Good morning, and welcome to Taboola's first ever Investor Day. I am Jennifer Horsley, Head of Investor Relations. We are excited to have so many of Taboola's investors and covering analysts here with us today in person at Nasdaq, and many more of you watching us virtually. We know how valuable your time is, and we appreciate you dedicating some of it to be with us today. We have a full morning of presentations, customer discussions, and product demos, so let me quickly get through the safe harbor. Certain statements we make today, including our expectations for future periods, are forward-looking statements. They are not facts and are subject to material risks and uncertainties described in our SEC filings. These statements are based on currently available information, and we undertake no duty to update them except as required by law.
Today's discussion is also subject to the forward-looking statement limitations as set forth in our presentation materials. Future events could differ materially and adversely from those anticipated. At times today, we'll refer to non-GAAP financial measures. For definitions and reconciliations to GAAP, please refer to the non-GAAP tables in the presentation materials on our website. With that, let's get things started by rolling the video and bringing to the stage Adam Singolda, our CEO and founder.
Good morning, everyone. Can you hear me well?
Yes.
Yeah. First of all, let me just tell you that today my four-year-old son, Ozzie, told me, "Abba," which is dad in Hebrew, he said, "If you'll do a good job today, and Taboola will do a good job at the Investor Day, you'll get a sticker." First of all, I mean, in his world, that is the highest way to reward anyone who's doing a fantastic job. Let me first welcome everyone to our very first Investor Day. Thank you, Nasdaq, for hosting us. This is really cool. I mean, too many COVID tests at the beginning, but it's fairly nice. We have our leadership here in town. We have our clients and partners, we have investors, analysts, and we have thousands of people watching remotely who could not have made it here today. I'm excited to kick this off.
We have a really exciting agenda for us today. I'll start with some of our story. Tom, our VP of Strategy, will talk about the market. Eldad, our President and COO, will go over what differentiates Taboola, what makes Taboola's core platform unique and special. Gil, our VP of Algo, will talk about our special sauce. We get so many questions, Gil. Now it's your turn to take it. Enough. You'll talk about everything that makes us so unique. We'll have Evangeline, Director of Publisher Accounts, talk about our newsroom, our editorial suite, as well as hosting McClatchy here with us today. We'll have Eldad again speaking about our strategy to recommend anything and anywhere over the next many, many years.
We'll have Nadav, our VP Global Brands and Agencies, talking about our High Impact Placements, how we attract premium demand from brands and agencies, as well as having a conversation with eToro client here with us. We'll have Bob, the President of Connexity, speaking about where we are with e-commerce, where is it going, our vision about bringing e-commerce superpowers to the open web. We'll have Jonny, our head of Taboola News, discussing how we distribute news to devices, as well as having Samsung on stage later discussing how they invest in services and how they work with Taboola. Kristy, our SVP of People Operations, will talk about what matters to us almost the most, which is our people, our culture, and things we care about. Of course, like all great events, we'll finish with our CFO talking about financials.
We'll have a Q&A session at the end. I started Taboola back in 2007 because I couldn't find information and TV shows on TV, and I thought I should not be looking for TV shows. TV shows should be looking for me. I thought, you know, there's just too much going on. There's only 24 hours a day. That will never change. I thought there's huge opportunity for us, you know, in terms of what to watch, what to read, and things of that nature. If you go back in time, 1990s, the way you ask yourself, how did people ever find information? We used to have portals. We used to go to Yahoo and AOL. It was like Yellow Pages online 14,400 modem. We would click on things and find information.
Back in the 2000s, search was invented with Google really building a huge category with the assumption that if you know what you're looking for and you can type it, you can find it. What happens when you have no idea what you're supposed to be searching next and what to do next? A lot of times we think about Taboola as kind of a search in reverse. Instead of expecting people to look for information, how can information be finding us? Actually, Larry Page, when he started Google, said that the perfect Google one day will be an empty page. You wouldn't have to search for anything. You would go to google.com, and the search would be surfacing itself to you based on what Google think you need to find out.
I'm happy to tell you that when you think about recommendations and search in reverse, many of us have been using recommendations all the time. In fact, we can't live without it. When you watch TV on Netflix, you're using recommendation engines. When you buying things on Amazon, you're using recommendations. Obviously, this is a page personalized for me. Who doesn't like a good lightsaber? When you're listening to music on Spotify, you're using recommendation engines. When you're just bored and you open a TikTok, you're using recommendation engines.
As I think about the future and the opportunity we have, our vision, I think about the opportunity to build a whole new category called discovery, bringing the power of recommendations that we're seeing in the walled gardens on Amazon, on TikTok, on Instagram, on Twitter to the open web, helping people discover things they may like but never knew existed. I'm convinced that our children and their children will have recommendation engines all around them all the time, making them and helping them making the big decision they ever need to make, what to watch, what to read, what to buy, and so forth. I'm happy to tell you we've been making some really good progress. Advertisers, some of the best ones in the world, are already diversifying outside of the walled garden with Taboola.
Some of the most amazing publishers in the world are already taking a technology-driven strategy with Taboola, bringing technology and analytics and insights into their editorial staff, so they can grow, drive revenue and engagement and audience growth in times when social networks are sucking oxygen back. Some of these publishers are with us today. Some of the most amazing OEMs, device manufacturers, are taking a service strategy similar to Apple News, but on Android devices, saying we need news to engage consumers, so we can go up the food chain, so we don't just sell devices, so that consumers interact with us all the time. Some of those OEM partners are with us today.
Now you'll notice that I spoke about the vision, why I started Taboola, about building a recommendation engine in the open web, helping people discover things they may like but never knew existed, but I didn't say ads once. The truth is that I didn't start Taboola to be an advertising company. I started because I thought there's huge opportunity to help people discover things they may love but never knew existed. When you look at the open web, the $64 billion advertising market, it was just huge and obvious we had such an opportunity ahead of us. The reason was that if you look in the open web and you look at the ads in the open web, they're just ugly. We're still seeing banners all over the internet. This advertising format was invented 30 years ago.
This is when Tamagotchi and DVDs were cool, and we still have banners up until now. None of you this morning woke up in the morning and say, "Oh, man, I really wish I can click on a really great banner today." That has never happened. In fact, most of you, if you clicked on a banner, was probably a mistake and years ago. That's still most of the $64 billion advertising market. We're still using the same advertising format as 30 years ago. When we think about the future, we think there's a huge opportunity to make this so much better. Now, if you take a step back and you're looking at walled gardens, which all of us are using every day, they've actually figured that out years ago. None of them is showing you a banner to build trillions of dollars of market cap companies.
They've already moved on to an integrated, relevant, personalized advertising experience. If you go to Facebook, some of the posts are organic, some of them are paid. Google, perhaps the very first native advertising company, invented a search experience. Some of the search results are organic, some of them are paid. Twitter, the same. Amazon, Jeff Bezos said it's so hard to build trust. It's so easy to ruin it. There are no banners on Amazon. Some of the recommendations are organic, some of them are paid. Same for Snap. This is exactly the revolution Taboola is bringing to the open web. As we think about that $64 billion of banners, which none of you clicks on, we think of an opportunity to reinvent, reimagine the internet in a way that's much more like a walled garden experience, but not evil. Much more relevant, integrated, personalized advertising experience.
Some of the things are organic, editorial, more from McClatchy. Some of it is paid from the internet. Now you'll hear Tom, our VP of Strategy, have a session on that right after me. Just from a high level, if you think about the open web, what is this market? People spend 25% of their time in the open web. What do they do? They watch videos. They buy products. They go on their connected TV. They open CNBC. They watch a show. They open a game on their mobile device. Where do they do it? On a browser, on their Samsung device, on their connected TV, maybe in their Tesla. Now, all of those moments compose billions of them altogether that advertisers spend $64 billion against those moments, but those moments are not relevant.
Many of those open web $64 billion moments are not relevant. We ignore the user. The open web ignores the value it can create for consumers in the open web. There's huge, huge opportunity for us to change that. Let me tell you something about those billions of moments in the open web that are not relevant, not personalized in the open web. It's really, really hard to do personalization. It's really, really hard to personalize to someone in less than a second, to make an educated guess as to what they might like to do next before they even told you who they are or what they wanna do. You don't know how hard it is to personalize recommendations? It's like sending spaceships to space. It's like rockets going to space hard. This is how difficult it is to personalize recommendations in the open web.
To do that, to send rockets to space, to drive personalization in the open web, you need two things. The first one you need, the first thing is an engine. To send rockets to space, you need an engine. In our world, that engine is the AI. You'll hear Gil, our VP of Algo, speaking about that later. What is that engine that many of you constantly ask us about? Why is it that we're able to drive differentiation, better yields, better margin, better growth than other companies? A lot of it's because we invest so much in technology. We got into deep learning about five years ago. You'll hear more about that later. We invest $100 million in R&D.
Eldad, our President and COO, will speak about why we invest so much in R&D, why we think that makes us special to go beyond just revenue in the world of the open web. We don't just show ads. We also recommend you editorial recommendations, right? About half the clicks on Taboola, people click on Taboola 30 billion times a year. Half of it is just more from the site that you're on. We need to build technology that's not only good for advertisers, it's really good for consumers and publishers. That's part of our engine. SmartBid. Of course, SmartBid, which create this autonomous experience for advertisers so that anyone can be successful outside of the walled garden. Now, I'll tell you, many people speak about deep learning as something that they do. They speak about AI and deep learning.
I'm telling you, it's really, really tough. It's really hard. We have the best people in the world doing that, and there's a world of difference, huge difference between machine learning, ML, DL, deep learning, and BS, bullshit. The truth is that to do real deep learning that is special and unique and works with a lot of data and a lot of scale is very scarce and a very rare event, which we do and we have. That makes us special, and you'll hear more about that later. To send rockets to space, the second thing you need is fuel. Our fuel is that unique data that we have, and I wanna tell you about that data. You've heard me speak about our publishers' relationships so many times. We used to have three-year partnerships, then it's five.
We recently started signing 10-year partnerships. I still wanna sign my first forever partnership. Publishers should sign forever partnerships with Taboola. Let's work forever together. The reason it's special is because it gives us unique access to consumers, about half a billion people every day. If we see you on ESPN today, we'll see you on ESPN tomorrow. It's like a consumer company only in the open web through those exclusive global relationships with publishers. We don't track you across apps. We don't go social dilemma on you. We look at a page, what you're reading about is a proxy for which market you're in. Contextual advertising, contextual signals as a strong proxy to drive the value we need to do. People click on us billions of times a year. I often compare our data set to kinda like Amazon, people who buy this also buy that.
For us, we see what you read, we see what you click on, so it gives us this cross-website experience, so we can say, "People who read about that also click about that." This is relevant across websites, so we don't have to use third-party cookies. 90% of our revenue is direct. This is super special also in the open web. You see it on Google, Amazon, or Facebook. When you buy search ads, you only buy it from Google. When you buy Facebook, you buy from Facebook. Historically in the open web, when you buy ad tech, you buy through a thousand channels. The company a lot of times never actually know who the client is. For us, 90% of our revenue is direct. We know who you are. We have a pixel on your page.
We're optimizing for you, and you're using our own technology to be successful. All of those things give us just an ocean of data that is very special to us and allows us to navigate privacy dynamics. It allows us to navigate cookies going away. You've all seen our Safari yield going up while Apple was deprecating cookies. That is very special to Taboola. To send rockets to space, you need fuel. I gotta tell you, we're building the best of them. We're building the best rockets out there in the open web with the best engine, the best people in the world working on building that engine, and we have a special relationship with publishers that gives us a special fuel. Now, don't believe me that we're good. Look at our partners and clients. Believe them.
We have some of the best publishers and the best advertisers in the world who choose Taboola. They trust us. Trust as a proxy for the journey we're talking about. This is not a vendor relationship. We're not buying inventory. We're not in a waterfall. We don't hope for the best. We're together building our strategy with our publisher partners. Some of them have been with us for 10-plus years. Advertisers, small and big, find Taboola as a real channel that can scale outside of the walled garden so they can rely on us, and you'll hear more about that later. As I finish, and I look into the next decade, 'cause I'm allowed, I told you about those moments that happen in the open web. The average person consumes eight hours of digital media every day. That's a lot. That's thousands of moments.
Some of those moments happen in the walled garden. We already said the walled garden's pretty good. Amazon is pretty good. Twitter is pretty good. They do a pretty good job surfacing things I like, giving me advertising experiences that I like. Those moments that happen outside of the walled garden in the open web. They're not where they need to be. I'm convinced that those moments, 10 years from now, the open web, every time you open a browser, every time you open an app on your phone, when you open your connected TV, all of those moments will be relevant, personalized. It's much like, you know, electric cars. It's obvious in 10 years we'll have electric cars, most of them. I know what you're thinking, TSLA, TBLA. I mean, don't let me stop you.
I'm just saying that I'm convinced that the future will be personalized. We will not live in a world where it's gonna be static for all of us. Our kids and their kids will be shocked. We used to live those days now in the open web, and I'm very excited about it. As I think about all of that, it's not only that we're having a good time, we're driving growth, working with amazing partners and clients, it also warrants us the opportunity to see financial success. We just announced today that we reaffirm our guidance in Q1. We assume an accelerated growth over the next three years to allow us to cross $1 billion in ex-TAC as a company, maintaining our margins. You know we like, we appreciate profitable growth.
All of those dynamics maintain while we accelerate our growth, crossing the $1 billion xTAC in three years. You see us on browser websites many, many times. We'll be on mobile devices, we'll be on a connected TV. Over time, we'll be on every watch, every car, every audio device in your home. Taboola will be powering recommendations for the open web in the future, powered by Taboola. Thanks for having me. On that note, Mr. Inbal.
Okay, let's see. Can you hear me now? Oh, there you go. Good? We're good. No.
No.
Am I saying no.
There we go.
There we go.
Mr. Bill Gates. There we go.
Good? All right. Thank you. That is the actual VP strategy. You know, he asked me to fill in. Really excited to be with all you this morning. Thanks, Adam. I'm gonna try and pick up where Adam left off more or less as if that's even possible. I think Adam talked to us about the open web, where we spend 25% of our time, and advertisers are already spending over $60 billion. Adam mentioned how that's kind of comprised of different segments according to what people are looking for, what they're kind of consuming, and where they're doing that. I wanna dive deeper into this concept of walled garden level kind of power and walled garden type abilities in the open web, because really that unlocks this market.
It's a very large market, and that's what unlocks it. It's what makes it relevant for Taboola. All of it, and probably more than all of it. When we say recommendation, that's kind of what we mean in that sense is, if we can do this right, then we have a really good claim to being relevant and useful and adding value to all of that. Let's try and dive in. What we're talking about is time, and we're talking about value, okay? I'm gonna show you a little bit of an analysis to try and share why we're so excited about the endless potential for growth that we think Taboola has and that we think we can deliver to our clients and partners.
In order to do this analysis, we're gonna be using data on U.S. adults because it's the most robust data set out there, and we're gonna focus on mobile, so internet usage through mobile devices. 'Cause that gives us more of a level playing field across different companies. Different companies have a very different profile in terms of how people kind of interact with them, and mobile is a good apples to apples way to look at things. We're gonna be looking at U.S. adults on mobile and thinking about time and value, right? Time and money. Let's start with the open web. The open web, the entirety of all of the you know the great companies in the open web, are capturing 25% of people's time on digital. That's around eight hours a day for a U.S. adult on digital.
25% of that time is on open web. 55% of that is on mobile. There are around 224 million U.S. adult mobile internet users. I know you're all already you know, you've done this in your head by now, but that works out to 5.3 trillion minutes a year. When we take you know, the portion of the 64 billion that's on U.S. mobile, which is around $17 billion, and we divide it by that time, it turns out that the open web on average creates just under a third of a cent per user minute. Zero point thirty-one cents per user minute for the open web with all of the great content that we all enjoy and all the great services and products that we all enjoy on the open web.
We've said that the walled gardens are doing much better, and let's see if that's true. Let's do that same thing we just did for the open web for Facebook. This is Facebook, again, U.S. only, and it's just Facebook, no Instagram or any other properties, just Facebook in the U.S. on mobile. Obviously, the majority of their revenue is on mobile. Less time, around 2.5 trillion minutes, right? They're making 6 times as much per user minute, $0.019 per user minute. Let's have a look at Google, right? Maybe one of the best at what we mean when we say a walled garden and walled garden power, we all think Google. Same technique for Google Search. This is just search, no YouTube or other of that stuff. Obviously, far less time. Shouldn't be surprising.
A huge jump in value to over $0.027 per user minute. That's 9 times as much as the open web. These are the same people. It's the same user across all of these different companies. It's the same user, but the minute is worth 9 times as much, and Adam's already told us why. It's because in order to extract that value, there's some stuff you need, and they have it. They have the AI rockets. They have the tech. They have tremendous amounts of unique fuel of the data that they have about all of us and from their direct relationships, massive scale of direct relationships, and they have the ability to give consumers what they want, the relevance. Nobody ever turned on a mobile device or a TV or anything and said, "Let's look for an ad. I want to see an ad." Right?
They give us what we're there to see. They, even when they are serving up an ad, it is relevant, and that's walled garden type ability. That's what we mean by walled garden type dynamics. It should all sound familiar 'cause it's exactly what Taboola brings to the table for the open web. It's exactly what we can do, and that's why it shouldn't be surprising that Taboola on this same chart is over here at around $0.022 per user minute. Obviously, the opportunity for Taboola is to have a quantum leap in the amount of time we're able to spend with people. That's where the potential lies because we're spending far less time with people. What do you need in order to spend more time with people?
Well, what I just said, nobody's gonna stick around with you because you're showing them a bunch of ads. You need to show them what they came to see. Well, good news. Taboola knows how to do that. The majority of the clicks on our network, the majority of the items we serve up, the majority of time on our network is spent on what we call organic. It's the stuff people came to see, and we do it in very long-term and strong partnership with our publisher partners who also know a lot about their audiences. We want more time. How much more time is there? What's realistic in that sense? Well, we said U.S., right? U.S. adults spend over 13 hours a day on media, according to eMarketer. Eight hours of that is on digital, so 480 minutes.
You can see the different segments that that's spent on, right? A lot of different ways to do digital. It turns out that out of that 480, there are three categories on this chart where Taboola is already well-established, and there's 130 minutes excluding social networks in those segments. That's CTV, you know, no gaming, none of the exciting stuff that we will do that we are relevant for. Even putting those aside, there's 130 minutes per day for the average U.S. adult in categories where our products are already well-established and our partnerships are already in place. Now, let's use this slide and say, this is the entirety of digital.
