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

Good afternoon, everyone. My name is Matt Coss. I work on the enterprise software team here at JPMorgan. I'm pleased to have Yext's CEO, Howard Lerman, joining us for this session this afternoon. Howard, welcome. Great to have you.

I'm glad to be on this amazing Zoom technology with you, Matt.

Pleasure's all mine. Before we begin, I would just like to inform everyone that Yext is actually reporting earnings this Thursday, and we're going to steer clear of Q1-related dynamics given that impending report. And also, if you're interested in asking a question, there should be a chat box beneath -- on the lower portion of your page. So to kick things off, Howard, why don't you start out with a very brief introduction of yourself and of the company.

Sure, Matt. I'm Howard Lerman, Founder and CEO of Yext. I started Yext in 2006 because I saw that we were witnessing a huge platform shift in the world of search. I've been saying the same thing for 15 years, which is that when you searched 20 years ago, you'd get 10 blue links back on a page. And today, when you run a search, increasingly, you get answers. And companies like Google have enabled search to change so much because of advancements in natural language processing, but most other search got stuck in the past. Keyword search based off an index that gives you those 10 blue links still powers the vast majority of search engines outside of Google within a company. And so Yext has made it its mission to, just like Google has brought natural language search to consumers, we're going to bring it to the enterprise.

Great. That's a great introduction. And maybe expanding further on that vision. So from where you've come from and sort of the keyword-based search for the consumer moving to the enterprise. So maybe how did you sort of start out in that role for search for the consumer? You can tell us about sort of the roots as sort of really in location and how you're diversifying the business to digital knowledge management and, like you said, that strategic shift towards enterprise search. Maybe just walk us through where you've been and where you're going.

What we're seeing is a platform shift where AI search is disrupting keyword search. Google started it, and now it's kind of picking up everywhere else. And there are 3 main components, main parts of AI search. The first part of AI search is natural language, and we talked a little bit about that. But essentially, natural language can take a full question, like who is Matt Coss, deduce it, figure it out and then give you an answer. And so that's pretty different than a keyword search, which simply relies on the exact string that you type into a search engine to give you an exact match of all the places in a document that contain the strength. That's the first thing that makes AI search different than keyword search. The second thing is that AI search uses multiple algorithms. So what I mean by that is when you run a Google search and you type in a question like McDonald's or a keyword like McDonald's, you don't just get one sort of result or one answer back, you get potentially maps or knowledge cards or snippets or blue links or parts of a car. And all these different, what Google calls, universal search. elements come together, they're all actually triggering a different algorithm on Google's back end. And it's not always obvious, which the best one is going to be. So Google gives multiple answers or using multiple algorithms back to the consumer. And that's the second thing that makes up an AI search. The third thing is the knowledge graph, and that's a brain-like database that contains all the facts about the world that we know and how they are related. And that is the central part of where Yext came from. You asked about sort of our origins and the location business. And it turned out that by building a knowledge graph by synchronizing information about a business, where their stores are, what the phone numbers of those stores are, all the facts about a company's wealth advisory agents, all that information. In doing so, we built a gigantic knowledge graph for every one of our customers, for their public stores, for their public agents, for their wealth advisers, for their insurance agents, whoever we were looking for. We put all that information together and stored it in the knowledge graph, and we were pushing it out to companies for their -- what we call, their answers. I'm sorry, when we were putting that into all these AI services, like Alexa and Google and others that were answering questions, and if you think about it, where do you see a listing, where do you see a business listing, you see that from an AI search because Google says, "Oh, you're looking for McDonald's. Here's the list of all McDonald's." And so we're sitting there, giving all the answers, if you will, or all the knowledge to Google and Alexa and Apple. And then one day, we said, well, wait a minute, the latest and greatest in NLP technologies available, let's help our own customers answer questions about themselves, and that's where the Answers possibly came from.

That's a great overview. I really appreciate that. And I think -- thinking back to the IPO, that's when we learned a lot about the knowledge graph and how it helps companies or sort of why companies need to understand why they need to organize their websites in a specific ways to leverage the way that Google will -- and Alexa will provide answers. And I guess is -- the knowledge graph has the foundation of what you do. Is that a fair characterization of its importance sort of in your role into how you develop?

