Yext, Inc. (YEXT) Earnings Call Transcript & Summary

Thank you for joining us today for the Yext’s 2023 Investor Day. Our presentation today will contain forward-looking statements, including statements related to our future financial performance, expectations regarding the growth of our business, our outlook for future periods, our strategy and estimates of financial and operating metrics, capital expenditures and other indications of future opportunities. These forward-looking statements are subject to certain risks, uncertainties and assumptions, which are discussed in our reports filed with the SEC, and we encourage you to review them. Yext is under no obligation to revise any of these financial statements to reflect changes that occur after this presentation. Throughout the day, we also refer to certain metrics, including non-GAAP financial metrics. Definitions of these metrics and reconciliations of non-GAAP financial measures with the most comparable historical GAAP measures are available in the appendix section of the Yext Investor Day 2023 slide presentation, which is available at investors.yext.com. It is now my pleasure to introduce Mike Walrath, Yext's Chief Executive Officer and Chair of the Board.

Thanks, Nils. Good afternoon, everybody. Welcome to Yext. I'd like to thank you for being with us today. It's great to see so many faces here in New York as well as those of you on the webcast. We're excited to share a little more about the future of Yext for you today. It's an incredible time at the company. I really look forward to sharing that with you. We find ourselves today at the center of one of the most interesting technology innovation cycles we've seen in a long time: unprecedented interest in AI; large language models; and ongoing transformation of how enterprises think about digital experiences. Here's what we hope to accomplish today. We'd like to start with a quick look at the last year and some of the fundamental restructuring changes we've made to the company. We'd like to talk about some of the progress we've made, and I'd like to share a high-level overview of our operating philosophy and plans for the years ahead. After that, I'd like to talk about, at the highest level, about the massive shifts we are seeing in awareness and focus around AI solutions for the enterprise and where we fit in that evolution. We've got a great program for you. When I'm done, you're going to hear from key members of our operating team on topics including our product road map, unified go-to-market, customer voices and financial updates. We'll wrap with questions at the end of each section, and we'll be hosting an informal cocktail reception at the end. We've got a beautiful day in New York. I think it's the first really spring-like day we've seen, and so we'll be taking advantage of our roof-top patio. We hope that you'll stay for that. By combining our focus today on 2 things, the enormous market opportunity ahead of us and more granular operations and execution topics, my hope is that you're going to get a picture of how bright the future is for Yext. But you're also going to have a really clear understanding about how we're going to attack the opportunity with discipline, transparency and operational efficiency. Let me share with you some of the progress that we've made this year. We've reduced our overall head count from about 1,400 full-time employees to about 1,100. We've reduced our executive roles further. We've reduced VP plus roles from about 104 to about 73. And this is directly related to becoming a more streamlined organization. When we look at Q4 of fiscal year '22 versus Q4 of fiscal year '23, we saw a 10% of revenue decrease in our sales and marketing expense from 51% in Q4 of '22 to 41% in Q4 of '23. We also reduced our total share count from approximately 130 million shares at the end of Q4 '22, to approximately 122 million shares at the end of Q4 '23, and we'll get into dilution- and share count-related topics as we go on. And finally, we talked a lot this year about our quarterly renewal rate. We saw a significant improvement from Q1, where we saw a quarterly renewal rate in the low 80s to a high watermark in Q4 for the year where we saw a quarterly renewal rate in the high 80s. We've historically talked about this metric as gross retention and we're going to be referring to it as quarterly renewal rate, to the extent we refer to it going forward. Darryl’s going to get into a lot more detail on this metric and how we're going to be talking about it going forward. What we do believe is that this progression is showing us that the -- is an important indication that the work that we're doing on customer satisfaction and reducing churn is taking hold. So the next thing I'd like to do is talk a little bit about how our organization has changed. Just under 13 months ago, we announced significant management changes, which many of you are aware of, including myself as CEO; Darryl Bond as CFO; Marc Ferrentino, in the role of President and COO, to unify our product and go-to-market execution efforts. Since then, we've hired new executive leaders that include Raianne Reiss as our Global Chief Marketing Officer; Tom Nielson as our Global Chief Revenue Officer; and Yvette Martinez-Rea to oversee the newly created corporate development position. We've also retained key leaders with long tenure at Yext, consolidated responsibilities and created a much more efficient organization. I am totally confident in each member of this team's ability to deliver results for their area of functional responsibility. Most importantly, by consolidating functions and eliminating operation silos, we've created a global, unified operating team. The culture of performance and accountability across this group is stronger than it's ever been. Each of the leaders that you see on this slide has global responsibility for their function. The team is aligned, the team feels responsible to deliver results, and results is exactly how we expect to be judged by our shareholders, employees, customers and partners. The question we get a lot and the question we like to talk to is obviously going to be growth, and I'm going to come to that in a moment. Now that we've got the company structured properly from an operational perspective, the natural question is how do we get it growing again. As I've said many times to many of you, the underlying operating philosophy we've installed here is that growth starts with efficiency and using the data to tell us the best way to reaccelerate the business. So how do we get there? This involves understanding when key metrics like sales productivity, qualified pipeline, ARR retention and the macro environment are lined up to allow us to reaccelerate growth. We use Rule of 40 as a guide, and we think that the growth opportunity, both revenue and EBITDA growth, are best framed through an understanding that a company can't force revenue growth and remain efficient at the same time. This entire team is focused on returning to revenue growth when the data tells us that we can achieve it efficiently. This approach ensures that we're going to continue to make progress against the Rule of 40 metric in different operating environments and that we're prepared to accelerate growth as the opportunity to do so becomes clear. So now I'd like to talk a little bit about the evolution of our platform, and we're going to get into this in a lot more detail as we go through the presentation. But I think the history is important here. So we started as a CMS to manage content for a company's listings. We then became a CMS for SEO landing pages, followed by a CMS for reviews and then a CMS for search experiences. Today, this platform can be a CMS across a variety of digital experiences, including full websites with the upgrade to our Pages offering. We were very early to AI. Some might say too early. We started to develop our Search offering in 2017, and we wanted to bring something differentiated to the market. At the time, Google had just put out a paper on large-scale language models. You might have heard of these models like BERT, more recently, GPT and there's many others. We decided to base our Search technology on these large language models and wanted to understand language and semantic understanding and build that into our models. Over the last 5 years, we've gotten very good at creating and training these models. It turns out these models are good for much more than search. The investments we've made over the years have allowed us to embed machine learning and AI capabilities into the core of our platform. For example, we've added AI to our connector framework a few years back. Within the transformation layer, we have the ability to do zero-shot or few-shot models that clean, organize and enhance the data coming from other systems. We leverage AI in our Listings product to do matching and verification of each listing. We leverage AI to generate content in our headless CMS we call the Knowledge Graph. We have leveraged AI in our Search product since the very first day. We use many different transformer models to create vector embeddings and extract answers from large documents. Internally, we leverage these large language models within our review response team to increase productivity. And of course, our recently announced chat offering uses multiple large language models to deliver a great chat experience while allowing the enterprise control and governance over that experience. And we're going to talk about how important that is. What we have seen over the last 6 months is a sea change in interest in how AI will transform the digital industry. The lightning bolt of energy that was the launch and growth of ChatGPT has captivated the world. AI is touted as the technology that's going to change everything. And at Yext, we totally believe that's true. And that's why we've been working with models like ChatGPT, others and even our homegrown models for years. The excitement around AI is palpable and justified. However, like most new disruptive technologies, generative models and the advances in AI come with very real risks. As the use cases for AI become more visible, the risks are being better understood, particularly for businesses. I think this works best when we use a real example. So I recently asked ChatGPT to write a biography for me. It's a fairly simple thing. And this is what it wrote. And actually, at first glance it looks pretty good. It's impressive. It feels like it could be very accurate. I did found Right Media. It was acquired by Yahoo! for $850 million. I was born in 1975 in Brookfield, Connecticut, not Florida. I did attend a university in Virginia, but it wasn't the University of Virginia. I'm a proud Richmond Spider, although I was a UConn Huskies fan before I was a Richmond Spider. I have an English degree, not a degree in computer science. Brian O'Kelley worked for me at Right Media, but he was not a founder. I would have loved to have started LinkedIn. I had nothing to do with it. I was totally uninvolved in AppNexus. That was Brian O'Kelley. And I was never on the 40 under 40, and I'm no longer qualified. If I asked ChatGPT to write the biography again, it would do it again, and it would do it differently. And it would be a similar mix of fact and fiction. So you see the problem, right? What we're looking at is a statistical model of the probability of the next work. That's it. And it's an amazing piece of technology, but once you understand how it works, the parlor trick, while impressive, is a little bit less magical. And we can laugh about it today and it's amusing, because what does it matter if they get my biography right and if there are some misses in there. And they'll get better, right? It'll definitely get better. But put yourself in the shoes of the CEO of a health care system or a bank or a chain of donut shops. Can you allow AI to answer the questions of your customers, partners and employees unsupervised? The answer is clearly no. So let me tell you just one way generative AI will certainly get better. When a business is able to put the power of the generative technologies to work with a Knowledge Graph-led approach to create a set of authoritative answers curated and managed by the business itself, then we can start talking about delivering perfect answers across the full spectrum of digital experiences using AI. The more fragmented these digital experiences become, the more important it will be for businesses to control how answers to questions are provided. Today, you're going to see how we're putting the Yext Answers Platform, including Knowledge Graph and a flexible language model-agnostic approach to AI content generation to work for our customers. You're going to learn more about how our technology works, how we're going to take it to market and how we're going to use it to solve some of our customers' stickiest problems. You've heard us talk for years about the size of this opportunity and the opportunity to bring these technologies to the digital experiences. We're also going to talk about that, and we're very excited about that opportunity. But I want to finish where we started. We will not operate the business on hopes and dreams. We will continue to take a disciplined approach to how we run our business and bring as much transparency to our partnership with you, our investors, so that you can judge for yourself the opportunity and how we're executing against it. I hope you enjoy the content today. I look forward to taking your questions at the midpoint and again at the end, and I hope I'll see you all at the cocktail event afterwards. Now I'd like to turn it over to Marc Ferrentino.