This is the 480 minutes, and let's see how much of it Taboola has already tapped into so that we get a sense if we have a glass ceiling going on, if, you know, if we do or do not have enough room to grow. Okay, there's Taboola. If you're having trouble seeing it, there's Taboola. That's $519 million of ex-TAC gross profit in 2021 based on 10 seconds a day, U.S. mobile, okay? That's a fraction of a percent of digital. We talked about the 130 minutes where we're already well-established. It's a fraction of a percent of that too. We think there's endless room to grow. We think we're exceptionally well-positioned in order to capture all this space, definitely what's in this frame and probably a lot of what's outside it as well.
What happens if we do that? Well, what happens if we can capture just around 5% of open web U.S. mobile time? They're around 65 minutes that people spend. I said earlier, it's 55%, you know, of the open web time. So there's around 65 minutes. What if we can get just 5% of that? It works out to under 10% of what Facebook is doing in the U.S. Let's make another assumption that in order to spend that much more time with people, even if it is just 5% and far less than 10% of what Facebook is doing, we'll take a huge hit to our value because we'll need to go even bigger on organic. We'll need to decrease what people call ad load. Fine.
If we do that, it'll be around 3.5 minutes per day. The ex-TAC gross profit just on U.S. mobile will exceed $1 billion, and by extrapolation, the company on the whole will have around $3.5 billion of ex-TAC gross profit. We hope you agree that this is both exciting and achievable. We believe it is. We're working on this every day, and Eldad's about to come up here and tell you more about it. This is, you know, line of sight for this is what we're doing every day. If we wanna dream big, which is what we always do at Taboola, let's go for a really big number, 5 minutes.
Still under 10% of what people spend on the open web, and it's 10x our current ex-TAC gross profit if we can capture 5 minutes a day with all of the content of the open web at our disposal and all the relationships we have with our partners and clients. That's why we think there's huge potential, huge, you know, endless room to grow, amazing open web opportunity in front of us that Adam described. We have exactly the ingredients needed in order to build those rockets, create the value for our clients and partners, go after this, go to 3.5 minutes, and then go to 5x, and 10x the company. You know? Lucky for me, I'm VP strategy, so I get to say that and then introduce Eldad, our COO and president, who's gonna talk about how we do it. Thank you.
No. I'll start. Can you hear me?
Yes.
Good morning. Eldad Maniv, President and COO. Just about 10 years with Taboola. Science background, math, physics, and then into the business side, having worked in different industries, the first of which was aerospace. I really like Adam's rocket science part. Let's start with the video. Adam spoke to you about recommendation engines. Tom spoke about the opportunity. I want to talk about our most important partners, the publishers, and what we've learned. What we've learned working with thousands of publishers is that there's a very common theme. In essence, when you look at them, they have a very interesting goal. They are all trying to get audience, people to come on-site.
They want to engage them, so they want them to stay there, and once they're there, they want to generate revenue. That will generate the maximum value. There are many ways to actually get that done and do that. They have tools at their hands. Doing that, they're doing multiple things. They engage users by bringing them on-site in various ways. Once they're on-site, they create content that's text, that's videos. They try to monetize with video ads, with premium ads, with any type of ads. But it's extremely complex. Using all of those tools, it's really extracting the value is extremely difficult. Thank you. Now, what is the challenge? The challenge is that they use the same common set of tools. It's really a one-size-fits-all. What does it mean?
If we have different users coming in, I'm interested in economy. My partner is interested in politics. At the end of the day, we all get the exact same experience because, again, it's a one size fit all. That really doesn't make a lot of sense. What we're really interested in is personalization. By the way, the monitor is not synced. What makes the most sense is really to personalize. Let's look at this example. We have Jen here. Jen is a visitor to one of our publishers. She is interested in e-commerce. She's an avid shopper. She also likes to watch videos. On the other hand, we have James. James is a really avid follower of this publisher. He's a very loyal user, comes very frequently in. What do we do with each one?
What the ideal situation would be is to personalize the experience for each one. We would want Jen, who really is an infrequent visitor, to be presented with a video because she happens to like watching videos, but really offer her a purchasing, an e-commerce and a purchase e-experience. On the other end, James here really wants content. We want to minimize the amount of advertising, the amount of clutter, he gets and work with him and show him premium content step by step until at the end of the day, we want to actually get him to subscribe, to be a loyal user and generate the revenue from that in this way. We want a personalized experience for each one of them. The key to this is personalization, and this is exactly what we enable publishers to do.
Adam told you that walled gardens are financially successful because they know how to personalize. We have the leading recommendation engine. It's not moving anymore.
You're getting all the.
This is great. Let me give a second for this to recover, hopefully. Okay. See, test it again. Okay, now it works. So as I said, this is exactly what we enable publishers to do. Now it's jumping again. All set back there? Okay. Okay, so repeating myself. So this is exactly what we enable publishers to do. We allow them to do this. As I said, the walled gardens are exceptionally good at personalization, and that's where they're finally successful, where they're financially successful. We have built, as Adam mentioned earlier, the leading recommendation engine, and this is based on data, technology, and scale, very much like Tom showed in the presentation. If we look at data, we're working with over 16,000 digital properties, so 16,000 publishers.
What that enables us is to process huge amounts of data every single day. We've built the technology, and I'll talk about that in a minute, and that technology is driven by 500 employees in our R&D department and $100 million in annual spend on that technology. We'll mention in a minute what that includes. The scale. The scale is extremely important because we see so many users. We're using almost 10,000 servers and the trillion recommendations per month. What happened is about six years ago, we took a monumental decision and decided to build a platform. That platform was supposed to deliver value, not advertising value, but full value to our publisher partners and to our advertisers.
We came up with this. I'm not expecting everyone to go through all the details, but this is a very comprehensive platform that is addressing a significant portion of the business goals of our publisher partners. If we look at this overall, this box here, the ads, Taboola Ads at the top, this is what an SSP would do, a supply side platform. This is a very small part, as you can see, of this platform. If we look at some of the components, a major part of this is code on page. Code on page, which was mentioned earlier, means that our code is running on these thousands and thousands of publisher sites, allowing us to do multiple things.
One, it allows us to collect the data and see every user on there and understand what they're doing, how they're doing, what they're engaging. It also allows us at a later stage to adapt and personalize the experience because we're on that page, and we can adapt and show each user what they're really interested in. We have significant understanding of insights, so we know what all of these users as a collective are doing. We can tell a publisher what their audience is reading, what their audience is reading when they're not on their site, what people are interested in consuming, and what's trending right now. Leverage that with our artificial intelligence engine, our deep learning platform, and then that allows us to create significant amounts of value, which you can see at the top. It starts by building various types of experiences.
We can show any which way the content and the information we want in video, in text, exploring more in a Taboola News way. Then we allow really the publishers to generate organic value, which means show them more and more of the content that Tom mentioned that they're there for. We can allow them to grow their audience, to engage them on site, and to drive different formats of revenue and monetization through commerce, through direct advertising, et cetera. Now, couple this with our advertiser platform. Again, a similar aspect using the same artificial intelligence engine, in this case, called SmartBid Dimensions, which Adam likes to talk about, and we'll hear Gil, our Head of Algo, talk about in a minute.
Engage code on page, pixel on page allows us to see all that data and exactly understand what are the trends, what is happening, and leverage that all the way to drive audience and the entire marketing funnel from audience to engagement, to at the end of the day purchase. Now, all of this with a core component of that deep learning that always matches what the user is interested in with what the next step should be for them and how we can introduce them to what is relevant for them. Now, since the platform is so powerful, we have 15,000 advertisers using this platform and includes some of the top brands in the world. These are some stats that we're sharing here. Looking at the top 50 advertisers, their tenure with us is over 4.5 years.
They're spending, looking at last year, over $7 million with us. If we look at 2021 in general, we have more than 200 advertisers that spent more than $1 million with us. Now, it's important to note, it was mentioned earlier, that we work directly with our advertisers. 90% of our advertisers and advertising dollars work directly with us. Saying that, it's important for us to enable everyone to buy programmatically and to work through the channels that they want. We partnered and have integration with all the top players to allow that. Our goal in this is really to ensure that an advertiser can buy in whichever form, format they want with us.
Now, take all of the things I described, our comprehensive solution, relationship with the advertisers, and bring that into the publishing world. What we see is that enabled us to buy long-term relationships with publishers. We have more than five years average tenure with our publisher partners. Our contracts now are more than three years long at inception. Really all of these agreements are exclusive relationships. Again, we have long-term exclusive relationships with our publisher partners. Now, how do we bring all of this to market? We mentioned we work with the publishers, we work with the advertisers, both of them we work directly with. We segment the market into the enterprise publishers, starting with the publisher side, into the enterprise publishers. These enterprise publishers are the biggest, more premium publishers that we work with.
Those call for a direct relationship, outside sales teams, building relationships, feet on the ground, interacting, driving value to these publishers. We also have an SMB market segment, which is very efficient inside sales, working out of a hub, a type operation. Similarly, on the advertiser side, we work with an outside sales team on the enterprise publishers. These are the big brands and agencies that you're very familiar with. We have an inside sales team addressing the SMB advertisers. We also have an online, acquisition and self-service platform to work with these advertisers. All of this works via geo-based.
We have an operation serving the U.S. out of New York, serving EMEA out of London, serving Asia-Pacific via Bangkok, China in Beijing, LatAm in São Paulo, and all of these are spread within those regions to drive that value and have close relationships. Now, on top of all of that, we're now specializing. We have a specialized e-commerce team that was driven by our Connexity acquisition and now handles commerce in general. Bob will be talking about this later. We have a brands and agencies focus, a verticalized focus. Nadav will be speaking about that. Taboola News, Jonny will talk about that. In essence, a very specialized programmatic effort. What does this all bring us to? Really, with these long-term exclusive partnerships with the publishers, what we have is long-term insight into what our supply is.
We have a very good understanding of what our long-term supply is. With our direct relationships with the advertisers and always-on advertisers, since a significant portion of our advertising is performance, these are always on advertisers. You don't need to run and chase every single insertion order or order in general. We have very good visibility into our advertisers and into their business. We intimately understand their business. All of that together means predictability. That means we know the supply side, we know the demand side, we know them directly. It's long-term relationships. We have a very good sense of what the business will look like three months ahead, six months ahead, a year ahead.
Now, with all of this predictability, this is all based on our core technology, which is our deep learning stack, and this is what Gil will talk about now.
Can you hear me?
Yes.
Hi, everybody. I'm very, very excited to be here. Okay, here you go. I'm very, very excited to be using this clicker, if it works.
Let's see what happens. I'm Gil. I lead the algorithm AI group in Taboola. I've been in Taboola for a little bit more than 11 years, so I've seen the growth and being part of that has been amazing, an amazing journey for me. Before that, I did a PhD in Computer Science in the field. Today, I would like to share with you a little bit about how things work AI-wise. Was there enough build-up for AI from the previous speakers? I think there was. The clicker works. Here we go. We're gonna talk about four things. One, I wanna talk a little bit about in general the revolution of deep learning and AI on the world, and what we learned from it.
We're gonna talk about how we took that into our systems, our products. What do we do with it? Specifically, we're gonna talk about SmartBid and what we do with AI in SmartBid as an example. We'll wrap up by talking about a few things we've been working on, we are working on right now. Okay. Let's talk a little bit about that revolution everybody's talking about, that is called AI or deep learning and what it did to the world, and particularly about machines that learn to do things on their own. When I'm thinking about it, I think about what we can do with AI is two things.
One is to be able to observe the world in a meaningful way, and we'll talk in a second about what that means, and also learn a new skill. Those two things together enable us to do really amazing things. When we're talking about observing the world in a meaningful way, for instance, being able to look at images and videos and say something smart about them, like there is a car on the road for autonomous cars or for other applications of such a thing, or when you wanted to take faces of people in an image. All those kind of things are now relatively straightforward to do with deep learning.
What's amazing about it is that, up until a few years ago, people were doing computer vision, which is looking at images and getting information from them using various techniques, using traditional machine learning techniques, and within a few years, all those techniques became completely obsolete, and no one today is doing computer vision not using deep learning. When we saw that, we said, "Okay, this is interesting." This was a trend that's been going on for a few years now. Same thing with everything to do with language. With the natural language processing, being able to get semantic information from text or translation or text to speech, stuff like that. There's even a very cool project called DALL·E, where you can take text, and it creates an image.
These images were created by AI on its own, so stuff like that. Useful stuff like that. This same thing, people used to use all sorts of techniques, and today everybody's using deep learning. When we saw that, we said, "Okay, we need to do something with that in Taboola." Adam mentioned earlier the difference between machine learning and deep learning and the promise that. We'll forget about that. I don't have BS on the slide. Sorry about that. There are many differences, but in high level, I wanna pinpoint one. In classical machine learning, the flow is different. The way you work the problem is different. You take the input, let's say an image, and you need to do some what we call feature extraction.
You need to take the image and, for instance, segment it, so you know where an object might be, and then send it to another algorithm to tell you what the object is. That's very limiting, whereas in deep learning, you take the image, and you throw it into the deep learning algorithm, and on its own, it will find all these things autonomously. It sounds like it's easier. It sounds like I'm lazy, but actually, what's amazing about it is that it is able to find patterns that you didn't think about before, and we'll see some examples about that a little bit later on.
The work here is now to design a deep learning algorithm which is able to capture those things automatically without having to manually tweak it at all. For instance, deep learning, we use this term called neural networks, deep neural networks. I'm not gonna explain details about that, but it's a bunch of neurons. It's inspired by the brain, and information flowing between them, and they learn things. For instance, with deep learning, you take an image, and on its own, it learns where the lines are, it learns where the face starts and finishes, and slowly it makes it more and more granular. It'll detect where the eyes are, where the ears are, and finally it will get a full idea of what the picture is, and it does it on its own.
We wanna utilize that. A second thing is that really drives AI to be able to teach itself things is the ability to learn a new skill, and that's using a term called exploration. Let's say that I have this maze, and I want to solve this maze, this agent. Okay, we're computer scientists. We have weird terminology. We call things like agent, and we use terms like exploration, exploitation. Sorry about that, but this is the computer science lingo. Let's say that this agent or player wants to solve this maze, and we're looking at it from the top, so we can see it, but it can't. It can only see the next step, and he wants to figure out. What it does, and this is how AI works, it explores.
It tries right, left, and it tries to find a solution. With deep learning, you can do that in a smart way, in a way that will let you learn from one maze to many other mazes, and it's called generalization. You can take an agent. He learns a maze. It learns the concept of a maze and how to solve it, and it learns it also for new mazes he didn't see before. That's an amazing. Together with deep learning, that's an amazing ability to be able to learn a new skill using these paradigms called exploration, exploitation, where you sometimes explore and try something new to learn on your own, a skill, which is a very powerful tool.
Those things together, the ability to observe the world in a meaningful way and learning a new skill, gives the ability to develop amazing things like autonomous cars or beating a Go expert without having been taught Go, so learning it on its own. The way we see Taboola and the discovery problem is in a similar magnitude because we also need to observe the world, which is publishers' content and understand it in a meaningful way, and also to learn the skill of finding matches, what works for users, and doing personalization. Let's talk a little about the how more specifically about how it works. First of all, what are we trying to solve? We have a matching problem.
We're trying to match millions of content pieces with users with many, many users every month while we are making our publishers and advertisers successful at the same time. That's like the high level problem. We want and need to do that at scale. First of all, the scale enables us to do this. Without the scale, we wouldn't have the data. Data is crucial for any deep learning algorithm to be able to be established. We have a lot of it, so this is an opportunity, and it's great.
On the other hand, this also makes it hard because you need to have a platform that is able to capture the insights and do matching for 1.3 billion users and 100 TB of data to crunch every day, which is a lot of data. You got to do it very fast. You got to do 30 million predictions per second, and every prediction needs to happen within a few dozen milliseconds. It's a fairly challenging infrastructure to work with, and we do it. Second thing, another element of scale is what we call experimental throughput or the way we experiment. We train on software. We have 200 deep learning models trained every day. That's a lot of models.
We do that in order for us to be able to iteratively improve all the time. To try new things and to experiment with them on live users every day, which is another scalability issue, to be able to train a lot of algorithms at the same time and run them in a production system. Let's talk a little bit about the team. We are the AI team. We are 85 research engineers, and we like to think of our research engineers as two-headed beasts. One head is a data scientist, researcher, knows about machine learning, all the stuff I talked about before, probably better than me, which is good. The second is a software engineer who knows...
The second head, the same person, is able to code that into a production system and run it at scale. That's super important because in an ever-changing world like ours, where the publisher data changes all the time, people talk about different things on the internet. The advertisers change all the time. We have new ones, and they change their content all the time. We got to experiment with things online very quickly. We built this paradigm where we do some research, but don't stay with the research in the research phase for too long, but go and put our ideas in front of real users fairly quickly. For that, you need people to be able to both do the research and get the code into production and make it work.
Via that, we iterate, and we progress. We use and we are actually a part of an open source community of tools which are quite top-notch, I think. For instance, TensorFlow, it was originally developed in Google. Now it's a pretty big community. We're part of it. Spark, other things, to be able to do all this tech in a great way. We didn't start with deep learning. When I started, deep learning wasn't really a thing. We started with Bayesian inference models, the statistical models. Then we moved to do machine learning in a traditional way. In 2017, we decided strategically we're going to move into deep learning.
We got in relatively early, which is great because we built that skill. Today we are in a place where we can solve many problems with deep learning, and we really feel like we're established there. It's also worth saying that getting into that is very hard. We talked about the scale. You got to have the know-how. You got to have the computing power, which is tough. I don't believe there are many companies that made it all the way to being able to run deep learning and do AI in that scale. This is my intuition, and I really feel very fortunate that we were able to do that for us.
The impact of that is very, very significant. We see dramatic improvements using our technology on RPM revenue for publishers, on advertiser success metrics like conversion rate and CPA and on organic recirculation performance. The reason I know that is we keep track of it. We always take a very small portion of the traffic, and we serve it with a relatively simple but traditional machine learning algorithm, and then we track it over time. We see that we are okay and we are progressing and becoming better as we go along. All right. Let's talk a little bit more in depth about and with some examples about how does deep learning fit in with Taboola. First of all, the frame, the general framing.
