered briefly in my talk, the underlying data format is generally the open table format that we talked about for zero ETL, which allows various engines outside of CDW to also access that particular data. So it can be processed in any execution environment, generally speaking, it will be processed by -- within the data warehouse context or directly by the analytics layer on directly. And the solution or the use cases it's trying to address is the ad hoc query. The other part of the question was, can it be further joined to address that? That's where the latencies will come in. And that is why for enterprise scale analytics, that's not really an ideal solution, if you have to actually look at the data that is sitting in somebody else's cloud. And now you are bringing that dataset over, joining it and then surfacing the value. For real analytics, you want that data to be present where it is done. Security constraint, governance constraints are other reasons why those architectures are required. But just purely speaking technically, that is a hard thing to do, to join, to filter, to do all the processing that generally you do for analytics for as part of zero ETL.

So second part of that question is what about federated data or data sharing. What does it mean? How does it work?

Federated data and data sharing also is relied on the same thing. Zero ETL has a lot of terms today. Some people are -- some vendors are calling it zero copy. Some are calling it as links and references. This has been around in the past as data virtualization technology, so it's not really a new tech as such. And all it talks about is making that data available with a reference rather than actually physically copying the data, which is generally needed for large-scale analytics projects.

Great. Let me ask Brett a question. How widely adopted is data catalog within your customers? And how much of their data is cataloged within Informatica?

Great question. We don't talk specific numbers. I will say that we have a very large chunk of our IDMC customer base that is using our catalog and governance solutions in one way or another. In terms of how many of their data -- how much of their data state they cover, it widely varies. I mean, we often talk to customers and say, look, start simple. Start with one data element and expand. And then we have customers using it for -- to basically create catalog and data sharing across their entire organization. In that case, they have millions of data assets sitting in the catalog. And so it varies by customer quite a bit, but what we're seeing, especially fueled by GenAI is a much -- people are running to this to make sure that, man, I need to ramp up how much data I share across my organization, which means I need to have control over it, I need to have governance around it. So we're just seeing a great demand and a great adoption here lately, fueled by a lot of these GenAI projects.

Great. Pratik, can you please talk about the risks and opportunities for Informatica with the emergence of open tables?

As I said, formats are continuously evolving. For the last 2 decades, there has been a number of formats. Those are horizontal formats that I showed in this slide like XML and JSON that then gave rise to Avro, Parquet, ORC, and then now we are talking about Iceberg and Delta. With every generational shift, there has been -- those formats have been able to do more. In fact, for -- to convert from one format to another format, you need a data integration tool like Informatica. So we see that as a net positive for Informatica.

And then another question. With the hyperscalers leading the charge with LLMs, is there a risk that they capture more customer mind share, including being more competitive with Informatica?

They are not competitive. So LLMs serves a very, very specific purposes. Informatica is in the business of data integration and data management. So that's what we do the best. And good data is needed for good AI. And all of this data, 90% of the data that customers want to solve in the context of their AI project is sitting within their firewall, and that's what Informatica is able to unlock and then blend with the LLM data, the general models, to provide the value that is very specific to that enterprise. So from that context, Informatica, I don't see that as a threat or a competition. In fact, we are partnering with all the hyperscalers to bring that blueprint in terms of what the customers need to implement a good AI project internally.

Pratik, maybe I'll add on to that a little bit. We do compete against the hyperscalers and the cloud data warehouse providers to some extent. They have data management tools, and they're pretty good. As we talked about today, they generally work within the 4 walls of their environment, and their connectivity with external sources and targets is limited, but if you have a relatively simple use case and it's on a single cloud, those tools are sometimes sufficient. We don't even try to compete for those situations, for those use cases, because that's not where our value add lies. Our value add lies in the complexity in the multi-cloud world that the modern enterprise lives in, and where that is the situation, we compete and we generally win, and we partner with those hyperscalers and cloud data warehouse providers for their piece of the puzzle.

Great. So turning to catalog. Doesn't iceberg provide a metadata layer? Could continued open-source development in that project lead it to evolve into an independent enterprise catalog over time that people gravitate towards?

Yes. I mean, obviously, there are capabilities to develop a catalog on top of an open table format. A great example of that is Snowflake introduced Polaris and then introduced it an open-source project to extract metadata from that format. But a great -- another great example of how that -- maybe just the function of having so many different data sources in that environment is they use a different horizon to do all of their cataloging for other areas of their data sources and data lakes. So I think that's a perfect example that we don't see this consolidation of one format driving -- all of a sudden driving a consolidation in the catalog space. If anything, the formats -- Pratik spent a lot of time on this. The formats continue to kind of expand and follow that pattern, which means that our IP, our ability to scan all of those different sources to provide, extract rich metadata and do it in a consolidated way is what our customers continue to find valuable.