Thank you. Thank you, Mike. Thank you all for being here today. I put a jacket on for you guys. I hope you appreciate it. As for those of you who know me, I don't normally wear a jacket, but I thought it was the right occasion. So I want to talk a little bit about customer journey and really the evolution over the last few years. So it used to be that to make your business digital, all you had to do was build a website and buy some ad words. That was pretty much it. So if you wanted someone to learn about you, the entire customer journey would start with your URL. And the website was the center of your digital experience. But over time, things shifted. Over time, the customer journey fragmented. Now a customer interacts with a company over search, mobile, chat, social, messaging, maps and less and less your website. Now there is no real true center to the digital experience anymore. All of these new digital touch points are just as important as your website now. Many of these touch points like voice and search and chat and messaging are now driven by innovations in AI. So a user, for example, could come in through one channel, then switch to another and then another. A customer in theory could engage with all of the channels in a single day because each one is optimized for a different part of their daily life. At this point, your company needs to manage all the experiences across all the digital touch points. And while almost every business and the bulk of the experienced vendors out there focus on creating experiences on a company's own website, it turns out your brand is everywhere. Customers and prospects are experiencing your brand on third-party sites like Google and Facebook, Bing and Apple and more. In fact, we ran a study that showed that more than half of your brand interactions are happening off your website or mobile app. You can't just ignore half the journey. And user expectations have never been higher. All of the customer -- All these different consumer services that we use today, they set the bar for customers and prospects and what their expectations are and how their interactions with your company online will be. And actually, they set the expectations for your employees, too, and their work environment. ChatGPT, great example, Exhibit A, resetting the expectations of experience once again. Now all these great consumer-grade experiences, they all have a few things in common. One is they have incredible search, great search, right? Netflix, great search; Pinterest, great search, right? Great content. They all have great content, too. When was the last time you went to Spotify and found a song or an album that didn't have a description or some content along with it? They're all multimodal: website; mobile; messaging, right? They're leaning into this fragmented journey and meeting the customer where they are. They're all super-fast. I'm talking sub-100 millisecond response time, which for anyone who doesn't know what that means, it's fast. It's really fast, right? And most business experiences are not even close to that. Cutting-edge technology, they all have ML or AI groups inside their organization, and they're implementing those technologies right now. Except these companies, they also have tens of thousands of employees to make this awesome, to make the experience awesome. They have huge engineering teams, they have scalable infrastructure, they have armies of content people and, of course, they have data science and ML and AI teams. Most businesses do not have these sort of resources to deliver a comparable digital experience to these big guys. They need help. They need a partner to help them level up and to keep up. Now let's talk about -- I've said digital experience. Let's talk about what I mean by digital experience. Let's look at a modern digital experience. Let's take, for example here, a typical digital experience for a retail business. So we have, for example, a prospective customer might find something on Facebook or Instagram or Snap. They might find a new product that piques their interest. They then read reviews online for that product to see if other customers also thought the product was interesting and if they had a successful purchase. They then go to Google or Bing, right? And what do they do? They look for a competitive offer. They look through, and probably price shop after that. They then visit the website of the retailer, and they look up product information. They maybe read a blog or two. Then they want to try out the product typically, right? So they maybe go to a -- look up a store location, and they'll sign up for an in-store appointment potentially, which then leads them in-store, to a kiosk to sign in and check in for their appointment. And then ultimately, they go home, maybe check the return policy, read an FAQ or two, and then they purchase the product online. The product then arrives, and they begin using it, right? That may lead them to follow-up questions, which will lead them to support portals, to support articles and even to support agents. There is a team of people at the retailer who thinks through this every day, right? And they assemble this digital experience. They put this together very thoughtfully with many, many different vendors. And the entire experience was built on top of this set of digital building blocks. The digital team at the retailer cobbled together all these digital services to build this end-to-end digital experience that we just looked at. They did not use one vendor because there is no one vendor that does it all. It takes a village of vendors to deliver a proper digital experience. All these building blocks -- all these different building blocks here are what the industry calls composable services. You need vendors who subscribe to this architecture in order to plug and play best-of-breed technologies. You can see how hard this is for a business to accomplish this all and do it well. It's not easy. They need help, but help is on the way. Now I think it goes without saying that AI is going to disrupt every industry. I know we've been saying it for a few years. But with the innovations over the last, call it, 12 months, I think everyone can see how the time is finally now. And we believe that AI will disrupt a digital experience space and that really is our vision for Yext. Yext is leveraging AI to completely disrupt the digital experience space. Our vision is to leverage AI and graph technologies to enable every business to build a seamless, multichannel, consumer-grade digital experience. Sam Altman, CEO of OpenAI, recently said in an interview, too much of the processing power of AI is going into using the model as a database instead of using the model as a reasoning engine. And we couldn't agree more. And we built our platform with that in mind. We believe that the architecture of the future will consist of a data layer that will store the knowledge about your company, an AI layer that will handle generation, reasoning and orchestration and an experience layer where you will fine tune and deliver each experience against each channel. You want the inspection and [ artability ] of the data layer, in our case, that data layer is built on graph technology, and combine it with the logic, generation, orchestration and reasoning capabilities of AI to deliver these incredible experiences. And we want to make this available to companies of all sizes. Now look, this concept completely turns upside down the traditional way of how experiences are created. For example, a content management system, or CMS, is not a CMS anymore. You don't want to think of content as the web page, right? A CMS is now a representation of all the sum facts about the business and its relationships because the content is now generated by AI for each channel that a user might engage the business on. So in a world where I can generate 100 versions of an article about a topic or a product, in a world where I can generate 100 images to make your product most engaging and appealing, where I can create versions of articles and content that are personalized for each person, what is content? What is content in this world? So for example, the business wants to put an offer out for one of their products or services, to do that, all they need to know is what the offer is. That's pretty much it. The AI can write copy for every channel, web pages, mobile, social, conversational. The offer can then be communicated across any channel. The experience is generated and it's augmented by the AI. And digital experience platforms are a major pillar of enterprise software. It is foundational to a company's operations. It's a must-have. Even when there are downturns in the economy or maybe the business itself is having some troubles, they can't turn off their digital experience. It's not possible. So we are really excited to be part of this massive category. And Darryl will talk about it a little bit later today and go into some more detail. So Yext delivers a platform, ultimately, that helps the business build incredible consumer-grade, AI-first digital experiences, and we call it the Answers Platform. The Answers Platform is a composable, API-first digital experience platform that is made up of several composable services that work great together, but also can be used individually with other vendors' products. So at the center, we have our headless content management system, which we call the Knowledge Graph, collecting data and content from across the business is foundational, obviously. And so we built out our connector framework, which is a configuration as code ETL tool that can bring in information from anywhere across the enterprise. Once you've collected the data about the company, you can now create and deliver digital experiences anywhere. So for example, you can control and deliver a digital experience on third-party platforms like Google, Facebook, Bing, Apple and many, many more with our Listings product. You can monitor, respond and generate reviews about your locations, your products, your company with our reviews offering. You can build and serve multiple search experiences across any part of your organization with our Search offering. You can build high-performance SEO-optimized websites with our Pages product. And recently, we announced -- actually, being able to announce, our chat product, where you can create compliant AI chatbots, all built with the latest large-language models. Now our platform overall is built on top of a scalable, secure and compliant cloud infrastructure that you can deliver experiences globally with low latency. We have enterprise-grade user roles and permissions, an ecosystem of integrations and software partners to make it easy to work with other platforms and, at the core is our AI and machine learning foundation, which allows us to build and deploy new models and leverage them across any part of our technology stack. Michael, in a minute, will go into more detail about our AI infrastructure. And since we are a composable, API-first platform that will integrate with any other experience products, such as Adobe's CDP or Salesforce's Commerce Cloud, we want to make sure that consumers will be able to build their digital experience of their dreams. A customer can use as much or as little of our platform as they like. We hope they would use more, obviously. But the flexibility is an important part of a modern architecture and is customer friendly. Many vendors, they try to lock you in with one sort of monolithic, giant offering that is really not great for the customer long term. And you can see, when you compare us to existing digital experience vendors, you can see the difference. We help companies deliver consumer-grade experiences with modern AI-based technology. We are built for a multichannel world. We know having great search as part of your experience makes all the difference. We are built on a modern web architecture that delivers very, very fast SEO-optimized websites. We believe in open-source technologies and languages and we want to make sure that what you build is actually portable to other platforms. And we are built for this amazing new conversational future that we've been talking about for years and it's finally here. All right. Next, I'd like to dive into how we leverage AI in our products and our model-agnostic approach. So I'd like to invite up on stage our Director of Data Science, Michael Misiewicz.

Hello. Thanks, Marc. Good afternoon, everyone. My name is Michael, and I lead Data Science here at Yext. So what I'd like to talk to you about today is our strategy and outlook on the amazing breakthroughs that are taking place currently with large language models and how we're integrating these new technologies into our products. The key thing I'd like to communicate with all of you is that it's not really -- we've actually been in this game for a while. We're not new to this game. Our Search product actually was built from the ground up, starting in 2019 with large language models, way before anyone was even talking about GPT. So a question to ask is, well, why did we start early with this technology? Well, that was not long after the August 2018 publication of the BERT paper, which was a landmark paper in this field. And they showed some really key results in that paper that were very important. BERT was demonstrated to be dramatically better at understanding words and context and also retrieving data. And these are things you need to do really well to build a really good search experience. And large language models showed us the way. And we were able to dramatically improve the quality of the search experiences by using this technology. But there's still more of this. We spent a lot of time trying to figure out what makes a good search and what makes a bad search. Obviously, we want to make bad searches get better and one of the factors that we found, no matter what we did, we always found that content was the most important factor in a system. So that's why we started using large language models to improve the content in searches. And so we started to build up the connectors so that we could build better Knowledge Graphs for our clients, also powered by large language models. And we also spent a lot of time making sure that we were working on the boring stuff. And what do I mean by boring stuff? Well, it's things that never really make the demo, and it ends up being about 90% of the work behind the scenes. And I'm talking about things like warehousing, data ingestion, data labeling, model retraining. And if you want to pull off a really powerful algorithm-powered product, you need to get good at this stuff. And that led us to this concept called the virtuous cycle of AI-powered features. One of the things that you might notice about a lot of famous products that are powered by algorithms like Google Search, ChatGPT, Netflix, Instagram, all of them have this virtuous cycle, which is the more you use them, the better they get and the more of a strategic moat the company has, who makes them. So the cycle starts with good logging, warehousing and instrumentation. You need to be able to log predictions from your models and then all the interactions that end users end up having with those. In ChatGPT, for example, this is when you thumbs-down a bad response. Or in Netflix, it's when you click on a video they recommended, or Instagram, same thing. Next, once you have this data that's coming out of this instrumented system, you need to start getting it into the hands of humans to label and retrain and improve the data. And then you have to use that as the basis for what you do next. So we have a great labeling team here at Yext, headed up by a linguistics expert and she manages all the hard work that's required to write correct labeling guidelines, to hire the annotators, to make sure it's all done correctly and ethically. So once you have that, you can then be ready to deploy your models out into production, into new features and the cycle repeats it again. So more and more you use our products like Chat and Search, the better they get and the more of a differentiated data difference we'll have and a moat that'll be strategic for us. So how do we actually go about building one of these models? Well, I'd like to walk you through some of the technical and operational details of building a large language model or a product feature powered by them or, in other words, what the data science team actually does on a day-to-day basis. Well, at its core, there are 3 steps and 2 artifacts that are produced when you train a model. So I'll start with the 2 artifacts. Every large language model is comprised of 2 things: a computer program that defines the model; and a set of model weights. So all large language models these days are implemented as deep neural networks. And they are written in Python programs, which define all of the neurons and how they interact with each other in that artificial neural network. The weights define how strongly they are connected to each other. So when you have these 2 things together and you've obtained the weights through training, you can start to do inference for example, on a product feature. So in the first step of this process, you need to define your problem. Are you working on chat? Are you working on search? Are you working on review response? Are you working on content cleaning? Are you dealing with images? Are you dealing with text? All of these -- the answers to all these questions have a big impact on how you structure your program and the architecture of the network. That, in turn, is going to define what you put in your implementation. One thing I want to point out here is that large generative models aren't necessarily using a new problem framing. The idea of generation has been around for a while. But what has gotten different and better about them is that the architectures are way more complex and capable. But the more things change, the more they stay the same. You still need to do a really good job with the nitty-gritty details and execution, warehousing, labeling, that sort of thing. So in the second step, you need to get good. clean data to train your model. And that's where that virtuous cycle I was just talking about comes back in again and where our labeling team has made a big difference. They make sure all the data is labeled correctly, ethically and in a brand-safe manner. And we've made big investments here for a big difference. So your third step. Now you are ready to run the training or inference and you're ready to deploy your product -- deploy your feature in production. Building effective infrastructure for using graphics processing units is really important in this area, and it's in a place where we've focused quite a bit of energy. And you've seen in Mike and Marc's slides, how important the accuracy of these models can be, especially with generative models. So building product features powered by a large language model technology must take this into account. And you really need to make sure you're getting the balance right between speed and cost and prediction accuracy. Okay. So I mentioned these 2 artifacts, code and weights. How do you get them? You've got 3 options here. Option 1, is that you can build the program and the weights entirely from scratch. This is the most strategic and most differentiated, but also the most challenging and time consuming. Option 2 is you can borrow from somebody else. There's lots of people who release open-source implementations of many of these models, along with the weights. And this is a great middle ground because a lot of the models can be like building blocks and you can still get a massively differentiated product without having to go through all the effort of doing everything from scratch, even if other people have access to those same weights. The third option is you can buy through an API. Now, many companies are offering access to large language models. They're highly capable, although you have less options around costs and differentiation, but it's a really great way to get started quickly. So let's talk a little bit about the landscape of what's out there. There's a lot of models that are on the market now, and this mostly represents what's in the buy and borrow categories. And this slide is from last month and it's already out of date, from March. As you can see in the upper left corner here, we have an area of this landscape that is pretty well suited for search problems. These models are really accurate at understanding words and context, but they also run really fast and could be useful in a search experience because they can run in the path of a user query. But down on the sides here, we have the GPTs. Now these are available through OpenAI in a proprietary manner. And this is a lot better if you're doing something like chat because you don't have the same latency requirements and you want a better accuracy here. So they're a lot bigger and slower to run, but they are well suited for that application. Lastly, in the upper middle here, we have Bloom, which is a good research model if you're working on problems related to content extraction or summarization and review response. And that's also good if you're not on the critical path. But since it's open source, you can run and retrain it on your own infrastructure. So the point I want to make here is that we are all over this map right now. So last area, where do we see things coming next? Well, I think the first thing is that it's important to be flexible with many models. I think the space is going to continue to be commoditized and just being able to plug into a lot of them is going to be an important thing to do. Second, I think it is going to continue to remain important that we continue to execute correctly in terms of balancing cost, latency and accuracy and choosing the right model for the right job at the right time. And then third, we want to keep building out complex Knowledge Graphs for our clients because it's clear these models are becoming better and better at reasoning. And that opens up a pretty exciting possibility, like just imagine what you could do if you had all of your enterprise data stored in a Knowledge Graph and then you could use one of these large language models to reason about it, what kind of digital experiences could you build with that? It's pretty cool. And I will thank you, and I'm going to hand it back over to Marc.