The general framing is as following. A user opens a page on the open web. We know stuff. We know the context where they come from. We know what they're reading about. We know the time of day and where they come from in the world and stuff like that. We have access to the curiosity graph that Adam mentioned before. We have millions of possible recommendations, and we need to rank them. The way we like to think about it, to rank them by establishing this thing called estimated CPM, which is what an advertiser would have bid if they had the access to all this granular information on their own. We do it on their behalf. It's not what they're gonna pay. It's just a score.
That's what determines what's gonna be the winning ad, what are we gonna put there. This is the challenge. It breaks down to two things. We need to estimate two things. This is for the ads product and just an example. It can translate to other problems. We need to estimate the click-through rate, which is the probability for a user to click. In Taboola, vast majority of the demand is in a click, in cost per click. If user didn't click, there was no money exchanged anywhere. Very important to be able to estimate the click-through rate.
We have the cost per click that the advertiser gave us, and we change it according to, on their behalf, according to what is the likelihood for the user to buy something on their site, to actually engage with the advertising to purchase a product. That for us serves for doing good both for the publishers driving revenue and for the advertisers, giving them users that are actually going to engage with the content. Together that becomes that eCPM bid that we do, on behalf of the advertisers. This is the general framing. Let's talk a little bit about what it captures. I can talk in general and in theory about all sorts of different things.
I think the easiest way to look at it is via examples of what SmartBid is able to capture. We looked at the data. We took data from a couple of weeks earlier this month, and we wanted to see what SmartBid captures. First example is people reading about COVID-19. When they go on publishers' sites and they read about COVID-19, SmartBid discovered that they have high tendency to engage with travel campaigns. Travel campaigns will work quite well for them. What did it capture? Probably users are planning to travel, and they're going on sites and checking out what's going on COVID-wise, to understand what's going on. Then it's a good time to show them travel campaigns. We do see. SmartBid likes them a lot.
The ranking score goes up as opposed to an average context, and the CPA goes down. CPA goes down means good thing for the advertiser. They will pay less because the users are more engaged in general. On the other hand, they will have lower tendency to engage with financial, personal investing campaigns. Probably in that context, they feel a little bit like things are a little bit shaky still, COVID-19, the pandemic, and we see that the algorithm ranks those ads lower. If we would have showed them, the CPA would go up by 50%, which is a lot. This is not a winner ad in that context.
How do I know that the CPA would go up and it's a bad thing? It's because we explored that. How does this work? First of all, it's a little bit like SmartBid understood the properties of the pandemic. This is like really exciting when you look at it like that. It could change over time, but this is what it's seeing right now. The reason it was able to capture that is using this paradigm of exploration. It tried things on different contexts, and it saw what works. First of all, it discovered bad examples, like showing an ad that won't work so much in that context, but it wouldn't find the good things if you...
If it wasn't trying and exploring the bad things as well. That's for us like the solving the maze. Okay. Finding those patterns and being able to match users with the content in an efficient way. Another example. I'm fortunate to have. Okay, there you go. People reading about sports. People in that context, when you're reading about sports, not a very good time for you to get a campaign about books and literature, like a big, heavy novel is not a good time for that. We see the ranking score is going down. The CPA would have been up, which is a bad thing. On the other hand, lighter subjects like golf campaigns or stuff about scooters.
Call that hobbies and interest. In those specific examples we show them highlighted when we looked at the data, it worked really well. And that's also something that SmartBid understood. Another example. We looked at trends throughout the week. On the X-axis, we see the day of week, and the Y-axis, we see the ranking score that we have for these three categories compared to the average score they would have had. We see a few really nice things. First of all, campaigns about fine arts are really trending well towards the weekend, people have more time to look at that. It makes sense. On the other hand, people looking at career campaigns, it's actually they're doing it while they're at work.
They are doing it midweek, and towards the weekend, they are going to other things, which are probably more fun things. Another very kind of cute example is the campaigns about events and attractions. We see that they are trending up towards the weekend, but on Friday they really peak. On Friday, like, people are really, really interested in buying tickets for something for the weekend. On Saturday, they still have a chance to do it, but on Sunday it goes down. This is it, weekend's over. Not a good time to buy a ticket for an event. We saw that in the data. Another. That was another example. Last one of this kind of this type. Something which is even more pinpointed.
We looked at there was an article about the Amazon stock split, which is both about a tech company and about the stock market, usually in the tech section of the sites. We saw that SmartBid really went all in with the stock investing ads, where the score went really high. Just to note that this is a difficult problem to do. To be able to capture that signal very quickly is very difficult because that article or that story is going to be live and showing up in the internet in high numbers for a day, and then it will be different news.
It's very, very important to also capture those signals very, very quickly, which we are able to do, which is good fun. If you think that those examples were a little bit simplistic, it's a bit difficult to show under the hood what happens in a fairly complex model, then let's have a look at under the hood in a different way. First of all, it is indeed a challenge when you're working with deep learning and neural networks. They are fairly complex, and it's difficult to understand. They are a little black boxy. It's difficult to understand what's going on inside.
What we like to do is to plot some of that model on the screen and see if it captures something intuitive, and it gives us a sense that it's doing well beyond the revenue numbers and the CPA numbers. We plotted on the screen those titles of ads, and we wanted to see what it captured. We took the title How Diverse Technology Will End Our Water Crisis, and we wanted to see what the algorithm thinks are similar titles that might work for them. That might work for users who like this one. Let's see what it found. The first one was How Diverse Technology Will End Our Water Crisis. I really like the example on the top over there.
How smart meters will help meet future energy demand in cities. Let's see what it captures. First of all, the wording. We didn't tell it anything about the phrasing, how will something affect something. It understood it on its own, that it is similar and similar users will engage with them. Also, here it talks about diverse technologies. Here it talks about smart meters, a specific technology, and how it impacts future energy demand in cities, which is another issue in sustainability, like water crisis.
Quite subtle things that these algorithms can capture, which is a very powerful tool because then if I know stuff about the first one and let's say that another ad was just brought up into the platform, I can already say something smart about where it's going to perform, which user's gonna like it and so on. It's a very, very powerful tool for what we call generalization. Taking information that I gathered on one thing and learning about it from to another thing, similar to the maze example from before. Taking one maze and learning to solve other mazes. This is our maze.
Before I wrap up, a few things. We built this as a platform, so it is relatively easy for us to get into a new problem. For instance, when we got into Taboola News or e-commerce, we already had a platform set up, and a head start for getting it. Now we can focus on actually engineering the algorithm itself and not having to worry about coming up with a platform. It gives us, again, it is relatively easy for us to get into new problems, which is a very, very powerful thing for us. Finally, I want to talk a little bit about some things that we've been working on. We talked a lot about contextual intent.
Those examples, for instance, the Amazon stock split example, it was actually what we did. We did like retargeting without third-party cookies, just by using contextual information. We want to be able to to do that in a higher scale. We want to see that happening more and more. We want to do it faster. Also, we would like to allow the algorithm to capture insights and help advertisers of many kinds. Eldad spoke a little bit about lifetime value. We're modeling that for publishers, so we're working on that too. And finally, we're also using AI, improving our bidding capabilities. Recently we started working with Microsoft on a bidder that is able to bid on any supply. We've been working on that too.
A bunch of things that we are able to do with this AI technology. I gave most of the examples for SmartBid, but it really applies to many, many things. To wrap up, I think we are in a place where deep learning and AI is in a mature enough place where we can do really amazing things, and we see the industry coming up with these amazing tools and technology. We feel like we can do the same for the open web, making it successful by having products driven by AI and deep learning. I thank you all. I would like to invite Evangeline to talk a little bit about editorial tools also using AI. Thank you.
Let's test the mic. Can you guys hear me? Okay. Hi, everyone. Good morning. I'm Evangeline Axiotis, Director of Publisher Account Management. I've been with Taboola for seven years now, and I remember in the early days, Adam and I would be in his office working on QBR decks over many espressos. I'm happy to say that over the years, we've built out a really strong and successful account management team. Adam no longer has to join every QBR, although I still get the occasional FaceTime from him asking for updates. Yeah. The account management organization at Taboola, our responsibility is to drive growth for our publisher partners.
We do that by getting to know their business, their needs, listening to their goals, consulting them with the best way to use the Taboola platform, and ultimately, creating a feedback loop that then drives our product roadmap. Today, I'm honored to be joining the stage with Kristin Roberts, Chief Content Officer for McClatchy, one of the longest standing local news organizations in the U.S. We're really excited to showcase our partnership because it's a great example of how newsrooms are leveraging Taboola to make data-driven decisions and maximize lifetime value for their users. With that, we're gonna roll a short video on the Newsroom product, and then I'll invite Kristin to the stage.
Good morning. I'm Tim Ruder, and I manage a team of editorial engagement managers at Taboola. My background and the background of this team includes substantial careers with editorial staff at some of the most trusted and innovative news publishers in the world, including The Washington Post, BuzzFeed, Newsday, Time, Associated Newspapers, and more. Our substantial experience in editorial and audience development teams at premium publishers let us guide journalists around the world to understand and use audience metrics more effectively through our Taboola Newsroom Insights platform. The day in the life of an editor of a news website is chaotic, to say the least.
Following constant breaking news updates on Twitter and Slack, copy editing incoming stories, writing great headlines, making sure the politics team and the national team aren't sending reporters to cover the same press conference, rescuing the business from Facebook and Google's latest algorithm changes, pivoting to and un-pivoting from video, and maybe even finding time to warm up the lunch you'll inevitably eat at your desk. At Taboola, we know not only our success as a business, but the success of society relies on these journalists' ability to bring order to the chaos. That's why we built Taboola Newsroom. Newsroom is a real-time audience analytics and insights platform provided free of charge to our news partners that aligns the interests of not only Taboola and news publishers on the open web, but also the public at large.
Using our state-of-the-art AI, informed by data from over 500 million readers daily, Newsroom provides actionable alerts and recommendations to thousands of editors and journalists at the exact moment when there are opportunities to optimize that user experience. We're not trying to turn journalists into data scientists. Rather, by packaging data in ultra user-friendly visualizations and reports that fit seamlessly into the editorial workflow, Newsroom lets editors, reporters, and content strategists impact the three metrics that drive the publishing business. Engagement, increasing the time users spend on a publisher's website. Audience, bringing more readers to the publisher's content. And revenue, growing the revenue from that publisher's audience. Let me show you some of Newsroom's expansive feature set that help publishers drive these key business metrics.
Increasing article page views and engagement with easy to understand traffic and engagement metrics and actionable alerts and recommendations at that moment when those opportunities to optimize and grow the audience occur. Improve audience engagement with better headlines and images using simple and fast A/B testing tools. Choose the right stories to feature on the homepage. Move stories higher or lower, and understand when a story is past its prime and should be replaced with something new through an easy to read overlay of the homepage. Understand what topics are trending nationally and among a site's loyal readers, both on-site and across the Taboola network. With this understanding, editors can respond to fast-moving news, as well as find new areas of coverage that their readers will be interested in.
Some publishers, like The Independent, have created new sections on their site and hired additional reporters to expand coverage based on these insights. Grow subscription and reader revenue through both content strategy and AI. By providing real-time information about which articles are leading to conversions and which aren't, editors can take this into account when deciding if an article should be behind a paywall, which articles are better outside the paywall, and what content is best for engaging existing subscribers to reduce churn. With Taboola's machine learning algorithms, AI is able to automatically program organic recirculation based on propensity to subscribe scores that take into account both content and reader behavioral attributes. Apply Taboola's AI and deep learning to improve the homepage experience for a publisher's best and most loyal readers, and much more.
Many of the world's highest quality news, sports, entertainment, and lifestyle publishers rely on Newsroom as a foundational tool in their daily workflow because it impacts the three metrics publishers care about, engagement, audience, and revenue, while supporting the editorial mission and editorial values.
Let's fix these.
We have some high chairs.
We have to turn them a little bit.
Yeah.
We're gonna look at each other. I'm very short. Why did you give me this tall chair? It's very tall. I'm glad I'm not wearing a dress.
Thank you for being here with us.
It's my pleasure.
Kristin. McClatchy is defying the audience trend right now. Can you talk to me about what you're doing to achieve that outcome? What role is AI and technology playing there?
Okay. We are growing audience, and we have been for months now, and that's a really unusual statement to make and to hear in local publishing. You know, at this point, at the end of March, we have stacked up eight consecutive months of growth in both page views and uniques. I'm not talking about small gains. I'm talking about large double-digit percentages. Nobody's putting up numbers like that in local. Nobody's putting up numbers like that in digital writ large right now. At this point, 65 million uniques per month on average, and it's a direct result of a set of strategic decisions that we made a year ago when the entire industry was really looking at an environment of declining audience trend lines.
We rejected in that moment, this idea that there was some inevitability to the post-COVID slump in readership. We instead decided to make a series of targeted investments in news products and in news experiences. Those resources were deployed based on data. Based on data about what our users wanted right now, not generally what they wanted in the last few months, but what they were telling us based on their behavior on our platforms that they needed from us. You ask about the role of technology. Data and technology, automation and AI were critical. They were absolutely core to our success in growing audience, really in two ways. One, making sure that the content we were producing in that moment was immediately relevant.
Second, it allowed us to make sure that the human beings in our newsrooms could focus on things that truly differentiated us, real journalism that a bot can never produce.
It sounds like you're innovating rapidly. What, you're making big bets on content and product. How are you able to do this?
Well, we're different. We are intentionally different from what other people in our space are doing right now. I mean, in the last 18 months, we've executed a full transformation of McClatchy. We are today growing EBITDA. Just to put some context around that, at the end of 2021, we threw off more EBITDA than we did even in 2019, right? Our trailing 12-month EBITDA as of this moment is north of $100 million, and we're doing it at a 20% margin. A 20% margin in local publishing, all while injecting literally millions of dollars back into our newsrooms and into our news products and experiences. That's because under new ownership and new leadership, we've organized the entire company around one concept. It's the idea that the future of digital media is tied to audience growth.
In order to achieve that, we've got to be thinking differently. The way that we visualize it is as a flywheel that begins with our content produced by our newsrooms, supplemented with our partners' content, spun out into multiple products that create consumer revenue engagement opportunities, the data from which creates advertising opportunities. All of that grows EBITDA and gives me more money to put back into news. That's our cycle, right? That's our flywheel, and it's the vision that we crafted a year ago and that we have now turned into reality.
What are the biggest opportunities that you see for your editorial teams, and which areas are you focused on to drive that mission, that flywheel?
Well, the opportunity really is in multipoint monetization of every individual customer. To step back first, you asked very specifically about editorial. McClatchy's mission, our very purpose, is independent local journalism that serves the public. Our newsrooms create journalism that is essential in the lives of our readers, viewers, and listeners. That's why people come to our platform, right? That is our core competency, it is our differentiator, and the task right now is to amplify that. In order to amplify that content, we need to think differently about how we're doing our jobs. This entire so media industry has been focused on a very simple view of audience for more than a decade, and it's visualized as a funnel. I'm sure many of you have seen the audience funnel in media decks, right? Looks like this.
At the top of that funnel are what I call one and dones, drive-by readers. They come in for a story, and they're gone. You monetize them with programmatic ad revenue. Cool. You want to try to pull them down the funnel a little bit toward a subscription, which has a higher lifetime value. Great. The problem with the funnel is that it's actually a sieve. It's got all kinds of holes in it. People will tech trick your paywall, and they always will. They'll come in for an entry-level subscription offer, and then they will churn out predictably. Frankly, in some local markets, the base of potential subscribers is way too small. It's way too small to ever, on its own, support that individual business unit.
What we're doing is trying to build a runway of repeatable revenue-driving engagement that starts with increasing the size of your audience and then collecting revenue all throughout the journey that they take on your platform. Let me give you an example. We just talked about sports a couple minutes ago, right? Let's say a user comes to one of our sites for an NCAA tournament story. While they're there, Taboola kicks in with AI that creates a piece of relevant content, a second piece of relevant content. We present an opportunity for them to explore the betting line from our gambling partner, then the opportunity to buy team merchandise, and then another piece of journalism that could be from our core newsrooms or from a partner like Sports Illustrated. Bam, the subscription offer for the niche sports product. That's the future.
It's not subscriptions or advertising, right? It's leveraging a digital platform to create full experiences driven by AI that is amplifying great content to do two things. One, to drive diversified B2B revenue streams, and two, to retain subscribers. That's the opportunity.
You touched on this a little bit, but could you expand more on where you see Taboola fitting into this space?
Taboola is a partner. This is a partnership that we have, and I really do use that word intentionally. I'm not interested in transactional relationships. LinkedIn is the worst because all I get are salespeople with their opportunities for McClatchy, and it's all transactional crap. What we have with Taboola is an opportunity to experiment and iterate and play around and try to get from here to here. That's why it's exciting, and that's why we're working together, and I see us working together for a very long time. Right now, we've got three things going. We've got content rec. We've got the homepage personalization project. We're doing A/B testing on headlines as well as the analytics dashboard. It's been a few weeks. So far, we're seeing some results on the agreed KPIs.
I think the next couple of weeks are going to be very important. We need to get through user acceptance testing on the homepage product. We're gonna be sending live content in, what, the next 14 days. At that point, once we've worked out all the kinks and mastered that piece, I'll be ready for us to say, "What can we do next? How do we take homepage beyond the homepage? How do we apply this to our users who are out of DMA rather than just in DMA?" It's all opportunity.
You mentioned homepage personalization. What is the role of the homepage for your digital properties?
It's such a fun question. The answers are so different all the time. Homepage personalization. Okay, I hope that you will allow me to, Adam, share this story. I called Adam on Thanksgiving weekend, and now that I say that out loud, it was an incredibly thoughtless thing to do, so I'm sorry.
I didn't know what that means.
Cool. I was obsessing. In fact, I think the entire McClatchy leadership team had been obsessing for a while about this idea of consumption intent. The social platforms brilliantly present to members of my family exactly the right product at exactly the right moment. Fortunately or unfortunately for me, it typically yields an immediate sale. Why can't my platforms as adeptly intuit what my users want to consume around content? That conversation is what launched the McClatchy-Taboola partnership on homepage personalization. Our loyal readers, the ones who have a stake in a community or an issue, they're homepage readers. Now, of course, I want to hold on to everybody. That is why we're investing in trying to create better experiences no matter how you come into our platform.