A knowledge graph is the foundation AI search. Knowledge graph is where all of the structured information is stored, and that's what makes it possible for an AI search engine to use an algorithm like named entity recognition to retrieve a direct answer. Contrast that to keyword search, which is the old type of search. Keyword search doesn't have any structured information. They have a giant index that contains all the places like a document, almost like a library index, and they know where all the places on a given document they look for almost like hitting control F, all the different places in their index that contain your exact keyword. So there's no structure, there's no understanding. It's actually simply just a match. And then with keyword Search, what happens is usually, there's so many exact matches for a given question, for a given string that the differentiation comes from the ranking algorithm. And that's why Google took off, because initially they built the biggest and baddest search engine, but it was still blue links, but it just turned out that PageRank was the best way to rank all of those billions of results for a question like McDonald's. And over time, as search evolved from at least Google search, keyword search to natural language, now most of the time, you're not even clicking the links in Google. You're just getting the answers you need from snippets, from maps, from knowledge card. And all of that content, those answers are all in the Google knowledge graph. They don't even consult that giant index. They still have it, and you still see the web results. But usually, the answers you get at the top are all from the Google knowledge graph. And so what we have is kind of like a Google for every -- we give our customers their very own Google. That's the best way to think about it. We give them a knowledge graph. We give them their own AI search with natural language and multiple algorithm and universal search on top of that knowledge graph that we made for them. So the knowledge graph has always been the foundation of Yext. It used to be that we simply gave Google and Apple and Facebook, et cetera, all the facts about a company. But now on top of that knowledge graph, the company can answer questions about -- to customers for marketing. And even this week, we launched our support answers so that when you're looking for customer service questions, which is a huge category, on a help site, for example, you can ask a question and get an answer. And the ROI for that for customers is tremendous. They reduce call volumes. They can have case deflection, the implementation, the application of AI search is huge. We see, Matt, 5 categories of AI search. Marketing answers, support answers, e-commerce answers, workplace answers and then developer answers, which is being able to build a search engine into your app, and it's our intention to participate in all 5 of those categories of AI search.

Great. That's a great overview of sort of the foundational importance of the knowledge graph. And speaking of those categories, you intend -- I think you also mentioned this at your Analyst Day, competing in each area. Where do you compete now and sort of what's like the next logical place for you to go?

We're really competitive in marketing answers. That's the first area that we're an industry leader already with just our sort of first foray into the product. And we bucket our listing solution as part of marketing answers because if you think about it, you get an answer about a company and their marketing department publishes that to Google and to Apple and to Alexa. And so that's all part of our marketing answers solution along with our site search or AI site search solution. Today, that's been where we've been built up, and we continue to see tremendous opportunities ahead, Matt, in that particular segment of the market. But last week, we launched an adjacent market. The second category of search or AI search that we're going to be participating in, and that's customer support. And customer support has exploded during the pandemic, digital customer support took off like nothing we've ever seen. People instead of walking into stores and having their phones or issues troubleshooted with people in person, they just called or even particularly, they wanted to resolve the issue online. And what this caused was an explosion, a spike in tickets being made. A huge technique for deflecting customer support tickets is giving your customer the answer before they even write in and be able to do that requires natural language. Now today, the vast majority of help search sites, help desk sites use a keyword-based search engine. So you type in a question, like how do I log in, and you get jumbled results back. And so what do you end up doing, you write in a ticket, I forgot my password, I don't know what to do. A better approach is to just simply be able to answer the question and do so with natural language. And for customer support questions, natural language search or AI search is extremely important because we need to be able to understand what the customer is asking and then be able to give it back to them. In fact, in our investor site, we've posted, right now on investors.yext.com, an overview of our support answers solution, which talks about kind of the 6 areas. That search or AI search can apply in our -- in the world of support. And the help site search is one. But just to give you an example of a second is as a user is filling out we call this case deflection, as the user is beginning to fill out the support ticket, they start to type, we can analyze what they're typing using natural language and suggest answers on the right. And if they find the answer that they want, based off what the user is typing, then, bam, you don't have to -- they might not write in. They might get their answer. This improves customer satisfaction. This reduces the calls in. It's just honestly, Matt, a huge deal for any company. And every company on the planet that is dealing with this massive influx of tickets is going to want to try to deflect those and increase customer satisfaction. In AI search, support answers with Yext is the way to do that, and that is our second category of search. We feel really, really excited about that.