Thank you, Michael. All right. So now I'd like to show a brief, 10-minute demo. And I can just tell you right now, 10 minutes is not enough time to show all the things that our platform does. But we picked out some of the highlights and we wanted to show you an example of what a customer journey looks like on our platform. And then what does it look like to make or to basically change some of the parts of that journey. [Presentation]

All right. So I hope that gave everyone a little 10-minute taste of what's possible, the platform, what we mean by digital experience, the sort of broadness of that definition, but we'd like to take it a step further. Next up on stage. I want to bring Christian Ward up. He's our Chief Data Officer, and he's going to talk about all the things that our customers are building on our platform. And I think this is a really exciting session.

Excellent. Thank you. Hello. My name is Christian Ward. I'm the Chief Data Officer here at Yext, and my role here at Yext is to work very closely with our customers and our partners on their data strategy to optimize their usage of the Yext platform on their journey toward AI-powered digital experiences. Today, the digital experience at all of our customers suffers from 2 key challenges. The first, called the information overload paradox, states that as we all have access to more information, our ability to process that information, as humans, actually declines. This is also known as cognitive overload or analysis paralysis. But fundamentally, this means that if I were to Google search something like dinner near me, I'm met with over 5 billion search results, just for dinner. This content and sensory overload, it's slowing down the customer journey, and it floods the customer with too much information. The second major challenge is the explore versus exploit trade-off. Consumers are typically balancing their time through different mindsets as they engage in any digital experience. The first is exploring their options, learning about a topic, a product, or a service. From there, though, the individual typically shifts their mindset to more of an exploit mindset, meaning, to take an action for their own benefit. This is where many digital experiences struggle. Better digital experiences bring the exploit options much closer to the exploration, so that people know what to do next. Essentially, digital experiences have overloaded human beings with content, while also making it very difficult to know what to do next. And this is where our customers leverage the Yext platform. We address both of these challenges to build better digital experiences. By centralizing the authoritative, human-curated content data and knowledge of every organization, companies are building on a platform that can leverage AI tools in a multi-model design across all the different digital experiences that they need. Let's look at some of those experiences from our customers. When the State of New Jersey needed to build their COVID-19 information hub, they leveraged the Yext platform to store their knowledge very efficiently. This knowledge is then deployed through this website. Now I'd like to point something out. This is probably the only government website you've ever seen that does not have drop-down navigation. This entire website is designed to engage the citizen to ask a question, directly, of the State of New Jersey. Now, inside the platform, the State of New Jersey can analyze all of the questions and engagements from their citizens to identify areas of new questions, concerns, or even opportunities to better the entire citizen experience. For example, by listening to the citizens during the pandemic and every query typed into that search bar, they were able to identify that people had concerns about evictions, due to the financial hardships caused by the pandemic. They utilized the Yext platform, not only to answer the question on their site, but also to publish that, as Marc showed, on a site on a page that also brings that content directly to the Internet. So now when someone types something in like, "Can I be evicted in New Jersey if I have COVID issues?", they're actually met with the authoritative answer, the top result, where we're telling them directly from the State of New Jersey, this is the access point for where you can find more information on this. What's great here is, once they click this link, it takes them directly back to the State of New Jersey's COVID information hub, where they can ask their next question directly of the State of New Jersey. By centralizing knowledge from the authoritative source, our customers can leverage the Yext platform across the entire digital experience, both on their site and off their site. In health care, customers are using the Yext platform to link patient care to physician availability. In government, we are helping bring clarity to complex processes like the patent process at the United States Patent and Trade Office, directly surfacing the right information from their Knowledge Graph. Digital transformation in industries like finance is all about simplifying the dialog, cutting down to those lean data interactions with the customer. With First Citizens Bank, their home page, content and sites have undergone a massive transformation, where customers no longer have to hunt through countless drop-down menus. By leveraging the Yext platform, this digital experience is as simple as asking a question. No matter what the industry or what the customer type, the Yext platform provides the digital experience with an authoritative, curated source of knowledge. Now, here at Yext, we are our own best customer for our platform. To give you a sense of the power of the platform, let's dive into the actual Knowledge Graph we use here at Yext. Now, the first thing you'll notice in our own Knowledge Graph is we have close to over 160,000 different entities. This is the data where we're storing all of the different content, things like our information and all the linkages between that information. So I can see right here, our blog posts, our FAQs, our jobs, our locations, our products, our services, even our integrations, all stored as entities with their attributes and the relationships to each other to build those digital experiences. But as Marc just demonstrated, if I go just below this section, I can see all the digital experiences that we're also producing from the graph. So I can jump to any given website. For example, internally, we have an agent desktop to assist all of our internal teams with any customer service requests. And this entire system runs on the Yext Knowledge Graph on the platform. I can jump back into the Yext platform and launch into a completely different experience, one that I'm very fond of, which is where the ideas from our developer community are proposed by people that use the platform. They're submitted, stored and even voted upon inside the Yext platform. And again, we store all that knowledge here on Yext. And if that developer has a question in the idea section, they now can launch our new AI-powered chat. Again, where we're taking AI and AI tools and sitting them on top of the authoritative knowledge that we have curated for this digital experience. Here's a completely different digital experience, over at our Hitchhikers developer portal, where they can search through all of the API documentation on how to build on our platform, as well as go through all of the content in our learning modules. Again, all stored here. So from our marketing site to our internal intranet site, called Telescope, the Yext platform streamlines the entire process into different experiences, each designed to meet the person directly on their journey, wherever they are on their journey. And when I jump back to yet another Yext experience, this is our internal knowledge base, I'm reminded that all of the searches, all of the conversations, all the chat, all that content, the engagements within the journey, are being captured and stored in the graph where we're applying our AI research tools to then look at all of the questions. So we can see everything that our customers, our employees are asking us. And that, in turn, helps generate new knowledge and new experiences. In fact, similar to the way Michael was speaking about in his large language model discussion of the flywheel, there's a business flywheel here for every one of our customers. It's moving them from this process of the classic marketing monologue to a far more personalized digital dialog. The future of digital experiences are driven by the customer providing direct information to a company via search, chat, or other methods of conversation. This dialog creates a flywheel effect, where each experience helps optimize the knowledge that is available and primes us to build that next best experience. For example, at our customer, Cox Communications, we recently had the pleasure of participating in their digital summit. And we examined different types of questions users were asking through their search experience platform, which is powered by Yext. Cox Communications stores almost 60,000 different entities in their Knowledge Graph. And leveraging our AI-powered cluster analysis, we identified that there were hundreds of questions on how to forward phone calls, just in the previous 2 days to the conference. Now, if I ask an AI to write an answer to this question, as was demonstrated before, the results are a little disappointing. If we used a third-party AI-generated content at the conference to show them, and it generated this, on how to forward one's business phone, this was the question that was constantly being asked, and it generated the content. But everyone at the conference very quickly realized that this answer, while interesting, is, in fact, completely wrong. This is an answer, if you look closely, for -- remember the 1990s and 2000s office desk phones with the CW forward-all button? That's what it went out and found because it didn't have the authoritative data to answer this properly now in a much more current time frame. So alternatively, inside the Cox Communications' Knowledge Graph from Yext, if you ask that same question, you get the exact direct answer from them. So to hear more about our Cox Communications partnership with Yext, I'd like to welcome Jim Robinson, the AVP of Digital Operations of Cox Communications to the stage. Excellent. Good to see you, sir.

Thank you. Thanks for having me.

So first, thank you so much for being here and joining us, and it was a pleasure to be down with you all a couple months back.

First, to start off, I'd love for you to just tell everyone about yourself and your role and your mandate.

Sure. So those of us that are old enough to remember the dawn of the Internet, there's this thing called the webmaster and my organization is what a webmaster looks like in 2023. So it's a little bit of a junk drawer. We have UX and a lot of our content. And then, obviously, you know, tied to Yext and our partnership, SEO falls under my organization as well.

Excellent. And so knowing that it's all of these different digital areas, could you tell the audience a little bit about Cox Communications? Very large organization. We love working with you in this partnership, because you do touch so many different parts, different customer types. Tell us about Cox Communications.

Yes. We are privately held. So that means that we kind of fly, I think, below the radar a little bit, but we are the third largest cable television provider and fourth largest cable Internet provider. So we have fairly substantial markets in Phoenix, Las Vegas, Orange County, really across the U.S.

Excellent. And so we've been in partnership since about 2015. And in that process, that 8 years, you all have been one of the earliest adopters of each of the different technologies. So take me back to as we started working, what was the process for engaging with Yext and how we work together.

Yes. Unbeknownst to my team at the time, we were leveraging your listings product quite successfully. And that -- big companies kind of being what they are, was siloed away kind of with our retail stores, which makes perfect logical sense. When I first started with Cox on the digital team, we didn't really have any kind of established SEO discipline or practice. And so it took us about 2 years to clean up a lot of our technical SEO stuff, URL structures and page, crawl budgets and all that good stuff. And we had just gotten to the point where we were starting to look to exploit like all of the work that was kind of foundationally done and at the same time, just kind of coincidentally, we are primarily an Adobe shop. And so their tool, Search and Promote, which was kind of their internal site platform, had gone end of life. And so we had been shopping around for a new tool and Yext had started to kind of come up on our radar for that. Also, your local pages and -- or pages, the local pages that we use had hit our radar from an SEO's perspective. So as we started to work with supply chain, just ironically, left hand meet right hand, they're like, yes, we already have a relationship with these guys, and we had called you in to do a demo because Answers had just started to kind of become a -- you, I think, were even in beta, to be honest.

Yes.

Yes. And so it was just a perfect kind of timing and opportunity, and we needed it, and you guys had had it ready to go. So...

Well, and knowing -- so I want to say that was in 2019 as we started the process. And you utilized Answers both for -- the Answers platform for Search for the residential or the retail customer and the business customer. Tell me a little about working within the Yext platform. So how you use the platform itself.

Okay.

Because we have the Pages, we have the Listings, we have the Search.

Yes.

How do you leverage all those different pieces?

You know, you talked a lot about the Knowledge Graph, and I was a little bit disappointed to hear that you guys have 160,000 entries and we only have 60,000. So we're going to have to do better there because I feel like we should have at least as many as you guys. But my team is one that supports both lines of business. So we support residential and Cox business. Obviously, as the org goes, there's different teams that focus on that, but we really see the Knowledge Graph as kind of the underpinnings of this whole thing. And, again, kind of as we got started in this journey, didn't maybe realize the true power of that, just at first glance. But as you get more familiar with the product and the tools, you really see that that is kind of the linchpin to this thing and that Knowledge Graph really does power everything from our listings experience to our Answers experience, to our pages experience. Our focus from my team is primarily on Answers and the Pages. And it has allowed us -- the Knowledge Graph and the capabilities of the tool have allowed us to kind of segment different responses and different content to the type of audience that we're addressing. So if you think about the use cases that we have for a noncustomer that are searching for Internet service or TV service, those questions are really around price and availability and maybe what channels and speed. That's a vastly different use case than what a current customer would have, and so the power of the platform enables us to kind of pivot on several axis points, around customer versus noncustomer, even locality. A lot of our prices are different depending on the market. And the Knowledge Graph and the tool itself gives us a flexibility to kind of address those use cases and make sure that the content is relevant to the customer and that we're answering their questions.