And I'll just add on that, to reiterate a point I made in my presentation that convergence is a popular concept, and it sounds good, creating simplicity with a new standard or a new way of doing things, and that's just not the way enterprise data works. It's not what we've seen in the past. It's not what we see today, and it's not what independent analysts and third-party surveys say is going to happen in the future. The world of enterprise data is going to stay complex, likely get more complex, and it's going to be more multi-cloud, not less. That's where we play.

So Mike, staying on you for a second. I got a question. Mike, you mentioned your historical strength in gold tier data. Can you talk about more of the specific technical investments that you are making that could expand the universe of integration and data catalog to other tiers of data? Or is that not as relevant to your use cases?

Well, sure. And maybe I'll ask Pratik to fill in any of the technical details I missed because I'm not a technologist, but we are investing in our cloud data integration engineering capabilities all the time. We introduce on a regular cadence at Informatica World, and again, a big release this fall that exposed additional capabilities that we've been investing in over the last year or 2, and that's going to continue. And stay tuned for Informatica World, because there will be some exciting things that are displayed there. And in terms of the actual management and modification of the data within the data warehouse, within the data lake, within the data lakehouse. As you can see from Brett's presentation, we do that. We support that. That's a core way that our customers use that, with an ELT pushdown format and other categories. So absolutely, it will be a continued level of investment, and you will see us unveiling new capabilities of that nature on a regular cadence.

And Mike, with respect to the specific question about whether we do play a role beyond the gold tier, actually, we play -- we play a role across all the tiers in preparation to the gold tier. So bronze, silver and gold. That is what I covered in this slide as to how we help prepare and refine the data for the gold tier. Now that's just one part of it, which is for analytics and for data warehouse. What we did not cover today is how reverse ETL helps back to get that data and take it back into the system of record. We do data synchronization between 2 systems. We do application integration and data integration. So as from an integration solution point of view, Informatica provides value way beyond just the creation of the data warehouse. It allows you to do operational data integration. It allows you to do application data integration. It allows you to do real-time data integration. So all the various what I call as patterns of integration.

Yes. And similar on catalog, I showed that lineage chart. So we capture all data. So as it moves through and progresses through these tiers or it progresses into a data warehouse, data lake, we allow you to track all that and know where that data came from, what operations took place on it, right? How those -- how that data is related to other data in your environment, and make sure you can govern it, make sure you have control over it, visibility over it, et cetera.

So Pratik, going back to tables, are you seeing customers adopt Iceberg and Delta tables?

Yes. We see customers embracing this new format. It has a lot of promise. The unique differentiation, as I covered earlier, which was not the case in the previous formats, was Iceberg and Delta both allows you to transact on that particular data store or the format. So you are able to do asset transactions, you are able to do current operations, which is -- it is acting like a database and hence, you can do read, write, filters, queries, all of that on top of that. And I believe that has a lot of value for the customers to further democratize the storage tier into this open -- into the lakehouses, which can then be made available for analytics as well as various other use cases.

And at the risk of sounding like a broken record, I want to remind everybody again that we support Iceberg, we embrace it, and we view it as a net positive for our business in the long term.

Exactly right.

So switching topics, where does real-time streaming fit into data architecture?

Real-time streaming is another pattern of integration, which actually processes data in real-time. There are 2 parts of it. One is what is called as real-time analytics. Customers want data to be brought in as soon as possible for doing real-time analytics and for certain workloads, that's very, very important. So we do solve that in 2 ways. One is all the connectors that I showed in my slide. There are a lot of real-time connectors that allows sources and targets to be received. So the real-time sources and targets, we are able to extract the data in real time and take it to a real-time data store such as messaging systems, whether it's Kafka-based or Kinesis-based or Azure event-based, we can take to any of the event stores, and that allows us to participate in the real-time architecture. The other one is we support real-time processing. So that is called streaming sources as it comes in, whether it's from IoT or whether it is from change data capture, our engine is able to process their workloads at scale in real-time, processing that and taking it to the target. So both use cases, whether you are connecting to real-time sources and targets or whether you are processing real-time, Informatica is able to do that both ways.