If you are bookmarking one of my sites, you are indicating a level of interest that raises the stakes for me to have a high-performing product, right? So for us, the homepage is an opportunity to invest for subscribers and also for non-subscribers. For subscribers, we have an opportunity, and we are creating an altogether different experience that elevates subscriber-only content and increases opportunities for engagement around content types that we know from the data those individual users want. For all of our users, subs and non, we try to remember that our homepages are portals to individual communities, right? Remember, our core competency is local, and so the people coming to us want us to act as a gateway, and our sites need to create space for both news and information, but other types of content as well, a lot of it from partners.
That's going to help us serve people in DMA and out of DMA. There's a lot of investment in that space and a lot of good momentum there.
What other challenges does Taboola's Homepage For You product address for you?
Listen, Taboola has built smart technology, and the Taboola team has shown a willingness to tailor it and adjust it based on our needs, and that fits right into my strategy, right? I am a platform. I provide content. I know what I do better than anybody else. We do the best local journalism. We also know when we need to go out and find somebody who does something better than us, and that's where Taboola and homepage personalization come in. You know, we believe that this tool is going to best intuit what our readers want second, third, and fourth. If we get that right, of course, we're gonna grow audience, which is going to drive revenue, and it's going to increase EBITDA. But more importantly, we're going to create better digital experiences around content that is both high quality and trusted.
That's really what this whole thing is about.
Thank you, Kristin.
You're welcome.
With that, we're gonna roll a short video on our Homepage for You product, and then I'll kick it back over to Eldad for the next session.
Cool. Thanks.
Is everyone enjoying the video?
Okay, very quickly, just a housekeeping note. We're not having a break, so anyone who wants to grab coffee or drinks, there are some at the back. Please do so at any time. Or food outside, feel free at any point in time. To summarize, we covered the core piece. What you've heard is about Taboola's vision to be the recommendation engine for the open web and to connect billions of people with millions of pieces of content, and that is our core Taboola business. You've heard that it's a $64 billion business, $64 billion dollar market opportunity. You've heard that there's some complex technology to build into this. You've heard about direct relationships with publishers and with advertisers that drive this. You've learned about the deep learning technology that powers all of this.
What we're now going to talk about, you heard Adam speak about Taboola Anything & Anywhere strategy. We'll start covering that. This is a multibillion-dollar opportunity. The Taboola Anything is really to recommend anything. It could be content, it could be products, it could be solutions, it could be people, it could be anything. Taboola Anywhere, we want to recommend anywhere the open web exists. It started with doing this on your PC, continued with doing this on the mobile web. It will be in connected TV. It will be in autonomous car. It will be in your connected refrigerator. This is the strategy. What we're going to hear about is brands and agencies. We're going to hear about e-commerce.
We're going to hear about Taboola News, and we're going to start with Nadav talking about brands and agencies.
Thank you. Let's see the video. Video, please.
Your success is our success. We cannot succeed without you succeeding first. By educating key accounts, but also the general advertising public about what Taboola can bring to the table, we stand to grow together in the years to come.
We generally utilize Taboola for two different reasons. Number one is because the quality of the clicks that drives on it from a cost per engagement perspective, and then also what we get off the back of that is a lot of impressions which feeds our prospects.
We're always innovating, and their native advertising, or as I like to say, endemic content, really drives results.
Definitely Taboola is key in our online media mix. Metyis is focused on performance, and native is a great media to drive results.
Now, by making Taboola's platform a state-of-the-art technology with its SmartBid algorithm, we can program different dimensions in terms of campaign objectives such as cost per acquisition or return on investment to be able to personalize or have much more precision in the objectives we are pursuing in each of our campaigns.
Native is a great media to drive results. We started small and tested the results against our core KPIs. From there, we expanded to larger budgets and new lines of product.
Taboola delivers us fast campaign results. We love the most at Taboola, a performance-focused algorithm. We can reach a unique audience which is not reachable for us over other channels.
There are two main reasons. The first is that our clients, mainly direct clients, are a good match with Taboola. The second is that compared to other DSPs, Taboola's performance is more stable. These two reasons are the main reasons why we use Taboola.
Metyis is looking for performance at scale with long-term partners. That's very important. 100% of Taboola's inventory for Advisor is exclusive to Taboola, and that has a value for Metyis.
We love to cooperate with hands-on partners, and this is what we found with you guys. We can spread our campaigns across all relevant markets for ERGO.
The open web offers a, you know, it's much more measurable, good deliverables on it, and, you know, you can. Obviously, we can use all of the measurement tools to give us that source of truth, which we feel is a bit more reliable.
Many brands and advertisers and agencies don't see native or Taboola as their first choice when it comes to paid ads. I believe, strongly believe that it should be alongside Google and Facebook, not after it.
Taboola powers recommendations on more than 10,000 global publishers, so we use it on top of social platforms in terms of reach. It's not the same objective, not the same content, but we are sure to have a quality of content in this premium publishers.
I believe that Taboola is well-positioned for the future, especially the cookieless web, and that's because in the cookieless world, context will be king.
We will have to reimagine and realign how brands measure performance. That's where Taboola wins. In this new world, non-audience-based targeting solutions such as contextual targeting and hybrid management models are going to work.
User privacy is definitely a priority for Metyis. With Taboola, we can rely on key strengths, which is contextual marketing.
From the point of view of implementation, we have precise control over where we want the ads to be exposed, and more importantly, the platform is integrated with multiple verification technologies.
Very satisfied from a client perspective and very, very, bullish on the future for Taboola and how they will win in a programmatic future.
We can expect a high level of service from the Taboola team and the most innovative technology.
They really understand readership data, and they do a lot of reading and noting studies so that they can help us not only with the targeting the campaigns, but also insights about our brands.
Say two words. I would say first of all, service, and the other word is, technology, which definitely leads to a long-term partnership.
It does make us feel very confident to have this strategic partner and, of course, to continue building the relationship we have had over the last few years.
Hello, hello. Yes. Can we put the right presentation on, please? Perfect. Good morning. My name is Nadav Perry. It's so great to see testimonials from our clients from all over the world, from key agencies in the East, from Japan, from India, from partners and clients in Europe, in the U.S., in Latin America. I have the privilege of leading the Global Brand and Agency Development effort for Taboola. I've been in Taboola for three years. Before that, I spent somewhat of 15 years in the high tech industry, started my journey in small startups. One of those startups was Google. Spent about 10 years at Google. I was honored to be one of the first Googlers in Israel, opening the local office and working with clients and agencies in Israel.
I moved with Google to lead the YouTube brand go-to-market innovation for Europe, Middle East, and Africa. Three years ago, I joined Taboola, and we started really focusing on this area of growth, which is brands and agencies. I'm proud to say that today already, although we're just getting started, we are already. This part of our revenue is already 15% of our global revenue, and it's growing like a rocket ship, and we have a lot to do. What are we gonna do today? Today, we are going to talk about why should brands care. We're gonna start there. We're gonna talk about how we are investing in brands from a Taboola perspective, and we're gonna hear from our clients about success stories. Why should brands care?
Well, brands should care because consumers are on the open web. We talked about it in the previous sessions. Consumers are spending their time on the internet doing many different things. They're having many different moments. They're having moments of entertainment, right? When they're scrolling a feed on Instagram or they're scrolling a feed on TikTok. When I look at my daughter, it's more of moments of boredom, but it's moments of entertainment nonetheless. They're having moments of intents all the way down in the funnel when they know exactly what they wanna do, when they know what they wanna buy, what they're searching for, they're having moments of intent on Google Search, on Amazon, et cetera. As you know, they're having many, many moments of discovery.
Moments where they're genuinely curious about what it is that they might read about, what it is that they might find out about, and those moments of discovery are where Taboola has a part to play. Brands should care because consumers are there. We've already read this stat before. Consumers are spending about 25% of their internet time on the open web in those moments of discovery, and brands have an opportunity to engage with those consumers in those moments through our platform. Now, we coined that terminology, which is moments of discovery, because this allows us to present a native experience, which is a discovery experience. The moment of discovery is one that has a lot of psychology behind it because the brain, the human brain, is actually more receptive at that moment.
You're reading an article, you're at the end of it, and now you're looking for what's next, right? It turns out that the brain is more receptive. The cognition load is lower and the receptivity for new messages, including brand messages, is higher. In fact, it is 25% higher. This is a Nielsen study that we've done a few years ago, which shows that ads that are placed in those moments of discovery generate 25% higher attention. What you see here on the screen is an eye-tracking exam. We put special glasses on people, and we told them to look at the screen, and what you can see on this part, right? This is a video app, and when this video ad is presented, actually the user's gaze is not focusing on the ad.
The user is focusing on something else because they didn't go to this platform to look at the ad, and it's not a native experience, it's an interruptive experience. It's a pre-roll. However, when they look at the same ad in a native experience on a premium publisher website in a moment of discovery, the attention is much higher, and you can see that from the exam. Not only is the attention higher, actually, consumers that are looking at ads in that moment, in that moment of discovery, find those ads 2.4 times more favorable. Not only brands have an opportunity to engage with the consumers, not only those brand messages get higher attention, but are actually more favorable. Why should brands care? Brands should care because consumers are there.
They have an opportunity to engage with them in a way that drives more attention and that is more favorable. Now, this entire thing has been exacerbated in the pandemic era. Recently, the IAB just produced a survey called The Value of Quality Environments, and in that survey, the IAB says ad perception on high-quality sites are 74% higher versus the same ads on low-quality sites, right? In the pandemic era, 77% of the people said that, actually, they learned to value good journalism more. The quality of the site and the premium-ness of the publisher does matter to brands. It generates an environment where they can flourish in delivering the message to their consumer base. I hope I convinced you that brands should care.
We care, and because we care, we've made a lot of investment over the past years, and I wanna talk to you about how we invested in our product, in our partnerships, and in our people. I'll be brief, don't worry. Product. We've actually you know, we're deeply rooted in performance marketing. We heard from Gil how smart data and our algorithm always optimizes to customer results. Over the past years, we've invested in a full funnel solution. We can help advertisers meet their awareness objectives all the way up on the funnel, as well as their performance objectives, driving conversion in the bottom of the funnel. Let's deep dive into that just a bit. We've recently packaged a solution that's called High Impact Placements.
These are premium ad placements on premium publishers with high visibility, brand safety controls, and are great and are optimal really for achieving reach and brand awareness objectives. You can see an example here on the left-hand side, how visible the ad actually is and how much time the consumer gets to interact with it. We don't forget where we came from, and we're also providing continuous innovation on the performance side, right? You've heard about our data-driven algorithmic solution to actually focus on performance outcomes using image ads, motion ads, video ads, and really optimizing towards the relevant KPIs. If you think back to my introduction, I said that I spent quite a lot of time in Google and YouTube. The equivalent would be YouTube for awareness and search for performance. This is our world at Taboola.
We have High Impact Placement to drive awareness results, and we have performance marketing solutions to drive lower funnel conversion results. We're always innovating. We don't stop at the placement, and we don't stop at the solutions and the packages. We're always creating experiments and innovating, and what you can see on the screen is a new experiment of rich media ad experiences to really drive awareness. Some of our clients are already experimenting, and we're scaling this up, and you're seeing it here on the slide, probably for the first time. Not a lot of people have seen this. We're driving experiments to drive awareness. We're also innovating for performance, right? How do we help go beyond the ad? How do we create experiences that help a user engage with the advertising message?
With ads with description, we're allowing the advertiser to actually storytell, to get more qualified users after they've understood more about the brand message. With call to actions, we're helping the users understand exactly what's gonna happen after the click. Remember, if the user didn't click, they didn't pay. The brand didn't pay. With these tweaks and capabilities, we're increasing, for example, if an advertiser is driving purchases, we're increasing conversion rates by 26%. That's a lot. That's a big lift. We're increasing lead gen by over 6% on conversion rates. Always innovating, both on the awareness side and on the performance side. Moving from product to partnerships. We know that brand safety has become a paramount requirement for many brands around the world.
We've partnered with the top players in the brand safety space, with IAS, DoubleVerify, Moat, and others, to make sure that advertisers not only can trust our own technology but also verify it with a trusted third party. We have those verification built into our system, if they choose to work with Taboola Ads, or naturally, if they choose to work programmatically, they're built into the DSPs. Beyond the third-party verification, we've also built our own controls, where an advertiser can decide to show their ads not adjacent to any other advertising content, so to create an adjacency control unit where their ad is placed between organic content. Their ads are safe on the Taboola network, and they can feel that they have the best way to deliver their message.
We've also partnered with the IAB. We recently announced that we got the IAB Gold Standard 2.0 certification. This is the highest industry standard that exists for a partner in terms of premium brand safe environments and an environment that's respectful for user privacy. We're continuing to partner with more third parties to really position Taboola in this space as a brand safe and premium solution for advertisers. We have more. Eldad talked about how we value the partnerships with publishers. We heard from McClatchy how they see that partnership. We don't stop with publishers. We're actually expanding our canvases to new screens, new touch points between advertisers and their consumers.
We're gonna hear from Jonny very soon about how we're working with OEM and device manufacturers to create new engagement, new touch points, where brands can tell their stories and reach millions and millions of new users through partnerships with many device manufacturers. Finally, our people. We truly believe that we're privileged to work with the biggest brands and agencies in the world. You've seen them on the video. We invested in our people to be able to give them the best service possible. We have industry experts now on the team with people from the auto industry, the retail industry, the finance industry, to really be able to work with our clients in the best way possible. We have agency leaders that can work with agencies and really understand how the agencies tick. Client success leads.
We have creative experts to help build the best messages and creatives on the network. We have analytical experts and many more. We're building a very strong service team to help our partners. That allows us to go beyond just selling media and selling ads. That allows us to really take our data. Remember, we're sitting on mountains and mountains of readership data based on thousands and thousands of publisher partnerships, and that data can be translated to insights, right? Insights that go beyond the media buy-in, insights that can help advertisers craft their stories, design their creatives, target specific audiences that are more likely to convert. All of this is what the team does all over the world in all our offices with all our partners and clients and agencies. We started with success stories, and we'll end with success stories.
I mean, we have been able to build relationships and partners, and partnerships with agencies all over the world. I could tell you about how we work with Xaxis globally. I could tell you about how we work with Dentsu in many, many different markets. I could tell you about the Stagwell Group relationship that we have in the U.S. and many more. I could tell you about direct clients. I could tell you about how Dell is using Taboola across multiple markets to sell their computers to businesses through the Taboola network and how Kia has expanded their relationship with Taboola to drive brand awareness in many, many markets. Instead of drip-feeding little examples of many, many clients, I'm actually delighted and honored to invite Lule to the stage.
Lule is the CEO of eToro in the U.S., and I'm gonna invite her for a fireside chat so you hear from her and not from me. Lule, please join me on stage. Thank you very much. If we can get her set up here.
Hello, everyone.
Hello.
Hi. Where would you like me to be?
We're setting up a chair for you.
Oh, okay. I don't like standing either. Oops. Sorry.
Thank you very much. Perfect.
Thank you.
You're welcome. Cheers.
Thanks.
Okay. Stable. Good.
Yeah.
Welcome. Good morning.
Good morning, everyone.
All right. First of all, congratulations on the Super Bowl ad.
Thank you.
It was amazing. It was very, very good.
You are officially part of the Crypto Bowl, I guess.
I know.
Yeah.
Look, I looked at the stats today. 16 million views on YouTube, and probably millions and millions more.
Indeed.
On the big screen and on your website. For those of you who are not familiar with eToro, why don't we start with a bit of a brief explanation of what eToro is?
My pleasure. I love talking about eToro. Good morning, everyone. My name is Lule Demmissie. I'm the U.S. CEO of eToro, which is an online fintech for individual investors. It is designed to be a multi-asset platform. We offer crypto, over 25 coins and growing. We offer stocks we just launched recently. It's a global company. We're in many countries, including in many parts of Europe, Asia, Latin America, and now the Americas. It's the secret sauce, the differentiation for us really is we are very much committed to a social framework for investing. Actually, when you open an account with us, you actually get a social handle. It's, you know, whatever you call yourself is what your account is at eToro.
You engage in our platform and our social platform not only to invest, but to converse with like-minded people who are interested in investing in different assets. We have about 27 million users now, and growing. It is an exciting platform to be part of the future.
Amazing.
Yeah.
Look, I think there's a lot of things in common. Look, we talked about how Taboola is bringing the power of personalization into the open web.
Yep.
eToro is bringing the power of social into the investing world.
Indeed. It's an exciting venture.
Okay. Look, we have been working together as two companies for years now, and we have a joint business plan that is really promising for the future. I know you use many different marketing tactics to drive new customers onto the platform. You mentioned 27 million, and there's many more to come.
Yep.
You measure success by a metric that you called FTDs, so first-time depositors, among other metrics.
Yep.
How could you tell us how Taboola is helping you reach those objectives?
For us, it's really important to have a partner not only that has the media assets, the creative assets, and the quality of partnership we have, but it's really important that it is a data-driven venture. It is impossible to be a challenger brand like we are without using data as your mechanism to really break through. The Taboola partnership has been very valuable because it's such a strong data-centered, content-driven partnership, and that is rare to find the firm that does both so well. One of the reasons I think it's really been a successful partnership is we are very much performance driven. Taboola is really very strong in this space, and it's been a symbiotic partnership.
Great. We talked a lot about innovation today. We heard from Gil earlier about algorithms and smart bids, and we talked about how we're changing and developing our formats. Could you tell us, if you think about the type of investment that Taboola is putting for innovation for clients like eToro, what excites you? What is one innovation that you know?
Yeah. You know, for us, you know, like part of it is you need to be with a partner that can give you surgical precision in terms of what you're gonna optimize.
Mm-hmm.
For example, I think there's a venture that you've started called SmartBid Dimensions.
Yeah.
-right?
Yes. Correct.
I got it right. What it allows us to do, for example, just as a simple example, is if we wanna optimize either for CPA or for cost for advertising, essentially that spend, we're able to pivot and quickly see the results of our content in a way that is very precise. You know, marketing is what I call a spaghetti on the wall exercise traditionally, and this is the least spaghetti on the wall kind of partnership we have given the level of precision it gives us.
It hasn't become simpler. It's becoming more and more difficult.
It's very complex.
Right, to manage the content.