Excellent. And I think 2020, we decided to do a case study together in partnership with Cox and couple of interesting data points off that, that you had 51% increase in site search conversion rates and a 59% decrease in repeat on-site search, meaning they had to refine it or type something else in. How does that impact the business when you can show percentages like that?

Yes. I mean I touched on it a little bit like from -- not just from a Yext perspective, but from a search perspective, generally, on our site. We have really tried to make strides in relevancy and trying to answer the customer's question. Cox being the brand that it is, it's pretty easy that Google's just going to promote some content of yours. That doesn't always mean that it's the right content. And I think you see that in those statistics. Right? We were serving pages because Cox Communications was in the query, but it wasn't really relevant content to the customer. And so that leads to frustration on their part. They may be getting content that is not necessarily germane to whatever question they're asking. And it's just a bad customer experience. And so from our perspective, search -- our organic search is some of the best converting traffic that we have from a sales perspective. And then, when you look at the current customer site, a lot of that is about call deflection. And so we're making -- if the customer gets the right answer and we're giving them the content that they need to solve their problem, that's a call that we don't have to handle. And realistically, most people don't want to call anyone, right, these days.

Right.

Like they just want their question answered. And so if we can make it simple for them, I think that just holistically, the better relevancy is somewhat of a halo effect for the brand, especially for somebody that's providing kind of a technical service. Like you want somebody that appears to know what they're talking about. And so that just adds credibility to us that, yes, we can answer your question. We can get you the information that you need and that you're looking for.

Well, the beauty of that is this is really how trust is built between the brand and the consumer, which is, as I ask questions, if I get good answers, great, let's keep the conversation going.

Yes.

Very much like if you walked into a store and talked to a human.

Yes.

So that obviously adds trust to the entire process. So when we were at your summit, we talked a lot about what was coming out in the spring release. And I'm thrilled with the spring release with generative AI, with the new Yext chat system. So knowing that you have this graph with this centralized knowledge, what are some of the things you see going forward? So how this platform can help in terms of utilizing these new release tools or these new options?

Yes. I mean we talked a little bit about this backstage, but I feel like the kids that like the Indie band before they got popular. And so we've been talking about this for 3 or more years now, and always trying to -- is it ready yet? And is it ready for prime time? And I think finally, we are on the cusp of being able to really put production-ready use cases out. And we've got some things that we've been working on. But from my perspective, being the operations guy, a lot of my world is consumed with efficiency and scale. And so internally and just human nature, like people think of innovation like Meta and the VR glasses and things like that because that is cool and hip. But there are real bottom-line use cases to being able to scale this type of stuff. And for us, it's difficult because we know the value of personalization. We know the value of one-to-one marketing. We know the value of these segments. But the reality is there is a return on investment equation in terms of like what my team can support for those use cases. And so AI really just kind of tosses that out the window. I can scale now almost infinitely. Obviously, there's some business concerns. And Michael, I think, did a good job earlier of showing like, hey, this thing is not always 100% right. And so we're going to work our way into that and get comfortable with it. But I think the product that you guys have, the way that it kind of creates somewhat of a walled garden gives us that safety net that you're not just out there just showing everything that's on the Internet. And so those capabilities, I think really open up just a whole new universe for us in terms of being able to scale personalized content. I mean, and that's just the easy button. I think, again, we're on the cusp of the early days of the Internet, where everybody was just -- your mind couldn't even work fast enough to put all the use cases together. And so we see opportunities and quality score with paid search landing pages and the list goes on. So there's a lot of stuff that we can do here. And I don't even think we've not even scratched the surface. So very exciting times.

It really is. And I can say from working with you for years, I truly appreciate the partnership. I very much appreciate you being here today. And I look forward to building those next few fun toys with you. So thank you again.

All right. Thanks.

We're now going to begin our Q&A session for the first half, and I'd like to invite Mike Walrath, Christian Ward and Marc Ferrentino back to the stage. While they're assembling, if you do have a question from the audience, please raise your hand. I'm going to bring the microphone to you. And please ask your question into the microphone for the sake of the folks that are listening in on the webcast. All right. Who has a question? Here we go.

Great. Tom White, Davidson. First off, thanks for doing this, guys. Really appreciate it. Maybe one for you, Mike, a higher level one. When you rejoined or, I guess, took over the CEO role, I think one of the things you talked about was trying to kind of simplify the customer experience, simplify the go-to-market.

Yes.

Flash forward a couple quarters in this explosion in AI as a topic and kind of just a dynamic backdrop there, can you talk a little bit about how that may or may not kind of complicate things for you in terms of kind of simplifying the business versus kind of weighing the opportunities that that creates.

Yes. No, thanks for the question. I think look, if this had happened 2 years ago, I'm not sure I'd be sitting here. I think what we've been working on and what we pivoted towards, and I've talked about this a lot, very aggressively in 2019 and 2020, was the expectation that AI was going to explode. Right? So we were super way too early on that. And I think it caused a lot of the challenges that we've seen, which is the conversation that we've been trying to have with customers around creating digital experiences, using AI and the burgeoning power of AI to do this, has been very difficult to have because they had -- there was just not a lot of interest in it. And you heard some of this. You heard complaints about I don't know why you're in here trying to sell me something that doesn't exist yet, when you have product that does in terms of Listings and Pages. And so, as you know, hopefully, what you're beginning to see is as we get through this cycle, how these things work together and how they fit together in that what we're doing today is, I think, we're taking another step in the articulation of how Listings, Pages, Reviews, Search and the idea of digital experience, the idea of a next-generation content management system, they all do fit together. And so I think a lot of that kind of wandering around in the woods of expecting this to happen faster was being well ahead of where the market was ready to be and having a really hard time explaining how Listings could possibly be related to support search or site search or things like that. And I mean there's no -- I don't think I'm telling anyone anything they don't know to say that, that was a struggle for us. What we're finding today is that when we present the type of discussion that we've shared so far today, and obviously, this has been very product-centric with customers, instead of blank stares, what we now get is a lot of head nodding, a lot of "yes, you're doing a good job of describing my problems and a good job of describing the challenges that I face, and, oh, yes, my CEO is very worried about what we're going to do about this AI thing because it feels like opportunity, but it also feels like risk." So to me, this is not a -- what we're doing here today and what we're doing in the market isn't a big bang, new release. It's not new messaging. It's the logical progression of our message. And hopefully, it makes things clearer in terms of how we're going into the market. And you'll hear more, obviously, in the second half about how we're taking it to market, marketing, go-to-market, things like that.

Please say your name and your affiliation into the microphone. Thanks.

Neil Gagnon, Gagnon Securities. I'm struck with the Cox example and wondering how much of the opportunity is with new logos or doing business with the existing ones you have?

Did you look ahead in the presentation?

Sounds like he did.

We're going to talk a lot about that in the second half. I would tell you both, and I think there are huge opportunities with the new logos and also with existing customers.

Yes, just to add. I mean, think about it. Every single customer that's using our Listings product, Listings sits on top of the Knowledge Graph. So they're already using a portion of our headless CMS system already. And that's really where the opportunity -- that's a big part of where the opportunity lies, in expanding the number of use cases, the number of channels that they can manage while already having that content management system that they're using today.

And I mean, in fairness, I told you in the first call, there was a little bit of taking a step back here, right? We had to get back to the basics. We had to reconvince our customers that we were committed to Listings and that Listings mattered to us. And I think we had to go back and start over and explain why Listings matters to us and why it's an incredibly important part of this. I mean, when you talk about the fragmentation of the digital experience, a lot of that's happening off your own sites. And so having the capabilities to do that on third-party experience is going to be an incredibly -- which is what Listings represents, amongst other things. So there's a big opportunity in both categories, and we'll talk about it a bit more.

Ryan MacDonald with Needham. Just starting with I thought it was interesting in the Cox conversation about the number of entities that was in their Knowledge Graph versus the Yext's. And it seems like as ChatGPT and generative AI continue to be utilized, that you're going to need to grow the number of structured facts with your customers. How do you sort of drive that conversation and maybe get customers comfortable with that so that they can take advantage of sort of the fuller platform?

Yes. I think it's -- what you don't do is you don't walk into the organization like Cox and say, "okay, we're going to take every piece of data right now, and we're going to put it all on the graph." That would scare the hell out of them. And it actually -- I think someone might have done that once or twice in the past. So ultimately, what you're doing is the best way to approach and the way that we're approaching it is you take it use case by use case. So today, it's site search; tomorrow, it's support search; and the next time, it's knowledge base, the next solution may be a partner portal and so on and so forth. And then eventually, over time, the knowledge begins to build up in an incremental fashion. There is a virtuous cycle, of course, that happens here. There's the virtuous cycle of the customer or employee interfacing with the different experiences, just feeding intent and requests for knowledge into the system, which then, of course, the response to that is to generate the knowledge or find the knowledge inside of the organization. The -- also, the other thing that happens over time is that you become the place where all the knowledge is. And that has obviously a virtuous cycle that's driven there. And so the next project that come along is the next use case they need to solve, it becomes sort of very easy to just solve it on the Yext platform.

Yes. I mean I'd just say Christian, and Christian may want to say something about this, but the thing he showed you with the clusters and the analytics on the search, there is no more -- there's no better way, from my point of view, for a customer to realize that they need to put more information into the Knowledge Graph, than seeing that they have questions being asked that are not -- that don't have answers.

Yes, one of my favorite parts about working so closely with our clients is it's usually a few weeks into the launch, and they're starting to get the data in from the customer journey. And we're sitting, looking at that screen I was showing you, and they kind of cock their head and go, "huh, I didn't think people would ask us that." And it's enough traffic to then realize that there's a huge opportunity if you can answer that. And so that cycle really accelerates but it's one of the best parts of actually sitting and watching customer engagement to start to build the prioritization of the next entries into the Knowledge Graph.

Rohit Kulkarni from ROTH MKM. Thanks for the demo. I think the demo was very cool. Whoever made that was very cool. I think philosophical question for you, Mike. I think you said companies cannot force revenue growth and balance efficiency at the same time. So you have this new, very large potential opportunity in AI. How do you balance growth versus investments in AI as the year progresses, as the next 2, 3 years progress? And then maybe for Christian, I think you've been at the cusp of BERT and all LLMs over the last so many years. Next -- maybe talk about next couple of years. Where do you see this acceleration in AI kind of adoption evolve into? We saw some use cases here from CMS and you have a CMS and you can create a lot of new use cases. How do you see that evolve in terms of not just internal but external CMSs that all these companies can have?

So I'll take the first question. So philosophically, I mean, you notice, I chose my words there very carefully. I looked down and I read them off the monitor because I wanted those words to -- this is the English major in me. I wanted those words to be specific. And force is the key word in that sentence. So if you take anything away from today, from a go-to-market standpoint, a selling and marketing standpoint, it's this, that the attempt to force growth when the conditions for growth aren't good, aren't there, is what creates inefficiency, right? And so there's an acceptable amount of inefficiency when the conditions for growth are right. And you're going to hear Tom and Raianne talk about what the attributes of this are. And clearly, businesses that are growing faster, they can afford to be a little less efficient. What we can't afford to do is out of a sense of desperation to drive the growth faster than it's accessible to us is try to force it because what winds up happening there is you're going to see our expenses are going to go up, particularly from a selling and marketing standpoint, and the growth doesn't materialize because the conditions aren't right for it. So a lot of what you're going to hear me talk about as we go forward is trying to help everyone see more about how those conditions are developing and how things like productivity and retention and pipeline in the macro environment all have to work together to get to a place where you can have those investments actually pay off. When it comes to -- and Darryl will get into this in his section, when it comes to the investment that we made in AI, and we've talked about this the last few quarters, we're increasing our R&D spend. We're making the core investments that are required to stay ahead of the market because we're way ahead of where I think any competitor is when it comes to being able to deliver the types of experiences that you heard Jim and Christian talk about and many of our other customers. So philosophically speaking, I think if you take anything away from that statement, it's know that we're not going to overspend to force growth that's not here when it comes to selling and marketing, but that we will continue to invest in R&D to make sure that we stay way ahead of the rest of the market. Does that makes sense?