Great. So another question. We've touched on this a little bit earlier. It came back again. We understand that some hyperscalers have their own data management capabilities. Can you talk about what Snowflake and Databricks are doing? Do you compete with them? Where do you see them?

On the -- on what particular area?

The question is on data management.

Data management, what are these vendors doing? So both Databricks, data warehouse has a data platform. They started to provide a very strong analytical data science and now AI use cases. So in that context, they have to work. Their core competency is in how they manage the data store and how they process that data. Now in the context of that data, they are investing a little bit on subparts of data management, as Brett covered earlier. That includes cataloging, that includes some level of data transformation that can be used once the data is landed in those systems. So we see them investing in those areas, but their core priority is ensuring that they take the analytics, data science and AI use cases.

We partner really well with Databricks. We partner really well with Snowflake. In fact, this year, Databricks named Informatica is their Data Integration Partner of the Year. It may sound like we compete against them and we have duplicative surfaces here and there, and as Pratik talked about, yes, they do a good job inside their environment in doing the data management necessary for some very important workloads. But we're not in a converging world.

We are not in a converging world.

We're in a world of complexity that's increasing. We're in a multi-cloud world that is becoming even more multi-cloud. That's where we come in, and that's the value we add, and that's why we partner so well with all of those providers in that tier.

Well, I think they recognize our strength in the cross ecosystem, right? So when they come to us, it's really because, oh, we realize that you can help us work with customers who have these more multi-ecosystem-type environments. And so that to me is what fuels a lot of their interest in us. The reason -- part of the reason they view us as a great partner is, again, you said earlier, if they're using only them, if they're only using one ecosystem, it's where all their data is, it's where all their data stores are, we're not trying to steal that business. We are trying to be a value-add for a cross ecosystem where they see us as strong.

And our enterprise value, back to the slide that I showed, the 5 value proposition, especially in data integration and engineering, it's about the high productivity. It's about the scale that we provide. It's about the various connectivity options, both within the enterprise, outside the enterprise where they support. The workload management was a huge differentiator. And those are the reasons why we continue to provide -- we provide that high value, and hence, we partner with them closely when customers are looking for that type of enterprise-grade solution.

Great. So Brett, a question for you on data governance. How are customers thinking about data governance for their day-to-day regular use cases? And then how are they thinking about it for their AI initiatives?

Great question. So I would say, look, enterprises have -- they live in a world where all of their data is under some kind of regulatory or even corporate-driven compliance guidelines. And that's kind of the -- right -- that's kind of a level setting of, hey, I need to start having control and making sure that I'm compliant, that I can prove that we're taking care of this data, protecting sensitive information. So that's -- that's kind of any data. Now I mentioned this earlier, which is that AI has been around for a while, but generative AI is certainly new. All the GPT stuff is driving much more -- kind of data feeding these models in a much more -- much higher volume. One of the things when I talk to customers, I find that they are seeing people inside their companies who a year ago or 2 years ago would not be asking for data are suddenly asking for data. They want access to the data. Well, how do you -- and so if you had manual processes where you're trying to, hey, get data, people the data that should have it and make sure it was protected, those manual processes don't scale, all of a sudden, you have hundreds of people asking for it. So what we're seeing is, hey, maybe I used governance for some of my regulatory stuff, but now I need to scale up my data sharing, right? We talked about kind of data regulatory. We talked about data sharing, we're talking about observability. Well, as I need to share more, I got to bring in more of these tools. And as I'm sharing more, I also want to have observability as to what I'm sharing, how it's being used and am I in compliance, all those kinds of things. So I find that generative AI has turbocharged, if you will, the need for catalog and governance, but catalog and governance is needed for all data for most enterprises.

And I would just point you to the IDC survey data that was just released yesterday. We showed some sample results in the presentation today. At the beginning of our little show here, we started with our corporate video and our tagline, everybody is ready for AI except your data. Our customers are coming to that point of view very clearly. And all the survey data, all the conversations we're having with enterprise customers is that they agree.

Great. Well, that's all the questions that I've received. I don't know if we have any closing remarks we want to leave the audience with. I'll turn it back over to you, Mike.

Thank everybody for joining, and please with questions through Victoria or myself or our IR channels. And stay tuned. We only covered a couple of the key parts of our business today, cloud data integration and engineering and master data -- sorry, and cloud data governance. There's a lot more that we do, especially master data management, application integration and API management. And in 2025, we'll plan to have another session like this to shed more light on those parts of our business. Thank you, everyone, for joining, and happy holidays, everybody.

Thank you.

Thank you.