That's also an adventure in the sense that it makes it the better we are and the better we have partnerships like you, the more likely that we can make distance between our partner, our competitors and us.
Good. You know, we talk about media consumption and when, where people spend time, where your users, your consumers spend time, and we know that, you know, time spent on the internet, on social platform is huge. We talked about the stats earlier, but you know, we look around, we look at our kids, we look at investors, they're all on social platform. Why invest in the open web?
Yeah. You know, it's a... I think it's a fair question, because I think most of us think in terms of where the puck. You know, where is the shiny object? Social is where the shiny object is. That's where communities are being built, and the future is around communities. One cannot ignore that the open town square is still
Mm.
It's a very much important venture to be in. The web is the town square, right? It is not a precise science. When you think about it, I think something like 25% of time spent on the internet is actually on the web. This is not an insignificant, you know, part of the eyeballs on screens. Therefore, we feel that it's not something we can abandon, and it's not something we can under-invest in. What is important is to have a partner that can actually be precise in the way they do it, whether it's the level of sort of target content targeting that you do for the kind of users we want to reach is really phenomenal.
It becomes more of making sure we have a partner that knows how to navigate the web the way you do.
Okay. You talk about navigation. I think the past two years have probably been as chaotic as any two years could have been for the digital landscape. I mean.
Yeah.
We've seen regulation kick in in the States, in EMEA with GDPR and CCPA. We've seen the cookieless future emerge in front of our eyes. We've seen many different changes in programmatic buying.
Oh.
Direct buying, new technologies that are powered by AI. What keeps you up in terms of-
Well, try to be a crypto shop that also has stocks that is in the U.S.
Yeah.
Yeah. There's definitely a painting of me somewhere that's much older than I look.
Yeah.
Yeah, it's a really complex environment, and it's a really challenging space. It's actually that much harder to serve a customer like us than, let's say, a customer that is not as highly regulated, because of the fact that we have so many different nuances that we need to navigate. Again, it's that much more important to be with a partner that can help you navigate that in a way that has tools and services that have the intelligence and the algorithmic, the learning algorithmic logic that allows us to be able to adapt to that environment. I think probably financial services and healthcare is probably one of the two of the most highest regulated industries. To be able to win in this space with this kind of technology is a pretty miraculous effort.
Are you worried about the cookieless future? Is it affecting you?
I mean, I think it would be a lie to say that it's not gonna change things, right? The cookieless future is a change. It's moving the cookie, if you will.
Mm-hmm.
I think what I would say, though, is that it's more important to know that there's a reason we appreciate why that's happening, right? We are supportive of individuals being able to own their own data and being able to command their own future. I think what we need to make sure is to be partners with people who can adapt with that unpredictable future. For me, for example, the thing that's really important about Taboola is that, and I'm sure you've heard it all day through some of the videos, but, like, it's not just about third-party cookies that we source, cookies that we're reliant on when we're partnering with Taboola. It's coded into the very publishers, so it's first-party sort of sources. The quality's gonna be better in terms of targeting.
Yes, I am worried about cookieless future, but I think the more important thing is what are we doing about it, and I think that Taboola is doing something very important.
Thank you. I'm trying to think about eToro as a true global company with presence in multiple markets, and a very strong R&D foundation. You talked about what you look for in a partner in terms of scale. You talked about the precision of our capabilities. What else do you look for in a global partner like Taboola?
Yeah, you know, it's interesting because I feel like you can get wowed by technology. You could get wowed by a lot of things. At the end of the day, if the quality of the relationship is not there and you're not with like-minded people, whether it's intellectually or values-wise, I think it's really problematic. For us, of course, we look for the ability. I think you guys have said it over and over today, but quality at scale, right? That's essential. That's sweet. That's the price at the door, right? Taboola, I think, does deliver that in spades. The second is really making sure that we are partnering with people who are technology-centered at their core, you know. That doesn't mean just the rhetoric.
It means what are they investing in, and you guys are investing in hundreds of people that are invested in growing your algorithm and making it more smarter. Lastly, as I said, this is a people business. It's flesh and blood.
Hmm.
Partnership, service.
That was actually one of the inspiring things that I found in our prep call, 'cause, you know, Lule asked me, "Well, let's not talk only about ads and media. Let's talk about values." We talked about some of the things that are quite in common between the two companies in terms of a mindset and a value set. Would you like to elaborate on that a bit?
Yeah. I mean, there's so many places, I think, that investors in general. I'm in the investor services business, right? One of the things that we've learned, our customers are primarily what I call tomorrow's investors, right?
Mm-hmm.
55% of our customers are millennials, 33 are X-ers. It is all tomorrow's investors that we're servicing. What they have told us with their dollars is that it's not just about investing in the bottom line, but investing in values that are important to them. We are innovating in things that are important for us for showing how value is seen, whether it's understanding a company's footprint from its sort of its environmental impact, its social impact, or its digital impact, right? What kind of digital sort of responsibility it's taking. I think that on all three, you guys are very strong on making sure that you're creating digital impact that is responsible and not just commercially output oriented alone.
Of course, for me, like, you know, equity and inclusion is important and representation is important. It is a world where, you know, the future is browner. It's more women. Having a company that understands that we have to invest in talent that also brings those people in the room is important to us.
Yeah. Thank you for that. I mean, diversity, equity, inclusion is definitely an area of focus for us, and we're gonna hear later from Kristy as she comes up on stage, our H ead of People Operations and HR. But just seeing the amount of work that eToro puts in this space is also very inspiring. Thank you for leading the way.
Thank you.
You know, we don't have a lot of time, so I'm gonna ask a final question, which is if you were to ask Taboola one thing as we look into the bright future of our partnership, what would that thing be?
I think, don't lose your agility, your closeness to the relevance of the work you do. As companies get bigger, it can get they can start distancing from themselves from it. As wonderful as this gathering is, it's much more important to me that you perform well when we're together talking about our strategy and investing in your technology folks and your algorithms. I think just making sure that North Star is strong, I think is gonna be important.
Thank you very much. Ladies and gentlemen, Lule from eToro in the U.S. Thank you very much for joining us.
Thank you. Bye-bye.
Sorry. The most important part I forgot. I am proud to introduce Bob, my friend from Connexity, the President of Connexity, to talk about the future of e-commerce.
Thanks, Nadav. Thank you, guys, for helping out. Before we get started, I think we're gonna show a quick Connexity video that just kind of outlines the overall proposition of Connexity, and then we'll dig in further after that. Let's watch the video first. A sort of helpful view of the Connexity proposition. Before we dig in further there, I wanted to take a step back and really just talk about some of the macroeconomic trends that we're seeing related to e-commerce, and specifically how publishers on the open web are really leaning into those trends to drive growth. In our first slide here, I think the high note of this slide is really that e-commerce growth continues to accelerate at a faster pace than what analysts had previously projected. There's a couple quick quotes there from eMarketer.
eMarketer is saying that e-commerce will now surpass 20% of retail globally by the end of the year. A bigger milestone, though, is that U.S. e-commerce will actually hit $1 trillion in overall revenue by the end of this year. What's notable about that is that prior to the pandemic, the expectation for hitting that $1 trillion mark was actually gonna happen in 2024. That's shifted over, sort of two years. What really drove that? During the pandemic, what we saw was just a fundamental shift in the adoption of e-commerce online. You could say that was somewhat out of necessity, but a lot of folks were actually just exposed to the ease and convenience of online shopping, and we just don't see that shifting back. If anything, we see that accelerating further.
As I mentioned, publishers on the open web are seeing that trend. They're really leaning into that to invest in their own commerce strategies to drive growth. Some quick data points to validate that trend. 60% of publishers now cite e-commerce as their top revenue-generating opportunity. Those publishers that have actually leaned into e-commerce and are generating revenue through that source have actually gone 3x in the last year. Interestingly, LinkedIn actually stated that commerce editors are the fastest growing role in the newsroom. What that means is large publishers are just dedicating entire teams to this effort and really seeing a return on that investment. On the right-hand side, what you'll see is those are all examples of publishers that are really leaning into e-commerce. Probably notable that they're all working with Connexity.
All of those are building commerce content, operations at scale using our tools, monetization, reporting, API. Just some quick highlights there. The Independent grew their e-commerce revenue by 60% in 2021. Condé Nast now expects that their e-commerce revenue will be a third of their business by 2025. When you really think about you know Condé Nast, that's a pretty big shift for them. They have invested heavily in their offline publications for you know the majority of their history. I think the bigger takeaway though if you really think about this slide, Connexity is sitting at the nexus of that investment and helping premium publishers just create and monetize their commerce-oriented content. That's really the crux of what we're doing for our publishers.
If we go to the next slide, this sort of gives a snapshot of both the advertisers and the publishers that we work with. Much like Taboola, we operate a two-sided marketplace. The difference really is that Connexity is solely focused on the retail sector, so all advertisers are e-commerce online advertisers. On the left, it's the advertiser base. On the right, it's your publisher base. You can pretty quickly see a few things, though, out of this. The primary one is that these are all blue-chip partnerships. On the left-hand side, these are the top 100 retailers in the U.S. On the publisher side, it's also the biggest and the best publishers. We partner with 15 out of the top 20 global publishers worldwide. Probably worth noting as well the range of retail categories that we operate in with retailers.
It pretty much works, our solution works across every category. We have home and garden, electronics, automotive, sort of fashion, apparel and fashion in there. Everything sort of works within that. There are different types of retailers as well. When you think about it, there's omni-channel retailers, there's DTC brands in there, and then there's online only. Similar dynamic on the publisher side, where you have some publishers that are online only, think of the BuzzFeeds of the world, and then potentially those that are making much more of a transition to online, so think Hearst and Condé Nast. On the right-hand side, maybe just to sort of cover off on some helpful stats which kind of orient you to the history of Connexity.
Connexity has over 20+ years of experience in the retail space. A lot of institutional and industry knowledge about the complexities of how to run retail still within the management team. Would call out that we were one of the pioneers in sort of retail lead generation back in probably 2001, 2002. The other thing that's really interesting about our network is we call it sort of, you know, the long tenure, low churn rate network. The average tenure of our advertisers is 10+ years. If you actually look at some of the larger advertisers like Walmart, Macy's, and Lands' End, those are 15+ years. When we jump to this next slide, this gets into a little bit more about how the entire ecosystem works.
There's a lot on this slide. I'm gonna try to walk you through this. There are sort of four pillars in the slide, and each one has sort of a representation of how it provides value through the ecosystem. On the left-hand side, at a really high level, the primary goal that we're trying to achieve for all of our retailers is find them net new customers and incremental sales that they wouldn't otherwise get without being part of the Connexity network. The second is when you look at the scale on that left first column, we operate at a scale that really matters to retailers. We have 150 million monthly shoppers, $4 billion in GMS on an annualized basis.
For most retailers, what they tell us is that we're their single largest source of traffic outside of their direct campaigns in Google, Bing, and Facebook. A very meaningful part of their overall portfolio. The second column, where it says unique channels, essentially this represents where we get access to consumers. These are all the channels that we're getting access to consumers and driving them to retailers. Very much like Taboola, Connexity operates on the open web. We have direct partnerships with premium publishers like the Hearst and the Condé Nast, and essentially that's giving us access to those audiences. Unlike Taboola, though, we actually access audiences within the walled garden, within traditional paid search channels of Google and Bing specifically.
Connexity has a really long history of building its own proprietary technology to be able to build operations around buying traffic upstream. We've done that for years. The last channel here, which I think is important, is the influencer marketing channel. Obviously, a lot of talk around that type of social engagement around retail. About two years ago, we actually built our own influencer marketing platform, our own brand called ShopYourLikes. We now have over 1,500 direct influencer relationships on that platform. They have all cultivated really large audiences that are engaged in posting retail-oriented content in all of the major social platforms. Think YouTube, Twitter, Meta, and Instagram. I would call out, though, that these are all what we call performance-based partnerships.
In a lot of cases, what you hear about influencers, they're paid on likes or impressions, vanity metrics. That's not what our relationship is with our influencers. They actually have to deliver sales that meet our ROI targets of our retailers. I think when you look at just sort of staying on that second column, just to maybe give you some relative perspective in terms of the overall contribution. If you use ex-TAC revenue as the benchmark, 85% of our revenue actually comes from publishers on the open web and those influencers. 15% then comes from the rest of it, so it's basically within the walled gardens, so those paid search channels. The third column, which is using proprietary technology, this is really how we accomplish all of this.
We've facilitated all of this through our own proprietary technology. It's been built over the last 20 years. Obviously can't go through the entire stack here, but there are three sort of systems that I think are worth sort of highlighting. These are some of the foundational systems that really build our relationships with the retailers. The first is our inventory management platform. We ingest over 30,000 feeds on a daily basis. Those feeds house over 750 million offers. We update those offers multiple times per day to be able to get the latest in-stock and pricing information. We also enrich those offers. A lot of times what you see with retailers is they don't provide categorization, they don't provide attribute data, and so we have technology that fills that in.
We also have a way to provide tracking links across each one of those offers, 750 million of them. Probably more importantly, we add what we call estimated eCPCs to every offer traffic source combination. What that allows us to do is to signal to our downstream partners what you're going to potentially get paid if you're using the content from our feeds. That's an engine that sort of reproduces the estimated pricing on a daily basis based on all of the historical conversion data that we have. The second system there, in that third column, is our smart pricing system.
This has been developed over the last 12 years, really helps dimensionalize all of the clicks within our network and make sure that we're pricing each click based on what it's worth, again, using all of that sort of conversion data at every level of granularity within our ecosystem. I call this really the engine that drives performance. This is what really makes sure that we have our ROI targets met by every retailer that's participating in the Connexity network. The last sort of system here is our custom reporting and analytics system. The one thing that I would highlight is we have a lot of retail data that we sit on. We know what products consumers are clicking on. We know what consumers are transacting on.
More importantly, we're able to sort of package up all that data and be able to feed that back into our smart pricing system to be able to make better decisions on pricing. The other thing that is really notable is we have really strong partnerships with our publishers, and so we package up all that data as insights and information that help really drive and guide their e-commerce strategy. If we can tell them, these are the categories, the brands, you know, some of the products that are really working well, they'll start to lean into building more content around that, and that has been super successful in terms of their engagement with us. The last sort of column, this is really the ultimate goal for everything that we do for retailers. We have to deliver on performance.
We have a history and a track record of always hitting our retailers' ROI goals because of that other technology. We wouldn't have the long tenure and relationships that we do without having, that particular sort of focus. We have sort of an obsession around making sure that we're hitting the ROI goals there. If you think about the next slide, I'm gonna sort of walk you through a variety of different examples. There's multiple ways that we create touch points within our publisher base. The first one here, this is a lifestyle publisher. You probably all have heard of Hearst. They're basically starting to employ a commerce content strategy, across most of their brands. There's three that are highlighted here, so Good Housekeeping, Women's Health, and Bicycling.
Essentially, those are all sort of offline magazines which now have significant online presences. You can see in the pulled out links there, so there's three various links, whether it's the Shop Now button, the text link, or the price comparison widget. In all those cases, we're really helping the publisher earn money by driving links to our retailers, through those particular integrations, and that's really how they actually make money. Again, I would call out in this slide because it's really important with our editorial publishers, is that it's not just about monetization. It is really about the insights and the, guidance that we provide around what's working for their consumers. We know, again, products, brands, categories specifically that are converting for their audiences, and so we feed that back to them in terms of those insights.
We also have a ton of information around sale and promotional calendars, so all of our retailers send us their sale and promotional calendars. We synthesize that data, feed that back into our publishers, so they're writing about events that have higher conversion. The last thing I would say is, we have sort of the seasonal data that, you know, over the last 20 years, you have a really good understanding of what seasonal periods really drive conversion, and so you're really trying to funnel that back into the publishers so that they can make better decisions there. The next integration, this is a very different integration relative to the editorial example before. This is Yahoo Shopping. Most of the Yahoo Shopping portal was actually built off of our underlying feed data.
Essentially think of this as a full marketplace experience. Consumers can come in, search for any product, identify all the merchants that are selling that product, do some comparison shopping, make a more informed decision, and ultimately the conversion rates are higher there. What feeds this is all of the underlying product metadata that comes out of our feeds is really what's driving this entire experience. The final example, I mentioned our ShopYourLikes brand. We built our own influencer marketing platform. This is an example from a post of an influencer who's part of that ShopYourLikes partnership. She's been with us for a year, has over 100,000 followers on Instagram. In this example, she's basically highlighting a dress that was on sale from Lulus. That dress converted really well.
She put that in her Instagram Stories. The swipe up feature in Instagram Stories allows that lead to be generated to the retailer. She gets paid on those swipes, but more importantly, on the performance of those swipes, so how they convert over time. I think Adam and Steve have done a much better job of providing the overall comprehensive view of the Taboola and Connexity synergies. What I wanted to do is sort of share a quick update on a subset of those synergies that seem to be showing a lot of promise and are already getting traction in the first couple months here. First one is expansion of the Connexity publisher solution into APAC and EMEA.
I would call out that our core product was really well known in the U.S. and the EU, realistically not well known outside of that. We see a lot of opportunity to be able to take that core product and sell it globally. Taboola has amassed, you know, a large global footprint, which you guys have talked about, and more importantly, existing teams that are on the ground. The plan is to utilize those teams to be able to expand into 14 new countries. The teams are already trained up and are pitching actively two to five new publishers on a weekly basis to be able to fulfill on that synergy. The second big synergy here is leveraging the ad sales team, Taboola ad sales team to sell the Connexity advertiser solution.
Would again highlight just some context about Connexity. Our core DNA was really built around operations and analytics. We invested very little in new business. For context, we had one new business person in the U.S. in the last year in late stages of a PE environment. We see a ton of opportunity here to be able to utilize the you know the world-class enterprise sales team that Taboola has built. When you look at our team versus their team, there's just no comparison. There's just a ton of you know opportunity for us to lean into that team and be able to leverage that group to be able to sell the core Connexity product. To that end, we've actually created a retail vertical sales team. That team has been trained up.
The idea is for that team to actually go to market and start to pitch the core Connexity advertiser solution in Q2. The other thing probably just to call out on that slide, really quickly is that, we actually saw an opportunity with the China sales team. They have a large number of Chinese retailers that are trying to get access to the U.S. market, and so we've actually worked with that Chinese sales team to train them up. They've actually gone out and signed three new retailers in the first couple weeks that they've gone to market, more are being onboarded as we speak. Could be a decent size opportunity as well. The last one is really leveraging the Taboola supply network for our Connexity advertisers. Mentioned earlier, we have really strong relationships with our advertisers.