Excellent. And on your second question, I think Michael, our Director of Data Science, did a great job of showing this concept. He had a slide, and it was from March and he said it's already out of date. I think since you've been sitting here, it's out of date. The number of refinements to these models is just going to keep accelerating. I actually had that slide in December, and there was half as many entities in that large language model construct. And so what I think you will see, and it's definitely already happening, is continued refinement around particular use cases, right? So we're going to see that. There's the health care version, the financial version, all these different elements of making sure, because, as Michael demonstrated, what you train these models on is so critical to what they become. And so if you're looking for a model that understands a particular vertical, let's say, or let's -- how about UConn. Like if you went to UConn, you want a great knowledge interaction, you're going to have something that's trained on all the knowledge that we know about UConn and their dynasty, right? So when you go through this process, a lot of it is where do the models really add that differentiation? And then in terms of the next several years, it's everyone trying to identify the use cases that match to the right models. And I think that's actually quite an interesting area for us to examine, which is different models may touch at different points for different questions and different digital experiences. And so leveraging our platform as a center point of the knowledge that's available to use, but how that's used in each digital experience is going to be an area that continues to develop very quickly. I'm not sure, Marc, if you have anything to add?

Yes. I mean the only thing I want to add is just sort of, and Michael talked about this, the way we look at this and the way that -- the very simple way is I think people have been using these models wrong. They've been using them wrong for a little while. They are using them as a database. They use them to answer questions. Mike's example in the beginning about his bio is a great example of that. That is not how they should be used. What these models are really, really good at is reasoning. They're really good at instruction. They're really good at orchestration and generation once you give them the right facts. And so what you see, what this will evolve to over time is that everyone is going to start to figure out they're using the models wrong because they're fun. It's fun to have it write a rap song or something, but that's not really the point of it. And that's really where we see this going is we talked about that high-level architecture of the data layer and the AI layer and the experience layer. And that's where I think that will become the standard architecture of computing going forward in almost every realm of computing. And I like to think that we have a nice head start on it because we were able to see it earlier than most folks. And you're starting to see the benefits of that foresight showing up in the platform as we've showed in some of the demos today. And there is obviously more that these things can do.

Chet Kapoor from Tenzing. I'm curious to understand if you can differentiate the Mike's bio example relative to the Telluride blog post example, because as I understand, that wasn't pulling from a database of information about Telluride. So why was that able to pull data?

Very specific database.

In the case of Telluride, we were able to ask the model to write a blog about a certain topic. So we gave it the topic. It wasn't completely arbitrary. That one could have easily written a bad article, which is why the human being had to be involved, why he showed us an example of me inspecting it, looking at it and sort of fake reading it in the demo. But the reality is that, yes, someone would have needed to look at that. What we weren't able to show because we didn't have the time, but we would love to show it in more detail, is when you combine the graph with the actual models, they write very factually accurate articles and descriptions and things of that nature. And that's part of the feature that we just rolled out with computer fields was the ability to inject graph information directly into the creation of the generation process. And that's something that is completely up to the person configuring the platform. You can even go a step further and have it write in a different tone of voice and run and capture a certain brand uniqueness in how they write. But those are all sort of aspects of combining the 2 together. So if you wanted to write something factual like that, where maybe there needs less inspection or less approvals, you would want to make sure that it's writing from the data set itself.

Bob Johnston with Herald Investment Management. Could you follow up on your comments there about developing more of a brand? I mean do you feel that Yext helps do that enough now or will AI help take that forward and really differentiate how one customer uses the Knowledge Graph or other parts of your platform versus others?

You're saying, our brand or you're talking about their brand?

Their brand. I mean to what -- are you giving your -- currently, giving your customers the tools to really differentiate the brand they want to rather than just be able to check boxes that they're providing enough detail in the way that they want to do it, but not necessarily in the brand image.

Go ahead.

And so I mean, at the core of what we're offering them is a development platform, right, a platform for them to build the brand and experience of their choice, something that fits their products, fits their buyer, fits the market they want to be in. So there's lots of play for expression, lots of places for expression for that company, that business who's using the platform. I also look at brand and I say, on some level, brand is the sum facts of who you are sort of makes up your brand in many ways. And so as we're able to capture sort of all the information, all the facts, or as many as the company wants to put it in the graph, that will allow for a unique experience in itself. There's also lots of fine-tuning that comes into this where you can start to capture brand voice as part of this through the AI so you can show it. For example, a bunch of reviews that maybe you've written or kind of articles that have been written by your organization in the past and you can say, okay, well, this -- please emulate this and, please, this is what we want it to look like and how it to read. And so that gives another opportunity for each business to sort of uniquely express themselves on the platform.

Yes. I think part of the brand piece to this is if you -- Marc makes a great point about training models on brand voice so that they actually use your brand voice. And you start to think about use cases like chat and as they get more robust and they develop. The old problem with chat was that you had to basically preconfigure the answer to every question and it was a pretty limited universe, right? And so the experience wasn't very good. The new problem with chat will be that if you -- which is what most companies are doing, you put a skin over OpenAI and you put it up as a chatbot, it will answer every question. Right? It's not going to not answer any question. So the old problem was the chat didn't answer any questions. The new problem will be it answers every question. Right? And it gives you the answer whether it's true or not. So you can start to think -- when you think about the brand implications of allowing AI into your business without constraining the way that it answers, the way that it talks, the answers that it gives, imagine some of the questions that you could ask a brand's chat experience if the AI that's answering the question, the generative model, the placing of words that's answering the question, is unconstrained. Like would you like to -- it's a silly example or maybe not silly, but if you put an unconstrained AI chatbot on a website and I asked the chatbot to tell me why the management team is racist, it will. Right? I mean it's -- it will do it, right, if it's unconstrained. And so your brand has now basically published on the public Internet why your management team is racist, right? That's the downside of not using these technologies correctly. And I think people are just beginning to understand how big some of those downsides are. So as much of what we're doing, as Marc described, as a development platform, it's a development platform that allows you to utilize these things, but creates a brand safety that is going to be paramount to anyone who has -- does that make sense? Does that answer your question?

Yes. So putting in guardrails but not necessarily -- just putting in guardrails, but not necessarily enabling creativity or anything else as to how that data that you're passing on to your customer is being used?

Well, it's like Marc said, the way it should be used is for the creativity, for the semantic understanding, for the generative creativity, but it should be constrained by the facts. Right?

It shouldn't create facts.

Right. It shouldn't make up...

It's a pretty creative experience.

It shouldn't create a -- make up facts. So another made up example I like to use, you imagine the world of content generation, where, if I own 100 hotels in Southern California, and I can create infinite content about what do you do in Southern California, and I can do it in my brand voice, but not necessarily with my brand attached to it, but every one of those experiences is tied back to something you can do at one of my hospitality properties, that's an opportunity to create an unbranded brand experience, if you will, at scale, using the combination of generative AI technology and the facts that exist in the structured Knowledge Graph. You can brand it or unbrand it.

Okay. That concludes our Q&A session for now. We're going to take a 10-minute break. [Break]

I'd like to welcome to the stage, Tom Nielsen, Chief Revenue Officer.

Thank you, Mike. Okay. There's 3 main areas that we'd like to explore today with our revenue organization. First, how we're organized in FY '24; next, the big bets that we're going to be making around operational excellence and growth; and lastly, when it's time to accelerate our investments in growth again. So how we're organized in FY '24. Beginning in FY '24, we now have a global, unified go-to-market team. We're organized between our enterprise vertical and geo businesses, along with our mid-market teams. This also includes our reseller channel, which is a change this year, that now rolls up to our geo businesses in both North America and EMEA. We have flattened the organization this year in North America, essentially bringing our AEs, our customers and our executive teams closer together. Previously, we had 4 layers of VPs in sales. And we've collapsed that to 2. In Japan, we've implemented a partner-first model, where our direct sales teams are transitioning the renewal and the net new business over to our expanding partner ecosystem. And overall, what we've seen so far is more communication and collaboration between the teams. Secondly, we're making 6 big bets in FY '24 around operational excellence and growth. Now I understand when we're talking about operational rigor and inserting more discipline into a sales process, it's not always the most exciting topic, but we needed to act now if we were going to modernize around new responsibilities, around our move into DXP and to CMS and selling a full platform that customers want to build on. So our alignment and roles. We've discussed the alignment. And with our roles, we've had a change with our AEs, where the AEs now own all of the commercials with customers. So this means our renewals. This means our incremental ACV and our net new logos. So, so far, it's been well received by our customers, mainly from the CSM side, where customers want to see CSMs as trusted advisers. They want to see them focused on adoption, not necessarily dragged into the difficult quarterly renewals and negotiations that were happening. Also, this has given our AEs the opportunity to focus on the renewals and use the renewals as a compelling event to grow our incremental ACV. Secondly, we're going to be 100% data driven in FY '24. This means instrumenting all of our critical sales processes with KPIs that are measured weekly. The 2 primary measurements will be retention and incremental ACV. We have also put in place compensation plans that incent our AEs along these 2 metrics. Our CSMs continue to get compensated on retention. Other KPIs that we have aligned on are pipeline generation, new logo acquisition, early renewals and the optimal staffing levels. Next, we have a renewed focus on pipeline generation. So traditionally, at Yext, this was solely the responsibility of the AEs. Now PG is the responsibility of the entire company with sales, marketing and our partner organization all holding KPIs. Our sales team is aligned on targeted sales plays. Our marketing BDRs now have clearly defined goals and objectives and the investments that we're making in our partner org will begin to pay off in the latter part of the year. This year, our teams have extended their use of a value-based selling framework, because we know what our productivity numbers look like when we use our sales tools. We also know the outcomes when we lead with a point of view on our value drivers of gaining operational efficiencies by building on our platform, generating engagement and driving conversions. This has all come together into a common proposal framework that has increased the speed in which sellers provide our customers with compelling commercials, especially now that they have responsibility for the entire renewal book. A critical part of capturing our opportunity this year is our ability to sell the entire platform. We've enabled our sellers on more creative deal making that's aligned around land and expand and tool consolidation motions, along with doing many more multiyear deals with our customers. So value realization and the post-sale customer journey will extend the momentum that the team's built in the last 2 quarters of FY '23 and the positive impact that it had on retention. From our initial post-sale hand-offs, to implementation, to adoption, to growth, we're truly creating a flywheel effect between our AEs and our CSM organization. And all of this is wired together across our 6 big bets in creating a cross-functional winning team culture. Lastly, we'd like to discuss the criteria around when the time may come that we're going to accelerate our investments into growth again. So now that we have a lean and flexible organization and we're implementing our plan for efficient and sustained growth, let's discuss how we're measuring when it's time that we'll accelerate growth further and begin investing in new quota-carrying resources. We are doing this through closely measuring the relationship between sales productivity, pipeline generation and product innovation. Increasing AE productivity is the most efficient way that we can grow ARR without incurring much additional expense. We are seeing productivity gains and conversion gains through the adoption of our sales tools. However, we know that this curve will eventually begin to flatten as there's only so much and a natural limit that AEs can produce. So increasing productivity, singularly, is not enough for us to begin adding additional AEs. Next, we're also tracking pipeline generation into this equation. Our sellers have quarterly KPIs, and they're much more disciplined in how they're executing sales plays. We have grown our BDRs and we've implemented a lead flow process that Raianne is going to talk more about next. Also, we continue to invest in our partner organization, as we said, and expand the relationships with SIs. This is a foundational year for our partner team. While it's still early, too early to tell in our company-wide approach to demand generation, the results so far have shown very high-quality leads and faster progression into qualified pipeline. We don't expect to see the same limits or the flattening with pipeline generation as we do with AE productivity, and this should resemble more of a hockey stick type of curve. Now, we've added product innovation into the decision equation due to the impact that it may have on both sales productivity and pipeline generation. We expect our Spring release and move into the DXP and CMS technology category to be a leapfrogging event. It's simply a much higher spend and essential technology category in IT organizations, where we have never had this level of capability in previous releases. This, along with the early responses that we have received from customers with AI chat and content generation, makes us think it's going to have an exponential effect on the pipeline. Let's double-click into full platform selling for a moment and its relationship to product innovation. So with the confidence that we now have in our platform against the competition, we'll now expand on our full platform selling efforts because it's the opportunity that we have to capture this year. If we were to tell you that 76% of our direct customers have 2 or less products, you'd be inclined to agree that a large opportunity likely exists within our installed base. Now we're still very much focused on logo acquisition as [Audio Gap]. However, sellers and CSMs this year are enabled and expected to pursue tool consolidation, broader platform adoption, motions like that, respectively. Now Darryl is going to discuss this opportunity in a little more detail later in the program. So in summary, there's no single one event that's going to cause us to increase and invest more into our sales engine, but rather a combination of us reaching a limit with AE productivity, sustained increases in pipeline generation and the market reaction to our innovation. All of these together are going to signal the execution and the excellence that is expected at Yext. Now, one of the benefits for me of starting at the same time as a new CMO is we've had the opportunity to reset many of these things together. Now, one of these is how we work together as one go-to-market team. So it's my pleasure to introduce Raianne, who's going to talk about how marketing is helping to drive this support.