Great. That's very helpful overview about those key areas where you are and look forward to hearing more about additional areas where you'll enter. Now I want to get to the crux of the ROI of the Yext. Again, the roots are really helping people who are initiating searches sort of find what they're looking for. It's propagating the correct information across the web. Now it's propagating that correct information within an organization. So I just want you to share with investors here what is the ROI? Or what is the sort of bread-and-butter use case historically of Yext and sort of the tremendous ROI that you can see from having just correct information on the web? And then tying into that, one thing that we understand is Yext wasn't necessarily a COVID play. And so as COVID sort of crept up on all of us, what did that -- what impact did that have on Yext?

Let me answer your ROI question first, Matt, and then we'll get to COVID in a second. Search is the most important channel for any CMO. Search is the best way to acquire customers. You can tap into people that are already looking for your product or service. Companies make enormous investments in building sales forces and doing marketing to generate awareness of what they offer. But at the top of the funnel or the bottom of the funnel, depending on how you're looking at it, there is search where people are demonstrating intent. So when someone types in, I'm looking for a car product, you better believe that AutoZone and all the auto parts companies want to be there in the search results. And there's various ways that, that can happen. They want to be -- you want a blanket it. We were talking a minute about the universality of Google Search. You want to be there in the blue links, and you can do that with our Pages product. You want to be there in the maps, you do it with our Listings product. You want to be able to answer the question yourself. You do that with our Search product or our Answers product. So the ROI of Yext is measured by companies in increased transactions from search. And then in -- that's our marketing answers solution. In our support answer solution, it's measured by reduced or reduction in call volumes. I talked about this a couple of earnings calls ago, Krispy Kreme, the minute they put Yext Answers on their own site, saw a more than 40% reduction in contacts to their customer call center. The number one question people ask Krispy Kreme is how do I check my rewards card balance. And if you type that into Google, Google literally says, please call this number to check it. Now the result set that Google gives you is not even from Krispy Kreme. It's from some third-party site called giftcards.com, and it just causes it to rain phone calls to their call center, which is costly. As soon as we were able to answer the question directly on Krispy Kreme site, it's not like everyone goes there and asks, but it turns out a huge percentage of people do, and we are able to deflect those customer support calls. So the 2 ways that companies measure ROI in Yext are increase in revenue through more transactions driven from search; and secondly, reduction in support costs through support calls or case deflections. One other point in that. Let's talk about Maps specifically for a second, where that's kind of one of the things we're really good at. I don't know if you watched Google IO last week, half the presentation was about maps. Google Maps is one of the most important acquisition channels on the planet. And that is what people use all day every day and is an enormous driver of practically every location-based business's presence on the Internet. If you're not in Google Maps, you don't exist. And I'm going to now transition, Matt, to your second part of your question about COVID. We saw a huge decrease last year in Google Maps transactions. We went down, I think we said in our earnings call more than -- we saw Google Maps transactions year-over-year declined more than 50%, 5-0 percent. That is just an enormous hit. And it's because businesses were closed and people were forbidden from going out. And that was one of the core value props of what we did was offering ROI revenue per transaction or increases in transactions off of Google Maps to our customers as we help them put the correct information and synchronize the data everywhere. So when the traffic dropped by half, particularly in certain verticals, like retail and hospitality, we saw challenges last year. And you're right. I guess that means we were not a COVID play because a lot of these businesses were closed and a lot of our verticals, a lot of our customers were really hurt. And we really focused last year on a few different areas and mainly getting our AI search product answers cranking and also on efficiencies. And I think as you saw last year, we, I believe, did a really good job standing up an entirely new category of product built on our knowledge graph foundation and doing so and getting really efficient all at the same time.

Great. That's very helpful. Howard, on the sort of diversification with Answers, I know that some of your ARR comes from very impacted industries like retail and food services. But maybe can you talk about how Answers is allowing you to get into verticals that you just haven't been in before? And why that's acting in sort of the natural diversification for the entire business?