I think Eldad also mentioned this. This is similar to the Taboola network. We have the majority of our retailers on what we call uncapped budgets, so the ability for them to continue to sort of spend with us without us asking for more dollars. As long as we're hitting their ROI goals, that's a big advantage in terms of being able to continue to diversify that spend into different supply networks. We're working on doing that with our existing retailers. We're trying to get that into the Taboola supply network. We've done a lot of training in the first few weeks with our U.S. account management team. There are three weeks of pitching so far. 29 retailers have actually gone on and agreed to move forward with that.
What I mean by agreed to is there's a small amount of pixeling and tracking work that they need to do. Everybody has agreed to sign on to that. That's probably one of the more promising starts that I've seen. Many people have been, you know, hearing about the acquisition, but also wanting to take advantage of it. Now that's in market, and people are really responding to that. Maybe this is fair. I know I've been going on for a little while, and I'm probably a little bit over time, but I wanted to share maybe just a little bit of a personal reflection. I've been at Connexity for about 22 years. I've seen e-commerce go through a variety of different transitions. I believe at this point we're actually going through a more fundamental transformational shift in e-commerce.
The reason I believe that is because we have our retailers that are actually sharing a lot of information with us. I'll give you a quick anecdotal story. I went to a QBR, quarterly business review from one of our retailers, one of the larger retailers. They shared their forward-looking outlook for 2022, and one of the biggest themes in that outlook was growth outside of Google. What does that mean? What they're seeing is that historically they could plunk down a good percentage of their budget in Google and Bing, and they would see a consistent double-digit return year-over-year. That's not happening anymore.
What they're telling us is, "Hey, we've got, you know, slowing growth there or sometimes even negative growth, and we need to figure out a way to move away from our sole dependency on Google." In a lot of cases, retailers have a very, very heavy dependence on Google, and that's scary to them. What they're trying to do is figure out ways to buy outside of Google, and that means buy in the walled gardens. When you couple that with what's happening on the publisher side that we've talked about, publishers also need to diversify their revenue. I think their core DNA is around building really extremely compelling content. Where they're pointing that to is now towards e-commerce content.
When you look at, you know, the ability for us to, you know, showcase more compelling content on the open web versus what's happening in Google, it's not that people won't go to Google. It is that we believe that there will be many more opportunities to reach consumers on the open web with e-commerce journeys, and we will be a big part of that. When I think about the marriage between Connexity, which has really strong advertiser relationships, Taboola's global supply footprint, their world-class sales organization, and more importantly, their leading-edge technology, I believe that Taboola is really, at sort of the forefront of leading that transition into e-commerce on the open web. Excited to be a part of that mission, excited to be part of the larger Taboola family.
It's probably always better to hear from a client and not from somebody like myself. We're gonna close out with a video from Amy Larson. She's Senior Director of Growth Marketing at TOMS, and she's gonna sort of tell her story about Connexity.
We're here today with Amy Larson, Senior Director of Growth Marketing for TOMS, a shoe brand known around the world for both its philanthropic goals as well as its focus on sustainability.
I'm responsible for all growth marketing. What that really means is I'm responsible for generating top-line revenue for toms.com across paid, earned, and owned channel. Connexity has proven to be effective across multiple brands that I've worked with. I've found that Connexity really kind of augments our Google, and Bing, and Facebook more traditional media offering, and really helps us to grow the business and find new customers.
We've been really happy with Connexity, and that's really kind of diversifying our paid channel base and finding more ways to prospect and find new customers in the digital landscape. Connexity has been another great source of finding those new customers. I think every organization has slightly different KPIs and slightly different ways of measuring success for paid media channels. I mean, there's always the typical goals of generating more return on ad spend and driving lower cost per order and things like that. I think my relationship with Connexity has evolved over time in a way that I trust them, because I can go to Connexity and I know that they're gonna deliver for me over and over again. That's been really helpful.
I think Connexity can really find those customers that are interested in trying new brands and interested in engaging with us around our promotions and, you know, a lot of, just especially in footwear and apparel and so forth, Connexity can really find that sweet spot of customers who really want to engage with our brands. We wanna continue to grow the TOMS brand and the toms.com business, and I think Connexity is another great avenue for helping us do that. They can find those customers out there who really have an affinity with our type of brand, affordable footwear, that's really kind of, fashion-forward and, you know, help us to really grow our business overall.
Excellent. Well, Amy, I really do appreciate you sitting with us. Thank you very, very much for the interview.
Thank you.
Here to talk about Taboola News next.
Hi. Can you hear me?
Not yet.
Not yet? A second. Hello, hello.
Yes.
Hi, good morning, everyone. My name is Jonny Harvey, and I'm the Managing Director of Taboola News. Just a little bit of background, first of all. Before I joined Taboola, I was working at Microsoft in the Entertainment and Devices division. For the last 7.5 years, I've been working at Taboola. For the first six years of that time, I was the VP in APAC. I was based in Southeast Asia, which really is the heartland for all the OEM partners that we have today. For the last 1.5 years, I've been leading the Taboola News business alongside our head of product, Hanan Stein, who's also here. He's at the back of the room, who you'll also get to meet later on. We're here to introduce Taboola News.
Before I get started, I'm gonna run a short video and get into more details. Firstly, I just wanna say that we are really immensely proud of the partnerships that we've established now for the last four years. We work with the largest handset manufacturers all around the world, and I'm really excited to share for the very first time actually, that we now reach 100 million monthly active users all around the world across the partnerships that we have today. Now, if you just click on to the next slide. Just to talk a little bit more about Taboola News. I mentioned we see millions of people all around the world every single day, and they're coming to Taboola News to discover what's new and what's happening.
Whether that's from the lock screen of their device, so the first thing they see when they wake up their phone is content recommendations that's personalized for them. That's even before they get into their social apps, or they're getting into their browsers to read content from articles. Through to swiping from the home screen to the right, where they're going to their news feeds to learn about what's breaking, what's new, top news. Then also within the browser of the device itself. Once they're searching, and they get into those moments of discovery where they're ready to learn something new and interesting. We're also powering notifications when important events around the world are happening to keep users informed.
We're really excited about the partnerships that we've delivered with all of our OEMs. Now, a really big opportunity for me today, as well as to talk about Taboola News, is to introduce you all to one of our partners from Samsung. We have Avner Ronen with us here today. Now Avner is the VP of Product Development at Samsung Electronics. As well as that, he's a serial entrepreneur. He's the CEO and Founder of Boxee, which was a streaming technology, which was also acquired by Samsung. He's here to talk to you about his vision on Taboola News and Samsung. I would like to invite Avner to the stage.
Thank you for the kind introduction and great job on the video, by the way. So, you know, I work at Samsung on the product side, not on the hardware, but on the services group, which for Samsung is kind of a new thing, actually. You know, we're the number 1 TV manufacturer. You know, close to 1 billion phones, 150 million TVs. There's a lot of users touching our devices, not to mention washing their dishes and clothes in our machines as well. You know, we're mostly a hardware company. That's what we're known for. Over the past, I would say five years, the company has made significant investment in offering services to our users. You know, it's still a learning process for us, and it's a significantly growing business, but we're making inroads.
We're doing it both on our TVs and on our phones. On our TVs, if you use our TVs at home today, and you open the TV, you may see TV Plus, which is our streaming service, where news services are doing very well. Or you may see our art store if you have the frame. If you open the mobile phone, we have a Galaxy Store, and we have our Minus One screen where you can see content from Taboola. We have a browser, we have Samsung Pay, we have a variety of services. This is a part of the business that really is growing really nicely. The reality is people are not, you know, replacing their devices as often as they used to.
The relationship that we have as a manufacturer with the user, it's very important for us that we're not just the hardware that they use, but that they also use our services. On that front, you know, we've been working with Taboola now for a while. I think you've seen examples of Taboola is integrated into our browsers and powers our new service on the browser and into our Minus One. When you swipe, which people do very often, you can see news from Taboola, and we're looking forward to integrate into other parts of the user experience. For us, before I get into the product, the relationship with Taboola is quite amazing. You know, we work with many partners. I know it's a Taboola day event, and, you know, I didn't even get any, like, goodies yet.
I'm doing it without knowing if there's any reward here. You know, the relationship with Taboola has been quite amazing. Extremely responsive. The team both here in the U.S. and in Israel work with us both on the business side and the product side. They're proactive. They're extremely quick to respond if we have an issue. Extremely flexible, working with us as a real partner. If we have good ideas, they'll work with us. If we have bad ideas, they'll tell us it's a bad idea, but still work with us. You know, they're truly a partner rather than a vendor. We share with them a roadmap. We ask them questions. Do you think this is the right thing to do or not? How do we solve this problem? We've never heard like a no.
It's always a yes, always helpful, and this is the experience that we have. I'm based here in the U.S., but also the experience that our team in Korea has. You know, Samsung, like any big company, is a complicated company to work with. Taboola has to work with a few different people, getting sometimes, you know, different answers to the same questions. They manage it and everybody loves them. That's not true for most of our partners. First I want to say thank you to Taboola for the relationship and the investment and being such good friends. I know, you know, that's been mentioned by eToro as well, so maybe we're not unique, even though they make us feel very special. They really are a good partner.
You know, this is a relationship that hopefully will last for years. You know, like, the CEO of eToro U.S. mentioned, you wanna work with people that you feel comfortable with, that you trust, you know, ups and downs, frustrations. You wanna feel comfortable. With Taboola, we believe we have that in a big way, and it's appreciated throughout the company. Want to put that out there first. From a product experience, we're excited about the opportunity with news. This is a service we currently offer, and we intend to invest more in and make the experience better for users. You know, this is a service today available on our phones throughout the different models that we have. There's different touch points. We wanna improve on those touch points as well.
I think you mentioned notifications and, you know, the Minus One screen and the browser, but we think there's many other touch points that are relevant. We need to figure out the balance between, you know, the user experience and not being too pushy. At the end, you know, we need to win the user's trust when it comes to news, and people are not shy about telling us if we break their trust or they think we're leaning too much to the left or too much to the right or too much to the center, we can't do it right, I think, whatever we do. Taboola is helping us figure it out.
You know, the thing about news, you know, people are opening our devices, our phones, they're picking it up over 100 times a day. Most of those times, as you know, as users, you don't know why you're picking up the device. You're bored, you're at the restrooms, you're at the, you know, somewhere waiting in line somewhere. You just pick up the phone and you're, you know, wasting time in most cases. People scroll through their social feeds, and they play, you know, games, but also they get informed. For us to be there when people are just looking at their screen or opening it and be there with something that's relevant, we think is extremely important and it's a big opportunity for us.
We wanna do it in a way that is not intrusive. We're not trying to build a humongous business of just showing people ads all the time. We first and foremost wanna show people value, and have it be part of their menu. It's fine to scroll through the social feeds. It's fine to play mindless games. It's also fine to get educated about news. This year, I didn't watch the Oscars 'cause I said nothing interesting is gonna happen. I got the notification on Will Smith. So that's also news, you know, maybe more entertainment news, but that's fine too. It's not. Everything doesn't have to be about like the stock market or wars. Definitely for our younger users, they're more interested in, you know, entertainment type content.
The nice thing about working with Taboola is we can personalize the experience for the user. If you're really into sports, you'll get more sports. More into politics, you'll get more politics. I guess more into MMA fighting, you'll get Will Smith notification. It really depends on your interest. You know, we've been working with Taboola for a while, and what we see is, you know, they're constantly improving and working with us to make it better. We have more stuff in the works together with Taboola. We hope that news is gonna become something more meaningful as part of the services that we offer the users. You know, when we look at services that can be relevant to a good swath of our users, news is definitely right there at the top because almost everybody consumes it.
Again, maybe a different type of flavor of news, but we think this is a mass market service, so we hope that most of our users would develop a relationship with Samsung that they trust and they get news from, you know, hopefully on a daily basis. Yeah. You know, working with Taboola has been an easy, relatively easy integration to get started. Now we're going deeper. We hope that the ideas that we have, the ideas that Taboola have, is something we'll be able to bring out to the market later this year, next year, and again, a long-term relationship. We're excited about the partnership, excited about news, yeah, excited about the Oscars next year. I don't know.
I'm doing well on time, I guess, but happy to engage with questions, even though Adam told me you don't like, you know, asking questions. If you don't ask, I'm gonna pick at random.
I think we can definitely do some questions later on, right? I think.
No, I want questions now. Why later on? All right.
What else did you look at in terms of identifying other vendors? What else in the space that could possibly create an option for you guys to-
I mean, we could do direct relationships with publishers, but you know, we're a global company, you know, over 100 countries. To do all those direct relationships would have been a real challenge for us. So that probably would have been an option. Then, you know, we could work with somebody that's already aggregating news, like a brand, and do you know, either a white label or just a partnership, which we've done in the past. But then it's not really a Samsung service. We can't really personalize it, and we can't drive the roadmap.
We wanted to do something on our own, so we were committed to investing in it, but we needed a partner that would help us work with the different publishers, have experience with personalizing news, and that we thought is something that's gonna be very hard for us to start from scratch. Taboola was, you know, an obvious choice for us. Yep.
What does success look like?
I think success for us would mean that we want the, you know, significant number of users' trust, that they can rely on us for news, that they, you know, if we send them a notification, it's in the right time and for the right type of content, and they open it, and they come to rely on it as a major news source. Not the only news source. I still open The New York Times every day personally, and The Wall Street Journal. Then when I get notification, I'll open it, and hopefully Samsung could be there for some of our users some of the time. For us, that's success, that we become part of the menu for users' consumption of news, not the exclusive source.
Business-wise, you know, obviously, if we get a lot of usage, business will follow, but it's first and foremost, drive value for the users and win their trust and engagement. Yeah, so we're global with news. Not in every country, but tens of countries, North America, South America, Europe, several countries in Asia. We see some differences between countries. Mostly, honestly, it depends on how good is the list of publishers that we have in that country. Do we have the major publishers or not? But then the usage is actually pretty similar across the different countries. We haven't seen a huge difference in geography. All right.
Okay.
All right. Thank you for having me. I hear there's lunch coming. Yeah, looking forward to continue working with Taboola and maybe see you in Investor Day 2023, 2024, 2025. Make it a tradition.
Thanks.
Thank you.
Thank you, Avner. Thank you very much.
Thanks.
Thanks. Thanks ever so much to Avner for spending time with us today. I mean, today, Samsung, we've been working together for a number of years. We're live in over 80 countries all around the world, and obviously it's a very strategic and special partnership for us. I think it's pretty clear from what we've seen today that our focus right now and today has very much been around the OEM and the smartphone. We have many opportunities. I think our vision over time is really to be everywhere with Taboola News, whether that's your smartwatch, whether that's the autonomous cars, or anywhere really where a screen is connected to the open web. The future possibilities really are endless, and we're very excited about the opportunities. Again, thank you very much, Avner. Thank you for coming and joining us.
I believe so. We can talk about that later.
I'd like to introduce everybody now to hand over to Kristy, who is gonna talk to us about people at Taboola. Thank you very much.
Thank you, Jonny. I think you can hear me right. I just can't say how proud I was sitting at the front seat. I was always a good student. I went to engineering school and aced it all the way, and I sit at the front seat watching all the homework that got delivered, and I was so proud of it. I know Taboolars have been watching it globally. I got text, WeChat, WhatsApp telling me, "Wow, this is awesome." This really shows all of you, hopefully investors, that this is the work that we do behind the scene. This is what almost 2,000 people globally doing days and night, really working with our partners, our publishers, our advertisers in delivering our success. I'm Kristy Sundjaja. I lead people strategy and operations at Taboola.
I joined the company slightly more than two years ago, and if people ask me why did I join Taboola, I'll tell you, simple, it's the Taboolars. We call ourselves Taboolars, just add an R to our name. Until today, Taboolars are really the reason why I wake up every day coming back to work. We are a bunch of very fun people. You should see some of the chat. It was hilarious. We are very creative, we're very innovative, and we're very agile. You heard it from our partners, from our customers, in terms of how dedicated we are globally in solving our customers' problem.
That's truly drives me every day to come back to work and say, "What can we do to attract more diverse talent to come to Taboola and make sure that each of us find our path to succeed and really enjoy our work at Taboola?" Today, as you listen to the presenters, it give you a snapshot of who we are. Some of us has been with the company since the beginning of the day. Some of us just joined recently through an acquisition. Some of us has spent decades in one function mastering the art, becoming the world renowned, like, artist in that art that we do. Like Gil, who is, like, an expert in deep learning. There are people who are really adventurous.
I mean, Jonny talk about himself, like, traveling over the oceans, get into different functions, really trying to be this global entrepreneur leaders working with customers globally. That's who we are, and that's our people strategy. We want to be the place where the best people can discover their potential, enjoy their work, and succeed on a daily basis. We have big goal. We need to grow. Every day, every week, we are talking to hundreds of candidates trying to apply for a job at Taboola globally. We constantly try to raise the bar. We're not happy with where we are today. We wanna be happier. We want our customers, our publishers, advertisers to be happier. That's why we're bringing in diverse talent. We realize that diverse talent bring in different perspectives.
Sometimes we fight, sometimes we argue, but at the end of the day, when I leave the room, I know we have made a better decisions because of the diverse talent in the room. When I was asked to tell you more about people, I chose one specific topic that I wanna show you a video, is our Diversity, Equity, Inclusion efforts. When I first met with Adam and also our leadership team, I asked them, "Are you serious about this? Are you serious about investing in diversity?" They say, "Yes." To me, it's such a joy to be in this position to find partners who realize the importance of it. Diversity is important for our business, and when you bring in diverse talent, we want them to feel included. We want them to have a voice. We want them to be successful.
I'm not gonna show you hundreds of slides. I'm gonna show you a video, and you're gonna hear it from our Taboolars in terms of what we are doing in the DEI space.
In 2022, our theme and initiatives will highlight how we will work to come together to rise as one.
Each and everyone's unique perspective must feel valued in order to cultivate a sense of belonging and to create a workplace where everyone can be their authentic self.
It took me about three years. Today, I wish I came out earlier, but I guess I needed that time.
It was my first serious boyfriend, now her husband, about 20 years ago.
Empower her by showing her how much you appreciate her.