Thank you. I'm Raianne Reiss. I lead marketing at Yext. I'm going to focus on 2 main areas today. The first one is a recap of my first 6 months in the role, what observations have I had, what changes have we already made to put us on the path to success. The second area I'm going to cover is what are we focused on within marketing to drive growth for the business. It's hard to believe that it's already been 6 months for me at Yext. That old saying of time flying when you're having fun, definitely felt true for me here. We've made a lot of progress in a short amount of time. It's been super rewarding to see small changes have a really big impact. Shortly before I joined Yext, the marketing organization was actually 6 different organizations, reporting to 6 different leaders. We've done a lot of work to bring the marketing team into one integrated team. We've broken down silos, we've reorganized the team, and we've made sure that every single role is focused on the most impactful work to drive results for the business. As a part of this work, I've upleveled the talent and brought in a new senior layer of talent. We have new leaders and functions like demand generation, content marketing, partner marketing, brand and communications and the website. It was obvious to me, pretty quickly in my tenure, that the sales and marketing organization were not as connected as they should be. Sales had, frankly, all but given up that marketing would have a big impact on the organization. And marketing seem to be driving activities in a vacuum. There was a lot of stuff going on in marketing, but no one could articulate the impact of most of that work. I have a great partner in Tom. We share a common vision for what connected sales and marketing organization should look like, operating as one team with one set of shared goals. We're a lot closer to realizing that vision today than we were just 6 months ago. Speaking of shared goals, the marketing organization that I walked into was determining success based off a set of highly subjective measures. But I also found a team that was hungry for real goals, tangible goals and metrics. Today, we track over 100 KPIs across the marketing organization. Every single person on the team understands what their measurable goals are, how they ladder up to the marketing organization's goals and how they ladder up to the company goals. Every person in marketing now feels accountable for the success of the business. Six months ago, we had a marketing organization that was heavily focused on branding activities, messaging work, brand campaigns, acquisition of new logos. There was far less activity occurring that was focused on our customer base around how we would drive consumption and cross-sell across the platform. Today, we still care a lot about driving brand awareness. However, we are spending a heck of a lot more time on things like driving pipeline and growth. More about this in a bit. It's safe to say that just 6 months ago, marketing was having a much smaller impact on the business than we should be. One of the areas that we've been heavily focused on, as you've probably guessed, across the marketing organization, is pipeline and generating pipeline for the business. In one of the first meetings I had with the team, I painted a vision where we would move from a model of operating really as an in-house marketing agency to one of a growth driver for the business. It's a vision that the team got really excited about. We've made a lot of progress, and we're getting closer to executing on that vision. One of the areas of progress that I feel good about is rebooting the marketing demand gen engine. And when I say we rebooted the engine, I mean a literal reboot. We went to the ground floor of defining our lead flow processes, taking a look at the tools we were using, how those tools talked to each other, our lead scoring models, pretty much everything on the operational side. But in addition to the operational side, we also rebuilt our marketing demand engine from the top of the funnel, all the way to the bottom of the funnel. The activities and things that we were doing to target and identify new contracts, attract them into the top of the funnel, engage them and accelerate their buyer's journey. We've also aligned on a consistent messaging strategy and put mechanisms in place to make sure that we're delivering consistent messaging throughout all of our integrated marketing channels. We've implemented a new campaign framework. As a part of this campaign framework, it's really designed to drive alignment across the whole go-to-market organization and what our priorities are for the business for that year. Included in our new marketing campaign framework is a focus on reaching new audiences. One of those new key audiences is IT and developers. Those audiences are becoming increasingly important to us in the platform space. We're going to spend a lot more time focused on attracting and bringing these new audiences into the Yext family. On our path to accelerate growth, we took a hard look at the BDR organization, and we made some pretty substantial changes. The BDR org now has full accountability for pipeline, they're fully integrated into the marketing engine and really working as one comprehensive team. The changes that we've made in the BDR org, combined with the reboot of the demand engine, are already starting to show really positive signs. As was mentioned earlier, it's only been about 6 weeks since we turned this on and I'm hearing really great feedback from the sales team on the quality of the leads that we're producing, and that's always a really great sign. And last but not least, is turning on our partner ecosystem, from our alliances to our cloud hyper-scaler partners, to our systems integrators, we see a huge opportunity to go bigger here and drive co-marketing campaigns to develop mutual pipeline. We're going to be spending a lot more time focused on integrating our partners into our marketing demand engine. It's been a really great first 6 months for me. I'm super excited about the progress we've made, the changes we made and the opportunity ahead of us. I'd love to turn it over to Lexi Bohonnon who is going to -- the EVP of Global Client Success.

Thank you, Raianne. It's great to see all of you. I lead our customer success organization, which starts in the presale with our solution engineers, our solution architects, demo engineering and free trials team. And then our day-to-day customer solution management team, our CSM and, of course, our services and our support organization. As you heard from Tom earlier, a key go-to-market theme over the last year has been on organization and unification of our revenue team. And as a more unified and efficient revenue organization, we're certainly more equipped to better deliver on the expectations of our customers and deliver value at each step of their journey as trusted advisers. Our goal over that last year has really been to simplify the customer experience by eliminating those silos we mentioned and also upskilling the talent across all the various departments within CS. We're really proud. We've worked hard over that year to really enable the CS team to have more confidence in articulating the value of our platform and work better hand-in-hand with our customers each and every day to drive adoption, optimization and identify new use cases for them to use the Yext platform. As you also heard from Christian Ward and our customer, Jim Robinson from Cox Communications earlier, we are solving so many different types of challenges now from the Yext platform across different personas, use cases, et cetera. And so now to share a little bit more about her experience over the last couple of years working with Yext, please join me in welcoming the Director of Digital and e-Commerce at United Rentals, Courtney Versteeg. Welcome, Courtney. All right. Thank you so much for being here, Courtney.

Thank you for having me.

Yes.

It's a pleasure.

Can you start by sharing a little bit more your background and tell us about your experiences and your professional career?

Sure. So I'm Courtney Versteeg. I lead our digital marketing and e-commerce team at United Rentals. For those of you not familiar with United Rentals, we are the largest construction equipment rental company in the world. We have over 1,400 branches. I have been at United Rentals for 3 years leading the team. And then prior to that, I spent 9 years at GE in a variety of different businesses and different marketing roles.

Awesome. So United Rentals has been a customer with Yext for about 4 years now. Help us understand how you all use the Yext platform and how it helps to support some of your big organizational goals.

Sure. So Listings is really our bread-and-butter product that we use for Yext. As I mentioned, we have over 1,400 branches. Those branches operate across about 10 different businesses. We offer a variety of different equipment rentals and services at each of these branches. So it's really important for us to have local discoverability because our customers range, right, from our small contractors to our 2 large construction equipment -- sorry, construction customers. So we didn't have great discoverability, so we were really looking to partner with Yext on getting those branches visible and also being able to operate centrally at scale.

Awesome. Help us understand the first few years of our relationship together. What were we able to achieve? And what were some of those early wins?

Yes. So we were, again, very excited to onboard the Listings platform within United Rentals. We saw some great results in improving our metrics around map views, calls, website visits, and we were able to manage this centrally. We really didn't want to have our branches have to manage their own Listings and automation was also key, right? So whenever we opened a new branch, we wanted to be able to have a template, be able to have that branch launch quickly, so that way, the branch can already start getting local discoverability.

Thank you. So we had some success in the early years, but a few years ago, you noticed a little bit of a shift in our relationship. So help us understand what changed, and how did you share that feedback with us?

Sure. So something that was really integral, when we partnered with Yext originally, was we had a very strong customer success manager from Yext. He was extremely technical. He spent time getting to know our business, was extremely prepared, really helped us automate and really launch the platform. Unfortunately, there was some churn, and we lost that level of expertise within Yext -- from Yext, that really did kind of put us a little bit on the back burner. And we also felt that we were such a big investor in the Listings product, we weren't hearing about new features what Yext was investing in with Listings because, again, that was so important for our branches to be able to be discoverable on all search engines. So we escalated that to our account management team and SVP of customer success.

And how did the team respond? How did the team take the feedback and kind of help us understand what changed?

So I have to say I was quite impressed with how that happened. Within 2 to 3 weeks, we had a new support person who really spent time getting to know our business, understanding the setup that we already had, was extremely technical, reimplemented quarterly QBRs with us, support -- weekly support meetings, biweekly strategy meetings. So we felt that that really was really a change in our partnership. And additionally, we started hearing about new investments within the Listings platform. Again, for us, that is our bread and butter. That increases our discoverability. We want to be able to get as much content into Listings. So that way, our branches are discoverable.

Awesome. So let's talk about the future. What are you excited for with the platform? And how do you think we can further achieve some of those goals in the next year together?

Yes. So I'm excited, particularly for Listings because again, like that's our primary product. New partnerships that we're seeing within the Listings platform, I think there is an opportunity to really add more content, to automate some of the content from some of our systems into Yext that will hopefully help us grow our branch metrics, not just phone calls, but website visits and a lot of traffic and conversion on our site.

Awesome. Well, thank you so much…

Of course.

…for your honesty and transparency, and thanks for being here with us today.

Yes. Thank you for having me. Thanks so much, Lexi.

Thanks, Courtney.

Thanks, all.

All right. It is now my pleasure to introduce our Chief Financial Officer, Darryl Bond. Darryl?