I estimate -- we estimate about half of all companies in the world, enterprises, let's say, could purchase Listings from us because they care about Google Maps in a big way and they care about being Alexa in a big way and they care about being found online in a big way to consumers. 100% have a need for AI search within their customer success organizations, within their marketing organizations, within their developer organizations. So what we have done with expanding our product set from the knowledge graph and Listings to the knowledge graph and Listings and now Answers is dramatically expanded the type of customer that we could sell to them. And so you look at it, and you use the word diversification. I look at it as expansion. Never before were we able to sell to, for example, the state of New Jersey. And standing up a whole site for them with the whole Answers experience for them. Never before would we have been able to sell to the World Health Organization, where if you go to the COVID-19 site that they power, the Yext Answers bar is front and center there. Never before would we have been able to sell to something like the PGA tour, where if you go there, you'll see a Yext site search experience. They are in the top right. So the kinds of companies that we now can sell to is dramatically bigger and, also within our existing clients, deep in the finserve category. For example, you see us expanding with site search opportunities with support search opportunities. So we just have a ton more product to sell. And it's valuable stuff because every time you run a keyword search, you can picture it, picture keyword search from the 1990s. This is the -- it's like a Yahoo! or Lycos or AltaVista search. An AOL CD, think about getting one of those and getting 10 blue links back. It's funny. You've probably heard me say if you've been following Yext, used to be when you search 20 years ago, you'd get 10 blue links back on a page. I said that in our IPO page. I said that in our S1. Believe it or not, the Head of Google Search and Google IO literally used that exact quote in IO last week when he was introducing their new mom technology and said, 20 years ago, when you'd search on Google, you get 10 blue links back on a page. Well, that's great for Google. But all the other searches out there still use that same old blue link indexed keyword-based technology. So it is our mission to stamp out keyword search throughout the enterprise and put AI search everywhere.

Great. That leads me sort of into the next part of this. What variables are at play when the world fully reopens that could help your growth? And so just again, I know you're reporting this week, so not talking about any current trends. But just longer term, how should we think about what will support Yext's growth as the world opens back up?

I think the acceleration of the digital transformation is something that is causing so much more and an unprecedented amount of transactions to happen over digital channels. And within a digital channel, whether it's a marketing channel or support channel, we didn't even talk about workplace, which is the fifth category of search. We think that's huge. I don't doubt that within JPMorgan, your enterprise probably has all kinds of search boxes everywhere with little magnifying glasses that give you really frustrating sets of results about funds or employees and people give up and they go and they talk to their -- the rise of work from home. Think about the rise of work from home as it relates to being able to get and access information within the enterprise. Now we're an office-first company, and I know you guys are, too, based on what I've seen your CEO say online but -- or what I read him have said. But these are the trends, these are the forces that people are asking more and more questions. And the best way to answer a question is definitely not through trying to index a giant amount of text and then find an exact match keyword. The best way to give people what they want is to structure that data or have a set of known unstructured data and be able to mine it with an algorithm like our extractive Q&A, which can give answers from unstructured text. So keyword search is going to be dead everywhere. Keyword search is going to die everywhere, and it's our opportunity to compete with Elastic, with Solr and win in deals that companies otherwise would put those technologies in, they can use Yext instead.

Got it. That's great. And since you mentioned Elastic and people think of -- I think they think of Elastic when they think app search or website search. Can you sort of compare and contrast if I want an app search or if I want a website search, what am I getting with Yext and what am I getting something like Elastic?

Three things. If you're using Elastic, you are using keyword search. It's a big old index. Now there's a use case where you should use Elastic, and that is the use case in which you want to hire a team of machine learning specialists and build your own custom search engine from scratch. That is a really great use case for Elastic. So I think it sort of depends on the type of application. Spotify should use Elastic to build their personalization algorithms, to share with you the kinds of music that they want to have. Should a customer support site on help.x.com build their own custom search engine? No. So the 3 ways that we're different than an open source or sort of hosted version of an open-source technology, like Elastic. And by the way, their big competitor would be Solr, who is another huge player in the search market. They're all index based. They're all not low code. You have to show up with your own engineer. And then here's a big thing, too, which is that, as a consequence of being -- having to hire your own engineers to build a search engine on top of them, it takes a while. So the time to value is long. So I'd like to say that overall, if you want to build your own search engine, have a custom thing, you should definitely use one of the open-source thing, bring your own developer. But if you want to just turn something on and be able to answer questions off a support site, off a help site internally with AI search that gives answers, then Yext is going to be a better and a faster and a cheaper option.

Great. That's really helpful. And then when I think of sort of the alternatives to search, there's chats. It's a common one where I'd get on the website, try to find something quickly. Can you compare and contrast Yext Answers with maybe a chatbot or desktop agent chat?