Empower her by providing equal opportunities to pursue her dreams.
Strive towards creating an inclusive environment for women, ensuring that their efforts are appreciated and recognized, making Taboola a company where women just don't want to grow, but rather choose to grow. Being a part of our employee resource group, TaboolaHER, gave me and our allies an opportunity to move from well-meaning to well-doing, to help drive a lasting change that creates an equitable future for all.
As we work towards cultivating a culture of belonging, we were empowered and given the opportunity and support to create TaboolaInCulture, an employee resource group. TaboolaInCulture warmly welcomes all Taboolars, whether a member or aspiring ally.
Everyone is talking about D&I and inclusive environments, but actions speak louder than words. D&I is everyone's responsibility and doesn't fall on the heads of the ERGs and diversity managers. Being an ally is just as important.
As a member of the Talent Acquisition team, I'm excited to be a part of the changes that will impact the diversity in recruitment. Diversity of perspectives is a competitive advantage as we continue to innovate, scale, and definitely make an impact. In 2022, we will have four key focus areas. Number one, training and awareness. Through training and shared experiences, we will build awareness and work to develop one common language. Number two, diversity and recruitment. In support of our DE&I efforts and diversifying our talent, we will set measurable goals to ensure that we include qualified, diverse applicants from the earliest stages of our recruitment process and hiring pipeline. We will also partner with external vendors with access to highly skilled underrepresented groups who may not currently be in our recruiting platform. Number three, professional development programs for diverse talent.
Peer engagement and professional development are critical to the success of one's career. These programs will increase exposure for participants and help develop strategies and skills that will accelerate career development. Number four, community engagement. We will look to partner with organizations through volunteer and philanthropic efforts. All of these initiatives will highlight the importance of awareness and reflection with the understanding that each of us has a role to play in driving change.
We all have unique life experiences that make each and every one of us different and important to our organization. As a leader of the open web, with a global reach and influence, our work has the ability to impact the way the world sees people and how people see themselves.
Together, we'll continue to highlight and build awareness around the importance of allyship.
Together, we'll work on bridging the gap in representation.
Together, we'll create a culture of belonging.
Together, we'll work on unlearning old behaviors.
Together, we'll rise.
Together, we are Taboolars, and I know when Taboola set a goal to be successful in an area, we'll not only nail it, we're gonna exceed it. In the upcoming investor days, you'll see us, how we become a more diverse, more inclusive, more equitable company as we deliver more results to our partners, more impacts in the open web. With that, someone is gonna tell you about our performance and results, and that's our CFO, Steve Walker.
Thanks, Kristy. Thanks, everybody, for coming. Before I jump in, I think most of you know who I am. I'm Steve Walker, I'm CFO. Been with the company over seven years now. Came in through an acquisition of a company called Perfect Market that I had helped found, and I've been in the tech space for a number of years before that. I also wanna just thank our visitors today, so Kristin from McClatchy, Avner from Samsung, and Lule from eToro. That's kinda special. Special to have your partners here kind of repping on your behalf, so it was nice to have them here.
I'd also like to point out that Adam said some very nice things about how I was gonna tie everything together at the end, but I'd also like to point out that he didn't give me a snazzy video to lead off, and he put me right before lunch. Take it all with a grain of salt. I'm gonna dive into our financials here, are basically three parts. First, I'm going to give a quick review of our financial performance historically. Second, I'm gonna talk a little bit about what differentiates us from our competitors and tie in a lot of the things that you heard about today.
Third, I'm gonna talk about how we grow going forward. First, you've probably, many of you have heard me talk about this, but when we think about our business, we're really looking to grow but grow in a profitable way. The way we measure that is looking at two metrics. For growth, we look at how fast can we grow ex-TAC revenue. Ex-TAC is the revenue that we keep after what we pay to our publishers. Then on the profitability side, we look at adjusted EBITDA margins, so adjusted EBITDA divided by our ex-TAC. The model that we have, the long-term model that we've set, the goals for ourselves is we wanna grow, we wanna grow ex-TAC at 20%+, and we wanna maintain 30%+ adjusted EBITDA margins.
If you look at our history, we've done a very good job of that. We've grown 25.5% on an ex-TAC basis over the last five years, and we're currently above our long-term goal of 30%+ adjusted EBITDA margins. As Adam talked about earlier, if you kind of extend this out, we should be able to achieve more than $1 billion of ex-TAC by 2025. That's the kind of forward-looking part of that. Looking back at 2021, when we originally put out our PIPE deck, when we were raising money as part of our going public process, we had said that we wanted to do $445 million of ex-TAC in 2021. That would have represented 17.4% growth over 2020.
We actually did $519 million of ex-TAC. That was 36% growth. Obviously, we did the Connexity acquisition, so we had four months of Connexity in there, but it still would've been 25% growth on a pro forma basis. We're still exceeding our targets. Adjusted EBITDA, you know, we had said we wanted to have a 20%-20.5% adjusted EBITDA margin. We actually did a 34.6% adjusted EBITDA margin. Just wanted to kinda highlight that we're beating our numbers here out of the gate. If you look forward to 2022, in our pipe deck, we had said we wanted to do $516 million of ex-TAC in 2022.
We expect to do $665, obviously again aided by Connexity, but frankly, again, we would be well ahead of that even without Connexity. Again, our margins are above our long-term goals of 30%-30.7%. So far so good. We're beating our goals. One of the big reasons we've been able to beat our goals both this past, you know, last year and into 2022 and even before that is that you can see that over time, we've had a very steady increase in our ex-TAC margin. Our ex-TAC margin is the percentage of the gross revenue that we keep. The reason it's important to understand this is that is a pretty strong indicator of competitive advantage.
Obviously, we have to compete in a competitive environment for winning publisher deals with great publishers like McClatchy. Part of that is how much you can pay them, part of it's how much value add you bring on top of that, but it's a competitive environment. If you're able to improve your margins consistently over time, it's a pretty good indicator that you've got something that is a advantage versus your competitors. We believe this is a pretty strong indicator of competitive advantage. I'll also note that the competitive landscape has not really changed that much over this time. We face similar competitors that we have. We've grown faster than them, so we're bigger than them now. In fact, recently, we've actually brought on more competitors in many ways.
As we go into High Impact Placements and we start expanding what we're doing, we're taking on more and more competitors. We live in a very competitive environment, and yet we're able to continuously improve margin. How is that? The next thing I wanna talk to you about is how we've been able to achieve that. There's basically six reasons. I'm not gonna read you, the reasons from this slide, but I'll walk you through them one by one. The first reason, and we've shared this with many of our investors as part of our fundraising process, but the first reason is that there is a network effect in our business. Bigger is better in our business. The reason for that is basically what we're trying to express with this chart. It's the network effect.
The way to think about this is, as we win more publishers, we see consumers more frequently. We see people reading articles, clicking on things. We get more data, and that additional data allows us to do a better job of basically targeting ads to those people, targeting content too for that matter, but targeting ads to those users. What that does is it makes. If the targeting is better, it makes our advertisers more successful. We win more advertisers, we drive higher yield, and that higher yield then makes it easier to win the next publisher more profitably. The cycle begins again. You get bigger again, you get more data, and it just reinforces itself. This network effect in our business is one of the big things that over time has helped us improve our margins.
Even before we were the biggest, Adam likes to say we weren't first, but we're now the biggest in our space. Even before we were the biggest, we had an advantage. That advantage is that we invest heavily in technology that helps drive performance for us, yield. We talked about the fact that there's three components to yield: click-through rate, cost per click, and conversion rate. All three matters. You heard Gil speak today about how much we've invested in and really gone all in on deep learning and our AI, which helps drive this performance. 90% of our advertisers on a dollar basis use SmartBid. That's our AI that drives the performance that they're seeking for their ad dollars. That's a unique advantage that we have.
We're one of the few companies out there who has this much investment in AI and can drive that type of performance. Even before we were the biggest, we were investing in this type of technology that helps target ads better, and that's also helped us differentiate ourselves from our competitors. You'll notice if you look at our R&D investment versus our competitors, it's much bigger. The third differentiator is around privacy. We've talked about this extensively, but many companies in our space face a headwind from privacy. As cookies go away, it becomes harder for them to target ads. It costs them revenue over time. For us, we actually think it's a tailwind. We actually think over time, it's gonna give us an advantage because of the fact that we're heavily contextually oriented.
The proof is basically in the chart in the top right, I guess, from your perspective. What this shows is the orange line is our yield, how much revenue we make per ad spot for Safari, and the blue line is mobile Chrome browser. What it shows is that since Apple first blocked cookies back a few years ago, we've actually grown the yield on Safari faster than we've grown the yield on Chrome. Both have gone up, but Safari's gone up faster. We think the reason for that, and it's tough to prove perfectly, but we believe the reason is that advertisers who wanted to reach iPhone users had to find a way to reach those users in an effective way, so they found us because we use contextual data, we don't use cookies.
All those other channels that were working before stopped working, and they came to us, and that's helped us drive yields much faster on Safari. We think that the privacy landscape is actually a tailwind for us and is something that will help us over time. The fourth differentiator is related to what Eldad spoke about earlier. By the way, I'll just tell the joke. We thought these weren't busy enough, so I decided to put two on one slide and make them much smaller. But basically, what I'm trying to emphasize here is that Eldad talked about how much we invest in our platforms. This technology investment is a big differentiator for us.
Eldad mentioned on the publisher side, that small Taboola ads box, that's an SSP. That's their whole business. We invest in all sorts of things around that to have a much more rich ecosystem, more tools that we can give for publishers. You heard Kristin talk about Homepage For You and Newsroom and how it's gonna change the way they do business as a publisher. Those tools are unique to us. Other people don't provide that. You heard Avner talk about Taboola News and how we're working with Samsung to drive news onto devices, which then turns around and drives traffic to our publishers. That's unique to us. SSPs, our direct competitors like Outbrain, they don't do these things.
This tech stack is very unique to us and gives us a huge advantage when it goes to win publisher deals. Likewise, on the advertising side, we spent a lot of time today talking about our AI and how much we invested in the technology that targets our ads. Even beyond that, there's very few companies out there that have all the brand safety tools that Nadav mentioned earlier. There's very few companies that are integrated with all the different channels that people like to buy through and have all the tools available. Even beyond our AI, which frankly is our big advantage on the advertising side, we also have a lot of tools and things that advertisers need to be successful on our platform.
This investment in technology is also a huge differentiator for us. The fifth differentiator is a new one for us, and it's, frankly, growing and becoming a differentiator for us, which is Connexity. We believe that a third of revenue for publishers will come from e-commerce in the not too distant future. We think that if you want to compete for publisher business in the future, you better have an e-commerce strategy. We bought Connexity for that reason, and that's something that we believe will be a differentiator for us as we go forward. We bought Connexity for, you know, we thought there was significant value in the, their intrinsic business. They're a business that's growing about 20% per year. They had 50% adjusted EBITDA margins, very profitable. Obviously, there's intrinsic value.
We also have talked multiple times about the $100 million in ex-TAC synergies that we expect to be able to generate over the next four years. Even beyond that, the third value to the deal for us is the strategic value. When we go to win publishers in the future, we believe we'll be in an advantaged place because we have an e-commerce solution for publishers. That third value is one that we haven't talked a lot about, but is a big reason why this was a good acquisition for us. I won't dive into a lot of detail. Bob did a good job about talking about some of the progress on our synergies to date.
We're making progress on all five of the synergies that we've laid out here, and you know, we're excited about the opportunity here again to bring in the $100 million of annual ex-TAC that we think we can get in synergies. The sixth and last differentiator that I'll talk about here is really one that's more applicable to 2022 and going forward, which is that in tough economic times, there's no advertising business and frankly, very few businesses in general that won't be affected by an economic downturn or a tough economic period. We feel good at Taboola because 85% of our revenue comes from performance advertising, and usually, performance advertising is the last to be cut in an economic downturn and basically is, it's the most robust in tough economic times.
What this chart is basically talking about is that if you look back at when COVID came on, so obviously COVID hit in early 2020, what you saw from 2019 to 2020 was that the percentage of ad dollars spent in performance actually went up from 63% to 67% from 2019 to 2020. The reason for that is the quote on the right, which is that advertisers basically decided that if there, if there's a tough economic time, I'm gonna spend on what I know is working. Performance advertising, you can measure exactly the outcome you want. They doubled down on performance and basically cut back on branding and other things that were not as effective for them.
We think going into 2022, that will also be a bit more robust than other advertising businesses because of the fact that performance advertising tends to hold up better in tough economic times. Now I actually wanna move and look more forward and talk about how we grow going forward. First of all, I think you've all seen this slide before if you've followed our story, but we expect about 40% of our growth going forward to come from new publisher supply, so bringing in new publishers and signing them up. You can see that historically, we've had a pretty consistent track record of bringing on new supply. We expect that to continue in the future. We expect the other 60% of our growth to come from growing that existing supply.
We have average NDRs of 115%-120% over time. We expect that to continue, and we'll continue to grow our existing publisher relationships mostly in two ways. One, we grow yield, so we continue to make more per ad spot over time. Two, we continuously innovate and bring out new opportunities for publishers to make money with us, like High Impact Placements and other ad formats and placements that we can help them make money from and that we make more money from as well. That's how we grow our existing publisher relationships. On top of that, you know, you've heard today a lot about our Anything and Anywhere strategies. We do expect growth to come from these as well going forward.
On the anywhere side of things, you heard about from Avner about all the work that we're doing with Samsung. We expect to continue to grow that going forward. Also mentioned that connected TV is still somewhere in our future. We think that can drive growth. On the anything side, obviously, you heard from Bob about all the investment that's going into Connexity, and we think we can accelerate growth there. You heard from Nadav and Lule today about our brands and agencies business and how it's growing, and we think we can continue to drive growth there. These are also growth avenues for us that we expect to drive higher growth as we go forward. Lastly, I wanted to talk a little bit about where we are investing?
The best way to think about where we're investing our business is since a large portion of our costs are basically people-related, is to look at where we're hiring. What the graph on the left is basically showing is this is our hiring plan for from 2021 to 2022. We're trying to grow from 1,786 people to 2,162 people. You can see which areas we're hiring in. We expect to hire 17% more people in what we classify as other cost of revenues. These are service people who are directly working with our customers to service our revenue. We have guided to 15%-17% ex-TAC growth. You can see that's growing right in line with our revenue, which is what you'd expect.
On the other hand, in G&A, we're planning on growing G&A headcount this year 12%. That'll, that's not growing as fast as our revenue because that's an area where we expect increased operating leverage over time. We expect our sales and marketing teams to grow 20% this year. Generally, we would expect our sales and marketing teams to roughly grow in line with sales. Again, we're guiding to 15%-17% sales growth. We'd expect that to grow roughly in line, but we are investing in a couple of areas. Brands and agencies that Nadav talked about is a big investment area for us, and we're investing in more sales resources for Connexity. Those are the two areas where we'll be over-indexing, so to speak, on sales and marketing hiring.
Lastly, you can see that we're gonna be trying to increase our R&D staff by 28%. That's our biggest area of investment, and that's consistent with what we've said our philosophy is, which kind of our philosophy is on the right, which is R&D is basically an investment decision for us. When we go into a particular year and we're thinking about our budget and our plan, we think in terms of what are the opportunities out there? Where could we be investing in technology that we think will help us drive value and grow over time? That's how we decide on our hiring for R&D. We'll always try and maintain that 30% adjusted EBITDA margin. We view R&D as flexible. We hire more if we see big opportunities, we hire less if we see fewer opportunities.
The 28% that we're trying to grow that is a good indicator that we see lots of opportunities to invest in technologies that we think will add value over time. I also wanted to mention that while we focus a lot on adjusted EBITDA, we focus on ex-TAC growth, we don't forget about cash flow. Adam has said repeatedly that we expect our adjusted EBITDA to convert into free cash flow at about a 60% rate over any reasonably long period of time. In fact, if you look at the last two years, so 2020 and 2021, you can see that we generated about $286 million of adjusted EBITDA. We generated about $146 million of free cash flow, which is a 51% conversion rate.
If you factor out two kind of unusual items, the cost of us going public and the M&A costs related to, mostly Connexity, but also some, Outbrain merger costs that unfortunately didn't go through but were unusual costs. If you factor those two out, we had a 61% conversion rate of adjusted EBITDA into cash. Over any reasonably long period of time, that's what we'd expect. Now, I'll also note that I get lots of questions, obviously, all the time about what are you gonna do with your cash? How are you gonna use it? Couple of things. First, we do believe that there's opportunities to use our cash as a competitive advantage, to win more publisher deals.
For instance, we'll sometimes prepay publishers a decent portion of the value to them of a deal over time, pay it up front as a way to win them over, to win new customers, or frankly, to lock in longer-term deals or other things that we think add value. We will use cash for that. We also believe that there's gonna be opportunities for M&A going forward. Not large ones like Connexity, but we think there's gonna be opportunities for tuck-in acquisitions, smaller technologies, acqui-hires, things along those lines. That's how we're thinking about how we use the cash. In summary, and, yes, I went fast through that 'cause I was trying to get you to lunch. You know, we think we've got a strong model. We're growing fast.
We've grown 25.5% over the last five years. We have good margins. We're at 30.7%, projected this year for our adjusted EBITDA margin. Strong balance sheet. We have positive net cash of about $31 million and $319 million of total cash on the balance sheet. As I've said, we are generating positive cash flow, and we expect to continue to do that as we go forward. Net-net, we think we're in a pretty good position. With that, I'm actually gonna invite back up to the stage our executive officers that presented today, and we're actually gonna go into a little bit of Q&A for people who wanna ask questions. By the way, just feel free to raise your hand.
Jenn or somebody will come running to you with a mic, and then we can, we'll try to answer the questions. Adam's gonna be our master of ceremonies and to try and direct questions to the appropriate people. Yours is not on.
No, it doesn't matter. Oh. All right. There are two ways this can go about. One, we can have an engaging Q&A session for 30 minutes, or two, none of you will ask a question, and we'll just be here for 30 minutes on stage quietly. But we're not going anywhere, so this is it. But go ahead. If you have any question, please fire away.
Hi, Andrew Boone from JMP. I'll kick things off. Steve, just on the billion-dollar guide that you guys laid out, can you square that with the 20% long-term growth target that you talked about kind of in the summary slide in terms of financials? You know, on my math, that's 17% growth to 2025. Just talk about the difference there and how you view that in terms of conservatism.