Thank you, Lexi. Okay. I'm Darryl Bond, CFO here at Yext. And today, I'll begin by briefly reviewing our fiscal '23 performance and then talk a little bit about fiscal '24 strategic objectives and finish it up with a fiscal '25 and longer-term outlook. As a reminder, we may reference certain non-GAAP financial measures, which are reconciled to GAAP in the appendix in this presentation, which is also available at investors.yext.com. Fiscal '23 was a pivotal year for us. We had significant org changes. As Mike mentioned, we took our head count from 1,400 down to 1,100. Revenue was up to $401 million, up 5% on a constant currency basis, and we ended the year with $400 million of ARR, which was up 4% on a constant currency basis. Our direct business, which represents 82% of our total ARR, grew 6% on a constant currency basis, and we ended the year with net retention of 97%. We saw improvements in both gross and net retention during the year, and we'll talk more about this in a few moments. Our reseller business represents the remainder of our ARR and declined 6% on a constant currency basis and ended the year with net retention of 92%. Keeping in mind, reseller is the primary way we access the SMB market, which has been impacted by a challenging macroeconomic environment. We made significant improvements in profitability this year, which is illustrated by our adjusted EBITDA margin and our non-GAAP EPS. Now, since we made these improvements throughout fiscal '23, we think it makes sense to compare Q4 of fiscal '23 to Q4 of fiscal '22, to illustrate the progress we've made. We exited Q4 of fiscal '23 with $0.05 of positive EPS and an 11% adjusted EBITDA margin. Operating expense as a percentage of revenue improved from 80% to 69%, primarily driven by improvements in sales and marketing, which improved by 10 percentage points. We had positive operating cash flow for the third year in a row. We repurchased 13.8 million shares, which is just about over 10 -- just over 10% of the outstanding share balance, and we ended the year with $190 million in cash as we continued to maintain a strong balance sheet. So in fiscal '24, we're focused on 3 key strategic objectives. First, efficient growth within the Rule of 40 framework; continued focus on gross and net retention; and navigating the headwinds impacting revenue, ARR and retention. Rule of 40 is our north star metric. The past few quarters, we've seen an inflection on the Rule of 40, primarily driven by our adjusted EBITDA margins. Now, progress will continue, but in the near term, it will be driven by adjusted EBITDA expansion. Revenue growth will take time to impact Rule of 40 given our SaaS model. And over the longer term, we see the potential for more significant contribution from revenue, which will be driven by our platform expansion and the go-to-market changes that Tom and Raianne just talked about. And this chart here is showing our direct ARR, excluding our direct SMB customers. Direct SMB is $8 million of ARR at the end of Q4 of fiscal '23 and we've excluded direct SMB for the purposes of this analysis, because SMB customers generally have 1 or 2 products, and we recently deemphasized our direct sales into the SMB channel. We've segmented the ARR by the number of products the customers purchased and, for the purposes of this illustration, the products are Listings, Reviews, Pages and Search. As you can see, the CAGR from fiscal '20 to fiscal '23 was 10%. This illustrates the progress we've made on product adoption within our direct channel. Our Listings and Reviews ARR has grown at a 3% CAGR over that same time period. While Listings and Reviews was the majority of our direct ARR in fiscal '20, we have made progress with Pages and Search. This, despite the headwinds we've experienced the past few years. As Tom mentioned, 76% of our direct customers have 1 or 2 products. We have a large opportunity to drive further product adoption within our existing base of customers. This will drive revenue and ARR growth as well as net retention. Now looking at the average ARR per customer in these cohorts, enterprise customers with 4 products are 10x the size of a customer with 1 product. This represents a significant opportunity for us to expand ARR within the existing base. Customers that purchase more products generally have more data in the Knowledge Graph and show higher rates of retention. For the past few quarters, while we refocused on customer centricity, adoption and value, we provided renewal rates for our direct business, which we previously referred to as gross retention. We've seen these renewal rates recover from the low 80s in Q1 to the high 80s in Q4. Renewal rates are based on ARR up for renewal in that quarter, which is impacted by seasonality. For example, in Q4, we have bigger dollars up for renewal than we do in Q1. We will no longer mention the renewal rate in the quarter, as we believe it's not useful, since it only looks at a portion of the business and is not consistent with our software peers. Rather than renewal rate, we calculate a dollar-based gross retention rate. When we look at the gross retention on basis of ARR, we ended Q4 of fiscal '23 in the high 80s. And this methodology looks at ARR for a cohort of customers at the beginning of a period, then comparing the ARR balance from that same cohort of customers 12 months later, excluding upsells. We believe this basis is more consistent with software peers and provides a view of retention dynamics on our direct ARR balance. In addition, approximately 57% of our direct ARR is attributable to multiyear deals. Multiyear deals are a key part of our strategy, as Tom mentioned, and these will help with retention. And improvements in gross retention will help net retention, which is a key metric we're focusing on. We've made a lot of progress on renewal rates and gross retention this year. We're seeing improvement in the gross retention on single product customers, given our focus on customer value. We've seen higher gross retention from customers with multiple products as they generally have more data in the Knowledge Graph, and we can deliver more value. We've heard from Cox and United Rentals, 2 customers with multiple products that have seen a lot of success with the platform and customers with more products typically see more value and stay with us longer. Now, improvements in gross retention will help net retention and having a larger customer base provides us more opportunity for upsell. Our continued efforts to drive customer adoption provide an opportunity to bring more value to customers with more products. Now Marc highlighted the level of innovation and expansion within the platform that we've seen over the past few years. And we believe the DXP opportunity is significant. And with 76% of our customers on 1 or 2 products, we believe we have the opportunity to return to 110% net retention or more. As I mentioned earlier, reseller is about 18% of ARR and has been declining. However, we believe over the long term, we have the opportunity to grow ARR in this channel. Nearly half the resellers have multiyear contracts and resellers primarily only resell Listings. We have the opportunity to introduce additional products into the reseller channel. Now, as we mentioned in Q4 earnings, fiscal '24 will have some growth headwinds from strategic decisions we made in Q4. In total, this will impact our year-over-year growth rate by low single digits. Now first, we expect a headwind from services revenue as we are moving low-margin services work to our delivery partner network. We'll continue enabling our SI and partner network and, over time, this network will help generate pipeline. This will provide an immediate gross margin benefit in Q1 in fiscal '24, which we'll talk about in a few moments. Second, we shifted our strategy in Japan from Direct selling to partner led, as Tom mentioned, and we deemphasized our direct sales to SMB, and will use reseller as the primary means to access the SMB market going forward. And lastly, we expect the FX impact we saw in fiscal '23 to continue to be a headwind for the first half of the year, along with the uncertain macroeconomic environment we're working in and the elongated sales cycles we previously mentioned. Now these headwinds will also impact ARR and retention. In recent years, growth in our lower margin services revenue has pressured our overall gross margins. In Q4, we reduced the size of our services organization, as we're transitioning the lower-margin work to our delivery partner network. This will provide an immediate impact to Q1 gross margins, moving us to the middle of our 75% to 80% range and with continued improvement to the high end of the range throughout the rest of the fiscal year. In addition to our gross margin expansion, we also expect to increase adjusted EBITDA. The growth in adjusted EBITDA in fiscal '24 will come from incremental revenue, gross margin improvement that we just spoke about and also operating efficiencies as we now get the benefit of a full year of a leaner cost structure. And improvements in adjusted EBITDA and revenue growth will result in operating cash flow growth in fiscal '24. As growth has slowed and efficiency approved (sic) [ improved ], adjusted EBITDA and operating cash flow have come more in line. We expect this trend to continue. However, with acceleration in revenue growth, we would expect operating cash flow to outpace adjusted EBITDA given the upfront billing nature of our contracts. Our guidance for Q1 and fiscal '24 remains unchanged from what we provided during our fourth quarter earnings call. And we'll provide an update when we get to Q1. In recent quarters, we've increased our focus on reducing dilution. Stock-based compensation in terms of dollars and as a percentage of revenue has continued to decline and we expect that trend to continue in fiscal '24. Now this will be driven by lower head count compared to prior periods, as I mentioned, but also higher-priced grants rolling off. We've calculated dilution using the ending shares outstanding, excluding the impact of the secondary offering in fiscal '20 and the buyback activity in fiscal '23, and here, we can see steady improvements over the past few years. We've also calculated net issuance dilution, which is our grants, net of forfeitures, and we've seen a steady rate over the past few years. Now the delta between these 2 is primarily option exercises. And in fiscal '22 and prior, we saw higher levels of option exercises. At the end of fiscal '23, we had approximately 4.6 million options outstanding. Now going forward, we'll manage to a modest level of dilution. So now that we've spent some time talking about our expectations for fiscal '24, we'll spend a little time talking about the longer-term opportunity we see at Yext. Our TAM today is over $32 billion, growing to $60 billion in 2026, a 22% CAGR. Our platform expansion and investments in AI have enabled us to expand into additional markets. And here, we've got a slide that shows how the products are mapped into the different IDC categories. We're bullish on our opportunity and ability to capture TAM with our platform strategy. As you heard from Marc, we've significantly expanded the platform, adding new products and features. With our platform approach, we can target new logos, upsell existing customers and help our customers grow. We have the opportunity to displace multiple vendors as we offer a composable platform, enabling customers to build unique digital experiences. We have a robust product road map, and we'll continue our track record of strong innovation. Now with the breadth of our platform and our transition to DXP, we can now sell into more verticals. Since Q4 of fiscal '21, we've been able to expand into other verticals like services, consumer products and technology. In recent years, we've seen revenue growth deceleration. We heard from Tom and Raianne how they're rebuilding go-to-market around efficiency and productivity. We anticipate the business will grow again, and here, we've included versions of slower growth and faster growth, and we wanted to share with you how we think about it. While it's still early in our go-to-market efforts, and we are navigating macro uncertainty, we believe over the long term, we can get back to 10% growth and with better execution, get to 20% or more. We've talked about gross margins and the improvement we're expecting in fiscal '24 will put us in the higher end of this range. However, if we see higher revenue growth, we may reinvest. We've made a lot of progress on operating efficiency over the last 12 months, and it will continue to be a focus. In sales and marketing, we made great progress in fiscal '23, and we exited Q4 at 41%. Progress in the near term will be from better productivity, as Tom mentioned, and we expect our go-to-market changes to drive improvement over the longer term. With R&D, we've historically underspent here. However, we get a high level of productivity and innovation from the team, which is evident in the platform expansion you saw today. In G&A, we've seen modest improvement over the past few years and further leverage will come from revenue growth. Now we've laid out lower growth and higher growth scenarios. If we see lower growth, we'd expect higher margins. If we see higher growth, we may make additional investments. Both scenarios contemplate us operating as a Rule of 40 company. Now we've mentioned Rule of 40 a few times. We believe well-positioned software companies can achieve 40 or more, and our progress against our fiscal '24 objectives, plus the opportunity in fiscal '25 and beyond, should provide steady progress on an annual basis. Now, we will have some seasonality from quarter-to-quarter, primarily on adjusted EBITDA margins. However, on an annual basis, we will continue to make progress. And, as I mentioned earlier, how we get there will be a mix of growth and profitability. And as that mix becomes more clear, we'll provide further updates. So thank you for joining us today. We're excited about the opportunity ahead. Now we'd like to invite Tom, Raianne and Mike back to the stage for Q&A.

As a reminder, if you have a question, please raise your hand and wait until the microphone has been brought to you and please announce yourself and your affiliation.

Ryan MacDonald with Needham again. Maybe the first one's for Tom and Raianne. You obviously have joined the company recently and a very detailed sort of restructuring of sort of the organization, unifying of the organization. As you think about this year, you've got new categories that you're moving into, but then, obviously, a lot of maybe market demand for everything generative AI. Can you talk about how you're maybe structuring that lead gen engine to really take advantage of one, sort of the inbound demand, but then two, sort of moving into these new categories?

As I mentioned in my presentation, we rebooted the demand engine as a whole. We turned it on about 6 weeks ago. We rebuilt everything from all the operational processes to lead scoring and the tools that we use, our new campaign frameworks, as an example. We have a plan to go after both our existing customer base as well as new logo acquisition. And obviously, we're focused on selling the whole platform. AI is going to be a big theme throughout our campaigns as well, our AI-enabled technologies. But we're definitely focused on selling the full platform and our lead gen effort -- all of our lead-gen efforts.

Tom White at Davidson. Just maybe on the 10% to 20%-plus kind of long-term growth targets. Given the various different products you guys have and are probably going to continue to launch, I'd imagine there are a bunch of different ways that you can kind of get there. But curious just maybe, Mike, if you had kind of your crystal ball, what's your best guess around kind of what your product mix will look like once you kind of achieve or get back to those kind of long-term targets you put out?

Yes. So obviously, the fact that we put out 2 targets tells you that I don't have a crystal ball. And I think the point of that is several things. So maybe the news item here is that I think it's the first time we've told you that, with any level of confidence that we do see this as a growth business again. So as a headline, that's it. Obviously, we're also still being pretty cautious about those numbers and 20% plus doesn't have a ceiling on it. So it's a balance, right? To be as excited as we are about the product opportunity, to be -- to know the needs of the customers and the opportunity that we have there and to be patient with understanding that building the demand takes time, executing against that demand takes time and, obviously, the macro environment is -- it's always uncertain, but it feels particularly uncertain right now. I think part of the beauty of what we shared with you today is that we can be really flexible and thoughtful about the product mix because what's fundamentally differentiated about all the products is the CMS Knowledge Graph underpinning and the native AI natural language components of the platform. And so first of all, we're nowhere near done driving innovation through the platform and I think we've seen a lot of it just in the last quarter, with Chat, with Studio, with content generation. And there's a lot more to come on those fronts. What we're going to care about is the customer installing Knowledge Graph, the customer enriching the Knowledge Graph and the customer finding value in what's delivered through the Knowledge Graph and all the supporting products. So I wouldn't take a guess at what the product mix looks like. What I would tell you is that I think the position of our CMS and the DXP platform over the course of the next couple of years is only going to become more important to the enterprise and exactly how we break the value chunks of that out is a lot of the work that Raianne and Tom are doing and Tom's doing on the proposal side of things. But is that enough words to totally duck the question?

It's Bob Johnston with Herald Investment Management again. Could you talk a little bit more about the sales process at the enterprise level? I mean how does that evolve? And could you talk a little bit more about the competitive landscape? Do you run into anyone consistently or not at all? Or is it DIY or what?