I see in the future, all these different services kind of being used by people. Search is fundamentally a private thing. People are way more likely to search for something than they are willing to tell. If they think they're talking to someone, they might not truly bring up what they're looking for. So I think you're going to have both search, you're going to have chat. These are the 2 primary interfaces, but we have not seen -- there's a lot of companies out there doing chat. One problem with chat is it's very expensive. You still need to have agents on the other end answering the questions, and that can cost $6 to $8. So actually, as a strategy, we've seen a lot of companies employ even when they have chat is still try to deflect the chat with search because it costs $8 or $6 every time someone does a session to do a chat on the other end when you're using a service. So I think you're going to see both of them. The thing to look to, though, is Google, because they are always the leader in everything. And as you saw, Google launched a couple of years ago, remember Google Assistant, it was like kind of a chat app. It didn't take off because people didn't want to chat, they wanted to search and they just wanted an answer. So when you know what you want and you want an answer, we think search remains the most compelling way to get that. That said, you can integrate Yext search with a chatbot, you can do all kinds of stuff. We're going to see both of these be a pretty interesting way people get what they want.

Great. And then I understand there's companies who are very interested in search and they tend to sort of look at search and it sort of elevates the ASP, elevates close rates. Can you talk about sort of the impact that search has on some of your wins recently?

Well, there's no question in that. When you look at some of the big companies, they're only buying search from Yext. They're only buying our AI search product, but they always buy the knowledge graph platform. So I think that overall, search is a little bit lower of an ASP if you only buy search. But when you sort of buy our full platform solution, which is the answer of what we call the Answers platform, which is an AI search solution that comes with knowledge graft, that comes with Pages, if you wish. And an example of Pages is you can put your whole health site on SEO pages that will appear in Google. So when people search in Google for a question, they're going to get a snippet that kind of appears up there, hey, how do I log in to my x account. Well, when Google hears that, your answer should appear as an FAQ in Google. And so our -- one of our support search solutions for search enables that to happen. So that the answer can happen directly from Google or directly off that company's help site. And it all comes from the same single platform. And that's another huge advantage of Yext over an Elastic or Lucid or Solr, which is that you have one knowledge graph, you update it once and it updates on your own site, so you can answer the questions on your own site. It also updates in Google, also updates in Alexa across our network. That's pretty cool.

That is pretty cool. We do have one question from an investor, I'll go ahead and read that. Howard, what do you think investors are misunderstanding about Yext, which is not giving you enough credit for?

Well, gosh, I would love to hear the investor's answer to that question. But I think we had a challenging year last year given COVID-19. And Google Map SKUs were down 50%. I do not think everyone expects that to be the case permanently. But as a consequence of that, we had a challenging year, but we made the best of it, and I'm really proud of our team.

Got it. Okay. Very good. And then I know you're doing some things like enjoying some leverage in your model. So if I look at things like sales and marketing, that's really improved over the past year as a percent of revenue. To the extent you can, can you talk about sort of efficiency in your sales delivery, other efficiency or productivity gains you're seeing in an organization?

Well, remember, every deal, we used to -- 1 year ago or I should say, 1.5 years ago at this point, we used to be a heavy event-driven, heavy field sales-driven company that relied on events and sending reps to places to chase down contracts and customers walk the halls. Now every deal we did last year was not that. Every other -- a lot of other software companies did that, too. But we are not a product-led sales approach like a Slack or a DropBox. We're a field sales-led approach like a Salesforce. And so last year was disruptive for us because we couldn't travel, we couldn't do events, we couldn't be with customers as much. But as we -- but we took the time to kind of reprogram the DNA of our sales force. And you saw that reflected quantitatively in our results for the full year last year. Obviously, we can't talk about Q1. But if you look at our full year results last year, I think we were cash flow positive for the year, and we had good EPS results in the fourth quarter. So these improvements, these sales and marketing improvements, we believe they are sustainable because we adapted to a new selling notion. Furthermore, in the AI search business or the search category, there's a lot more existing demand that's happening. So we're less evangelical, and we can show up and catch the leads of the people that are sending out RFPs for support search or search, tap into that existing demand and then upsell kind of other things going forward.

Very good. I think that's a good note to end on. Looking forward to hearing about Q1 and what Answers will bring us. I just want to thank you for your time joining us this afternoon and really appreciate you being here.

Matt, I will say one last thing, which is, the next time we do this, we're doing it in person. I can't wait to see you, and I'm so sick of these damn Zoom boxes.

Hope it's in person, too. Thank you, Howard.

Thank you, Matt.