Yeah.
Adam, you know, one of the big takeaways from the strategy section was just increasing the time, right? Like, that's a big lever for you guys. Talk about how you guys do that, how you're thinking about that in terms of just adding more impressions overall.
Okay. We have a mic, by the way, to Jen to give people in the audience. Yeah, okay.
Yeah. I'll start with the first one. Your math is good. If we did exactly $1 billion in 2025 and we hit the midpoint of our guidance in 2022, that'd be 17.8% growth. Generally speaking, you know, Adam, I think when he talked about that, did say that we expect to cross that or to. You know, our goal is always to exceed our goal, our targets. We would hope to do better. But, you know, even at 17.8%, that's, you know, I think a decent acceleration beyond where we even are guiding this year. Overall, you know, it's a conservative goal that I think we believe we feel good we can beat.
Can you double-click on the acceleration of, like, that? That's inherently been my question.
You mean what will accelerate it?
Yeah. Talk about the leverage.
I think, first of all, I talked a little bit about some of those growth vectors. Our growth initiatives that we've had, we believe that's part of what will help us accelerate our growth over time. Frankly, you know, to be very blunt about it, Microsoft has been a drag on our growth historically. We think with the new arrangement and the new deal that we have, plus the bidder that we're building with them, that we actually can, it'll be less of a drag going forward. I think that'll also be part of it.
Hi. On time, there are many things that can increase user engagement. I think Tom spoke about to drive more time with consumers, you have to do more than just presenting ads. If you look at all of our investments, you know, Kristin spoke about Homepage For You as an example. That's Taboola spending time on the homepage and section fronts, surfacing things people might like. If you think about what Nadav has presented, our High Impact Placements, that's video on mid-article placements, and High Impact Placements on a publisher's site. That's video engagement. If you think about what Avner spoke about, distribution of news, most of that is organic news that can be surfaced to consumers through multiple touch points, notification on your screen. With Xiaomi, it's awake screen. It could be a Minus One in your browser.
All of those touch points can help us get more time with consumers. Gil spoke about AI, and AI essentially can just make sure that we're more relevant to you, so you get to use us more and you're more engaged with us. Really, all of those investments help us be more relevant for consumers wherever they might be. The combination of investing in anything, and then you saw Bob speak about commerce. If you can imagine over time, The New York Times has Wirecutter. I used it today has Reviews.com. If you can imagine over time, thousands of publishers will have an e-commerce section. That's another opportunity to engage with consumers around high intent moments of time.
All of those investments we're making are long-term investments, but on the other side of it, you're having better experiences for consumers, happier people, more time, better yield. Tom was conservative. He said only 10x. I told him, say 50x, 100x. He said, Adam, you have to be responsible. He did say, but 10x at least, you know, over time.
Hi. James Kopelman with Cowen. Considering your roughly 13,000 advertiser base and the disparity with Facebook or Google at millions of advertisers, it seems like a real opportunity for Taboola over time. As you look out over the next year or two, how much higher can the advertising base go? How do you view the longer tail opportunity? Should we expect to see a roughly 90% direct relationship remain relatively consistent as that evolves? Just a quick follow-up, maybe you guys could provide some additional color on in terms of geographies and regions as you look out in the world, what are the most important geos going forward? I'm not sure if you could provide an updated rough geo split between your biggest regions, U.S. versus international, maybe Europe and a couple others. Thank you.
Thanks. I'm gonna let Eldad take this one. It was a long question with a lot of details, and Eldad will take it.
Yeah. Way too detailed for you or I, Adam.
Yeah.
Maybe I'll start with the number of advertisers. As you mentioned, we're constantly growing the number of advertisers. It's a small fraction when compared to the Facebook and the Googles of the world. In order to evolve the number of advertisers, we need to invest in a number of places. One is obviously sales, account management, and that presence. The other is in different aspects of technology. We're not aggressively going after the long tail at this point. We're actually going after performance advertisers that can scale. As was mentioned earlier, we have enterprise and outside sales relationship based work with advertisers, brands and agencies, and those are growing nicely. We have inside sales, which we're actually ramping up and creating more of as a segmented business of ours.
On the online acquisition, we are continuing to invest in bringing them in, but we're not looking to grow these to a dramatic scale to mom and pop shops just yet. From a geo perspective, our operation, we currently operate in numerous markets. We operate in quite a few markets with feet on the ground. On the other end, we have local hubs. For every continent, pretty much we have a hub with inside sales and account management that support multiple countries within that continent.
You had asked about percentage of. I don't think I forget exactly what we've disclosed. I think we have talked about the fact that U.S. is about 40% of our business. Rest of world is about 60%. If you're asking in terms of exposure to kinda what's going on in Russia today, very small direct exposure to that. Our exposure to that would be more if it becomes an economic contagion or, you know, kinda affects the economy as a whole. But our direct exposure is sub 1%.
Great. Thanks, guys.
Going back to kind of the AI and the yield improvements that kind of Gil touched on, he had one slide that talked about the five things that you're working on AI in terms of going forward. I can list them if you guys probably know this off the memory better than I do. As we think about kind of contextual and intent at a higher scale, faster way to explore new matching patterns, models which allow scaling more diverse types of advertisers, model lifetime value and increased engagement over time, and then smarter bidding on any supply. Talk about, like, what's low-hanging fruit, what's more of a midterm goal that may be out there versus more near-term.
Like, how do we square that against what we saw in your yield chart as well as putting, you know, $1 billion of ex-TAC revenue going forward?
Gil, do you wanna start? Take a mic and.
Thanks. So all of them. It's all easy. No, I'm joking. I think we've learned that usually there is a low-hanging fruit when you lift an engine that you haven't opened yet, the engine cover that you haven't opened yet. New initiatives usually run faster. Homepage For You, we see an opportunity to make a big difference quite quickly. Hopefully that will unfold well. Same thing about other new initiatives like the bidder, which is progressing real well. Otherwise, it's one of the reasons we're going in multiple directions and the list there was partial, obviously, we're doing other things is that we diversify our work, so we increase the chances of success.
It's sometimes a bit difficult to know exactly what research initiative is going to unfold and in what magnitude. The reason to cope with that is by diversifying our work. I hope that answers the question.
This is much nicer than the usual investor meetings we have because I can just redirect questions to anyone else.
Gary Marks at [Celebra. I want to ask a question about kind of the video opportunity. I know historically you had kind of the indirect with ConvertMedia, and now you have kind of a lot more direct pub relationships you guys have talked about. Can you maybe kind of talk about the opportunity in direct video and potentially what that would look like for yield and monetization?
Yeah. Nadav, do you wanna?
Sure. Hi. I mean, we've seen the trend of advertisers using video to convey messages for years now, and we've seen many platforms exploit that trend. As I said in my presentation, we're investing in different placements on our network so that we can provide advertisers a high viewability and high impact experience. We've been working with many global agencies around the world so that they, when they have those buying decision moments, they consider Taboola as part of their video media mix, and we're seeing results. We're seeing results in the top of the feed placement, which is a very visible ad unit. We're seeing results in mid-article and homepage placements, and we're gonna continue to expand that opportunity.
We think that, depending on the ad format, the video ad format, we're seeing advertisers start to play around with it. We're seeing a lot of success with short form videos, kinda like the bumper ads on YouTube, the 6-second versions, where typically advertisers can just copy-paste from their social and use it on Taboola. The same video asset works very well from a performance standpoint, and we're seeing good traction on the long form video as well, so 30 seconds and beyond. Again, we offer many different types of buying models for those advertisers, CPM, VCPM, CPCV, and so advertisers feel very comfortable with the options that they have with us, and we're seeing that ramp up.
I think there was a question before about when do we think that 90% of direct revenue will remain? I'm not sure. I think someone asked that earlier, but yeah. We do think that one of the fact that most advertisers buy from Taboola directly using our own technology is something that will stay forever. Now whether that's 90 or close to that, we still think that's one of our strength. Investing in technology, having good data, the engine and fuel analogy in the rocket, and then that will help us always be successful and be predictable as a business versus traditionally how advertising companies build themselves, which is more hoping advertisers come their way but not necessarily working with them directly.
Steve, this is actually more for Adam versus Steve, but you kind of set up R&D as an allocation question in terms of allocating dollars towards R&D as well as the investments. Adam, as you know, as the CEO, how do you think about the return on investment from R&D, given the $100 million of spend that you guys have, which is, you know, in excess of peers more directly and then across ad tech more broadly? Just talk to us about, you know, where you guys do throw dollars at something. How do you think about peeling it off over time, or how do you scale the R&D spend that you guys have? Talk about it in terms of just a capital allocation question.
I can talk about the value of why we do it and how I see that, what's the return on investment, and we can talk more about other layers and that can chime in as well. I see it as there are two fundamental things it helps us achieve. The first one is, you know, generate growth in yield, which means that advertisers are successful and publishers create more value and revenue from Taboola. Which is a competitive advantage because it means that to take away business from Taboola, you have to be losing money a lot of times. We wanna make sure that we're very successful in attracting advertisers and making publishers grow revenue. Investing in R&D helps us drive yield growth.
That's the revenue component, and we believe we're top three revenue source most of the time to our publishers. You heard today from eToro that Taboola is a top channel as an advertiser outside of the walled garden. That's the first reason why, what's the return on investment from an R&D perspective. The second thing is differentiation. How are you different than just more than money? To me, the fact Kristin called me on a holiday day, which I hope you will forever continue to do.
For the record, Adam tells that story all the time.
I am.
You should feel free.
I am.
I was like, "Yes, she's telling it too now.
Yes. Validating the story. It is true. The reason I like that story is because she didn't call me about a CPM question or about a, you know, about growing revenue. She called me about something that she's passionate about, which is editorial growth, homepage, the journey of the consumer on McClatchy, which has value over the next many, many years, and that makes that conversation different than an ad tech conversation. That is also a return on investment because it means you're different, it means you win more than you lose, it means you can capture more value than other companies who has to pay more than what you pay. All those dynamics eventually make you be a company that can be here forever, you know? That's why R&D to us is an investment.
Because one, from the yield perspective, it makes us competitive. The revenue is growing, all the things. All good things happen when yield goes up. The second thing is that we're different because I think if you compare Taboola to many other companies in the history of time in the advertising space in the open web, none of them would get a call from Kristin. I always wanna get that call. That is a return on investment from an R&D perspective. I feel that if you have-
Yeah. I want to add to that. You've recently heard about the 10-year deal. The 10-year deal is only possible because we've built the right platform, continue to invest in it, and are building long-term value for our publisher partners and for our advertisers. That investment is what drives the long-term relationship, the predictability, and the growing margin that we're seeing.
Okay. Can you guys maybe talk a little bit about the different growth rates in the different parts of the Taboola business? It sounds like Microsoft's a little lower, other things are clearly a lot higher. Can you maybe go through some of those and break them out a little bit? The second question, I guess, is thinking about the mobile slide that we saw. Can you maybe talk about that relative to like what it means to McClatchy as they sort of go through time with you all, and how their journey will sort of tie into some of those numbers that you talked about?
Eldad, you wanna talk about growth?
I didn't hear the first question.
The growth engines, what drives growth for the board.
Relative growth rates of different segments of our business.
I think we separated the business into really core and then the recommend anything and recommend anywhere. When we look at our core business, we look at different types of geos on the one hand. Obviously we're seeing the growth rate in the emerging geos and where we have operated, I would say at a later stage is also the higher growth rate, which is common. On the other end of that, we look at enterprise and SMB. When we look at the advertiser side, it was mentioned earlier, there's performance advertisers at scale. There are millions of performance advertisers.
We believe that what we call growth advertising, which is in essence SMB and performance advertisers, is a huge opportunity for growth and will be the faster growth rates that we see in the core business. When we expand that beyond the core business, and Nadav mentioned the brands and agencies which are growing faster than our core or average growth rate for the company. We're not reporting results separately for Taboola News, but we see that as a very high growth opportunity as well.
What is kind of struggling to grow then at the corporate, overall corporate rate?
The one significant drag on our growth percentage that we've had historically has been frankly, Microsoft. You can. It's not like it's not public. You've seen what percentage of our business it is. When we did our initial F1, F4 for the filing, it was 20%. It's now 15% or 18%, sorry. 17%? What does Microsoft report at now? 17%? Yeah, so it's at 17%. You can do the math. You can see that it's been a drag on our business. But that's where we're kind of excited about the new bidder, and that has an opportunity to grow both our relationship with Microsoft, hopefully over time, but also use that bidder elsewhere.
On the web and another, you know, use it as another growth driver. That's Microsoft's really been the one, what I would call drag on our business. Even, you know, Eldad is correct, our newer geos like LATAM and APAC, which are two of our newer ones, are growing faster than say U.S. and EMEA, but not. It's all within a range of like they're similar. Even the U.S., which is our first real core market, is growing at a nice pace, not a big drag on our growth rates. Some of the new initiatives like Taboola News are our growth initiatives gonna be higher growth rates over time. Generally speaking, the one drag we have is pretty much Microsoft.
I actually have a question from an analyst that couldn't be here. Well, two questions actually. First is, how much of deep learning tech uses off-the-shelf tech from Google, Microsoft, or Amazon Cloud? Then the second is, how are you thinking about Google shifting to topics which is a more contextual format than FLoC?
I have the mic, so I'll start with the Google one, and Gil, by the time you get the mic, you'll get the.
Yeah.
Yeah. About Google evolving their privacy dynamics to topics, we think first of all they speak about this, their changes for the past year or two, and that will take more time. Overall, we do think that Google will be more in the business of helping advertisers be successful because they care. We think that will be somewhat beneficial to us and other companies because they'll provide some value in the context of helping people, you know, target using their cohorts and topics and things of that nature. Saying that, we're quite encouraged by the fact that we don't rely on third-party cookies and whichever future they land on, we believe that Apple represents probably the highest standard in terms of privacy, and we're doing very well in that environment.
Many of these things and changes are upside opportunity for us. We're trying to be conservative, but that's how we see the future. Overall, as you know, as a company, we're very happy that the world is moving to a more privacy-driven future because it's good for consumers, and it's good for the world, and we're ready for it. If you wanna talk about.
Yeah. The question was about whether we use off-the-shelf, I'm guessing models and stuff like that from Google, Facebook, and so on. I would say very, very little. Even when we do, we usually take those models which are built for a different problem, and we fine-tune them or retrain them on our own data to make it suitable for our own needs. We invested a lot in. There's a whole field called transfer learning, where you can take data that was or AI that was built for one problem, and you transfer it to another problem. We do it internally a lot, from one problem to another, which works really well. We do it a little bit also from models from the world, but it's very little.
I'm not gonna get into specifics, but it's very little.
Let me just add to that, so, you know, the genius of Gil's team is how you use those tools and how you set up the models and how you get the data to the, you know, to the model. He talked about the challenge of the massive scale and everything. To a certain extent, asking do you use off-the-shelf tools, it's like saying in my finance team, are we using Excel? Yeah, we are, but that doesn't. Excel doesn't do the job for you. There's still a job to be done. Sorry.
I want to clarify something. Maybe it wasn't so clear. Okay. We don't think it makes sense to build a tool that you can find elsewhere and it works well as a platform. We do. Like, I had a slide with the platform that we use. We do think it makes sense to use platforms that were built by other companies and then use them for our needs as a platform. That makes sense. Building a deep learning suite, let's say, that we have three out there already, which are really good, doesn't make sense. Makes more sense to take those as a platform and then make it work for your problem.
The platform, we definitely use platforms from that were built by other companies, usually open source platforms. Specifically for models, we do our own, that's more or less the thing.
We have about five minutes left, if anybody has more questions. Adam will sit here for five minutes and stare at you. He was serious about that.
On the 40% of revenue growth that's coming from new publishers, one of the questions we frequently get is what's new, right? It seems like everyone is now using recommendation engine. Can you just talk about just how penetrated kind of the market is, or how much is that coming from anywhere in that strategy?
How can I answer it? Are we in-
If I think about the growth
Thanks.
It's basically the new revenue, how much of that is greenfield versus steals from competitors versus whole new categories of publishers.
Oh.
That we're bringing in, like.
Yeah. You should holistically think about the publisher space in three tiers. You have top tier publishers, mid-tier publishers, long tail publishers, small ones. We haven't penetrated long tail at all. Mid-tail, we're small, but growing, and so there's a lot of greenfield. Top tail, like the enterprise, depends on the market. In the U.S., we could be tens of percents, or even 50%, depending which market. Then in international markets, we could be much smaller. There's always a lot of room to grow. Some of them are what we call liberation from other companies 'cause they're liberated, you know, to have a great future. Then other ones are new completely.
That's before we're getting into High Impact Placements where we're replacing SSPs and banners, a whole new types of companies, and that's before OEM, as well. On the publisher side, a lot of it is new, definitely if it's mid-tier. On in the top of the market, if it's a very mature market, it can be a liberation, and if it's international, it can be new. I don't think we give specific numbers, but that, those are the dynamics.
Kristin didn't know until just now that she was a liberation.
I'm sorry.
Okay. I think if there's no more questions, then we will let Adam wrap up the day for us.
I'll just wanna say thanks everyone for joining our journey back when we went public. Thank you is the main word in my mind, you know, for joining our journey back when we went public. Today, you know, you're very engaged with us, which is fun for us to learn about ourselves and what people care about. I wanna thank our partners and clients who choose Taboola and who joined us today. Thanks Kristin, Avner, Lule, who are here today with us. Our investors and analysts who are on our side. Taboolars, I'm so proud to be one of you all. Many of the people who are here today, we've been working together for 10 years. Evangeline, I will always FaceTime you when I have an issue, forever and ever.
I'm proud to be at Taboola. Three things that happen now, I mean, post-lunch. One, we have product demos, so our product leaders are here. If you wanna see, don't ask financial questions, please. You know, we should all behave. The idea is to learn about the product, so you get to see firsthand a great experience about Newsroom and our ad console and everything that we talked about today. You know, intimate setting, so feel free to enjoy that. We have cocktail session for those who like a great martini at 6:00 P.M., where you can spend time with us and talk about life. We have a goodie bag, which Avner mentioned.
He did not know that there is one, but there is a very special bag because it has my favorite LEGO inside of it. On your way out, make sure you, not Taboola employees, non-Taboola employees, take a Taboola goodie bag, enjoy it, and stay in touch with us. Thank you, everyone.