Yes. I think it's all of that, but first -- in terms of the competitive landscape. But the first part in terms of the sales process is we wanted to get back to some foundational items and some of the blocking and tackling and how we execute on sales place, how we define sales place. So we've started doing those things really, really well again. I think what the challenge is going to be for the team when we talk about full platform selling and when we talk about entering new markets, it's not just the product enablement. That will be important, but the folks are very smart. They've been doing this a long time. They can pick that up. It's also going into a different persona. And you're starting into a little bit more of an IT persona and an IT buyer as well. So right now, there's a portion of the sales team, certainly, that is very comfortable and has been successful for many years selling to more of the marketing folks and the CMO and things like that. We're now enabling and transitioning over because we know that this is coming. So that's really a big change and one of the enablement processes that we're going through right now in terms of the sales process. And then in terms of the competition, it's everything you would think it is. I mean, you can -- you probably have seen the quadrants, the waves. It's -- we're going to be entering, certainly going toe to toe with some of the largest companies out there. That's for sure. And I think everybody is excited about doing it.

You had an interesting slide in your presentation where you showed kind of multipliers of customers with different numbers of products. So can you talk a bit about how you plan to kind of get that mix from 1 or 2 products up to 3 or 4 and kind of expand that impact on the company from a financial perspective?

Tom can talk most authoritatively about this, but I want to tie it back to the presentation that we gave today for a second. So we brought 2 customers here today. We're incredibly grateful that they came. They're 2 very different customers and with 2 very different sets of products. Cox has been very much ahead of the curve, I think, at adopting the full platform. And when you hear Jim talk about, they're way out ahead of the market in terms of understanding the power of the Knowledge Graph and how much you can actually do with it. We really wanted to bring Courtney here because I think we've told you for quarters that we have had this problem in the business. And this problem was this perception that we had stopped investing in Listings, so we didn't care about it anymore, and that we had also dropped the ball from a lot of the servicing things. And United Rentals is, I think, a great example of how we've turned that around. And part of that was communications and part of that was reprioritizing our road map to make sure that our core Listings and customers were getting the value that we had promised them, and that we had delivered for many years, and we had dropped the ball, so you're asking, what does this have to do with your question. So I think the point is that like the customers have different life cycles and obviously, we'd love to -- and I told you this today, and if she's still listening, she's not going to be surprised, we would love the opportunity to go to United Rentals and talk to them about CMS and other solutions that are not part of their suite today. But there was no scenario where that was happening at a time when they felt like we had deprioritized their core product, and we had dropped the ball on servicing them. So first, with the customer in that situation, first, you earn the right to have that conversation before you can go in there and you can sell a broader platform. And so the reality is that the customer personas and the customers that we sell to, they're anywhere on the spectrum from they've never heard of Yext or talked to us before, to a Cox who's always been way out ahead of the market and has been an early adopter of our products. And we just have to get real about how we approach the discussion with those different customers, and it's a unique and different discussion every time. So there's no one size fits all approach to this. And there's no sort of one path to -- There's another customer who we talked about in our Q4 call without naming them, that was a retail brand who had fired us for Listings. And we went back in there and basically begged for a meeting to go back and give a chance to compete for their listings business, and we wound up selling them the whole platform because they were ready for the full platform discussion. But first, we had to earn our right back into the room. So in a very long-winded way, what I'm saying is there's no one size fits for this. It's -- you have to meet your customer where they are and you have to earn the right to sell them the bigger thing. I don't know if you want to talk about the actual tactics or anything, but...

I mean the actual tactics, full-platform selling, we, of course, focus on our strengths, which in a lot of customers is list and maybe in the example you gave were 2 or less. Say they have Listings and Reviews, then we go pretty aggressively towards another insertion point, right? We find a pain point and whether that's in Pages or whether that's in Search. And of course, always talking about the value of the Knowledge Graph and how easy the extensions are after that. But we're doing a lot of that now, specifically with Pages. And when you see when you see the Spring release and when you see what I feel is the stickiness, I think what we feel the stickiness of Pages, it's a great insertion point and onramp onto that platform. And then once you have 3 of the 4 or, let's say, 4 out of the 5, if you count the Knowledge Graph, it's not exceedingly difficult to create a compelling commercial proposal to just have a full platform adoption, especially in this environment where customers are increasingly looking to essentially save dollars and reduce technology complexity.

Yes. I mean, while we get to the next question of, you know, increasingly tie into customers, the question isn't, do you use the CMS, it's how many do you have?

Yes.

Because they're fragmented and none of them are capable of doing everything that we showed you that we can do today. So that'll be a really interesting conversation as we go forward because if you have multiple, you're spending too much money.

A couple questions on sales productivity and pipeline. I'm looking at both of you. I think maybe just basic tactical question. How are you measuring sales productivity? Is it through upsells to existing ones? Or is there some level of new versus existing customer sales that you think, at some point, when you think flattening or the plateauing of sales productivity will happen not so far out in the distant future, which is when you then go back to Mike and ask for more salespeople? So I would love to understand how do you see that unfold? And then on the pipeline, both of you talked about pipeline innovation, in working closely with each other. I guess the question is how do the determine the quality of pipeline, and over what period would you see that quality versus quantity balance shake out in terms of the various different things that you're doing with the partners, resellers, Sis? It feels like a lot of things are happening. But over -- how do you figure out that the pipeline quality and quantity balance is kind of coming together?

First, on the productivity piece, we measure it through essentially usage of sales tools, right? So we have a set of sales tools now and whether that's what we call business value assessments and really leading with value to looking at white space tools. So we have a bunch of automated tools that are already in-house. And then we are creating another set of, you can think of them as common reusable assets that essentially, the way we do account planning, the way we do proposals, right? They all look the same now. The way we do same themes, right, where we lead with value. We talk about bottoms-up builds of demand. We have compelling commercials, right? And then if you look at value realization and post-sale customer journey, right, we have common assets there. So it's all about efficiency and it's all about speed. When you talk about the plateauing effect, it's more just, I think, a natural limit that we're going to hit with how many -- with AEs and their customers that you can only do so much effectively, right, to get -- they can only produce so much output.

The quantity versus quality is an interesting conversation. It's something we look at all the time within the marketing lead funnel. I mentioned that sales had given really great feedback over the past 6 weeks following the reboot on the quality. So we're now having conversations around opening up the quantity. We can be maybe less conservative on our lead scoring models as an example because the quality is so high. So that's something that we look at all the time. And then as far as increasing investment and putting more fuel on that engine, the same as Tom. Tom will be asking for resources. I will be as well as we see those indicators start to kick in.

Yes. The only thing I'd add to that is like, it's not that Glengarry Glen Ross movie, where there's a secret tour that has the good leads in it. It's a funnel of stuff and Raianne has got a machine built, in the process of being built to understand that. Technically, I think part of your question may have been just like how do you measure it financially, and I don't want to answer for Darryl, but I think when we think about productivity, it's in the incremental ACV bucket, right? So it doesn't mean we don't -- because we've moved the commercials to the reps as well. So they're in charge of the renewals also. But the point of that is to make the upsell path more frictionless and more opportunity there. So from a mathematical standpoint, the way I think about it is when we have average productivity near the top across the selling organization and a high degree of productivity, that doesn't -- and this is the point we're trying to make, that doesn't mean you turn around and start hiring. That means you ask the next question, which is, is there enough pipe for us to grow? And when you start deconstructing this, like it's not one answer because every channel and every group has their own dynamics, right? So we talk about all the verticals groups and the geos groups and their different customer segmentations and things like that. So it could be that you're ready to grow in some areas and not ready to grow in other areas. And so what you're hearing me unrelentingly hammer on here is that we're going to do this by the numbers when the numbers tell us to do it, not because we feel like emotionally, we need to find a way to charge the growth.

You've stayed really tight on script on not wanting to say when you're going to get to your 10% number. But it seems like you had demos out there that would be natural growth drivers. And what percentage of your customers have a Knowledge Graph that would get you to that growth if you rolled out that beta. Just the Chat product, for example, it seems like a no-brainer on top of your Knowledge Graph customers.

Yes. I mean I think the stat that -- the most telling stat is the one we showed you a couple of times, around 76% of our customers being 1- or 2-product customers, right? So -- and when you look at the multiples on the ARR of customers with higher products, obviously, those are average numbers and blended numbers, but it just gives you a sense of the magnitude of the upsell opportunity that's here. I would do a -- happily do a deal with you. You tell me exactly what the macro environment's going to be like for the next 24 months, and then I'll commit to a growth rate. And I'm not trying to be flip about it. I'm trying -- like part of the reason why I won't do that and why I stay so close to the script is because I just can't predict whether the buying persona's going to be really excited about this conversation around centralizing and using these AI things or whether they're going to be really focused on cutting cost. If I knew that, then I would be less coy about presenting you with a range within a 24-month period.

Is it safe to say that once those are out of beta, you'll have a pretty rapid sense of upsell? I mean it seems to me like that's a pretty easy thing to lead.

Yes. I mean I would say that for every product that we offer, and Tom and Raianne, they want to -- there's a lot of pieces that go into generating demand and then accessing that demand and how fast we come out of beta, and some of that stuff with some of those products involves other people's technology, and we don't know how fast that's going to move. So I think the data dashboard we're building that would tell us -- so for example, if a massive amount of pipeline shows up for any opportunity, and we know that that pipeline is high quality and qualified, that allows us to predict the ability to sell against that pipeline, which allows us to start looking forward. So we would, yes, expect to see that and that would drive some of those decisions that I'm describing.

As a reminder, please say your name and your affiliation and wait for the microphone so -- for the benefit of those in the webcast.

Daniel from Tenzing. Tom, I had a question on sales productivity. When you think your team is at the natural limit, do you think the growth of the business is at that 10% level? Do you think it's still below? Is it above 10%? If you can kind of sketch out where you think the growth of that natural business will be with the productivity at the levels that you want. And then a second question on the Chat product. Do you think that can be a land product? Or do you think that it's mainly going to be expansionary?

I think -- so on the Chat one first, I think it can be both because we've seen that reaction from customers so far, right? Customers that we are pursuing, as, call it, greenfield customers that we've had conversations with for quite a while, just haven't landed yet, have wanted to get into a beta program. Customers that we have, and we have an installed base with, there's been an overwhelming reaction to a beta program. So that makes me think that we can do it through both of those elements. Now...

I can take the growth one.

Okay.

Yes, that's a Darryl question.

Yes, it's actually -- it's tough to answer because we've got a mix of sales reps across a wide spectrum. We've got reseller, we've got mid-market and we've got enterprise. All of them have different quotas. They sell into different sizes. There's different average deal sizes. So it's not really a mathematical way that we can get to an answer to say, okay, when we have full productivity on all these reps, that's going to equate to growth rates of this because all of it is a very fluid motion with pipeline, right? If we see really strong pipeline and pipeline quality and quantity coming from Raianne in mid-market, that's going to change the dynamics. And if we have the opposite with our enterprise business, because the ticket sizes are different, deal sizes are different, and it's just there's a mix within that.

Can I just ask something?

Yes. Yes, go ahead.

On the reseller SMB, what can you do to get it better? I can see that Direct is clearly outperforming reseller in SMB. Is it just macro? Is there something else in there?

Yes. I think there are 2 primary things there. One is these environments are going to be harder for SMBs. They just -- more of them go out of business. And so our reseller partners, you know, predominantly sell to SMBs. So they're going to just -- it's just going to be harder to sell stuff to SMBs in an environment where they're fighting for their lives more than large corporations are. The second is that there -- and that's just -- that's the environment, right? I think the second is that the products we mentioned going through that channel is primarily Listings. It's not entirely Listings, but it's primarily Listings. So the opportunity there is to -- even if there is some level of pressure, is to expand the product that we're putting into that channel through the reseller partners. And as we launch more of this innovation, there are things that small businesses, for example, they're less inclined to be really excited about support Search or things like that. But I suspect small businesses will be really interested in a chatbot or a chat solution. So I'm not making necessarily predictions today on where we're going there. But I think as our innovation and our R&D continues to produce product innovation, there will be opportunity to put more through that channel, which will benefit both us and our partners.

At this time, there are no further questions.

Okay. Thank you.

That concludes our Investor Day and for the folks on the webcast, thank you very much for joining us, and thanks to those who are here in person for joining us as well.