【數位時代Podcast】EP237｜膽大妄為的「虎之霸」（Chutzpah) 文化是什麼？為何決定以色列和台灣公司的截然不同路徑？ft. Taboola總裁暨營運長Eldad Maniv

Hello. Thank you for having me.

Sure. So Taboola is an international company, NASDAQ traded, 2,000 employees, close to $2 billion in revenue, and we serve brands by providing them outcomes at scale. So when an advertiser or a brand wants to grow their business and get to new customers, we're able to provide them performance-based outcomes beyond social and beyond search.

Right. Yeah. So we're in essence a two-sided network. We work with brands on the one end and with publishers, media companies on the other end. For the brands, we deliver performance to find the users that are likely to buy, subscribe, register with them as that's what they need as new customers. For the publisher side, we provide an ability to monetize, to generate revenue, and to be able to generate an audience, engage the audience that comes to their sites.

So it's a bit different than search because the user doesn't ask or write a query of what they're searching for. What we in essence do is when you go to a publisher site to read the news, to consume content, we actually are able to see what you're interested in and we know a lot about you as you browse the open web, different sites. And then based on that, using our data that we collect and using recommendation algorithms, we're able to predict which users are going to be interested in what and serve them the right content and the right advertising.

Yeah. So advertisers generally, when they need performance, they go to Google search, they go to social like Meta, and there's the other third option which is the open web. We're the open web solution for that. So when you need scale and you see diminishing returns from your search and social campaigns, we're the option to provide more performance conversions at scale on the open web.

So the program started back in the 80s of the previous century where the defense ministry wanted to create R&D leadership. And so they chose 30 to 50 young individuals before starting their military service and had them sign an 8-year contract where the first three years were training, doing their bachelor's degree in either math, physics, or computer science along with training in the military to learn the needs of the different units, and then followed by five years of being R&D leaders.

Well, I think that the startup bug is very common in Israel. It's very common to want to be part of a startup, to start a startup yourself. It's a thriving community. There are thousands of startups. I think what we've seen in the last 5 to 10 years is there are also a lot of mature companies. Israeli founders used to start companies and sell them very quickly as quickly as they possibly could. Now I think that the trend is build longer-lasting companies, take them public and grow them.

So I don't know much about the Taiwan startup business in detail, but I do know from our operation here that there is very talented, there's amazing talent here. There's good engineering talent. There is good business understanding. So I understand that there is amazing potential. I think what we see in Israel is a limitation that actually drove success. One limitation is there is no local market. The Israeli market is too small for any company to really build its business. So any Israeli startup from the very first day is an international global company because customers are either in the US, in Europe or in Asia. So that mindset means that from day one you know how to interact. Now from access to capital markets and investments, the investment community in Israel is very developed. There are local venture capital firms but there are also global venture capital firms who have offices in Israel. So as you start even from round A, you have access to global money because it's available in Israel. I think the third part of this is really that there's a term called in Hebrew or in Israel which means really being out there and being aggressive in a good way, asking for something even if you will get a hard no. So, Israeli founders just try and a lot of times fail. But even if you get one out of 10 times a positive answer, you've tried. Whereas in a lot of other cultures, if you're not completely certain that you will succeed, you will not go and take that path. So, I think it's a lot of risk-taking.

They don't really teach you. I think it's slightly negative term because you're out of line in some of what you're doing. But somehow it comes with being in Israel.

It's a noun of having a... I'm not sure what the right translation is. You can look it up, but it's really having the nerve to do something.

Mhm. Okay. But most founders have this kind of chutzpah. So a lot of the founders in Israel, the younger ones are graduates of military programs and have learned usually at a relatively early age, young age, to manage large projects or large big units and be managers. So people get the managerial entrepreneurial experience through military service early which is one factor. The other is the culture of having that. So trying things and even if you fail trying again and again and having the nerve to try and to ask for things that maybe you don't even deserve but you try to get.

I'm not sure. I think it's just in there. I think so. Who to choose with you, I think is one of the most important decisions that any founder and any manager can take. I think the team is the most important part because businesses change and markets change and everything changes but the people around you, if you chose the right people, you're able to make the changes needed and to adapt and to pivot and take the right steps. So by the way, I believe in it to the level when Taboola started, I joined when it was 40 people, I started interviewing every single person up until we were almost a thousand people. I interviewed every new person that was about to join because choosing the right people I thought was the most important thing.

So I think Taboola like most other companies try to raise money. The founder Adam Singolda came with an idea that initially we will work with publishers and we will put links on the publisher sites that will direct the visitor to the site to a different site. And the venture capital community in this case in Silicon Valley said are you crazy? No publisher will ever put a link on their site to go to their competitor's site and said we're not funding it. Eventually with enough push and enough he managed to raise funding and if you look at the market today the publishers sites are full with links that exit the site and go to other sites including competitor publishers etc. So it seemed to work. So again, having the right idea and pushing forward even when you get numerous nos is really what a founder needs to be able to do, right?

Yes, absolutely.

Yes, we have a few. Our CTO is a graduate. We have a leader in our... Tal is our SVP of R&D. We have our CTO named Leor who's a Talpiot graduate. We have one of the leaders in our ALGO department which is a Talpiot graduate. But I would say this is one program. There are a lot of very capable leaders in Taboola on the tech organization side both in Israel and here locally in Taiwan. We have more than 40 AI specialists working here. We also have an office in LA and one in Budapest, Hungary.

So the five years indeed have put us through a test of how we operate through a quite difficult environment as you said starting with COVID that put certain limitations and continuing with a war in Israel. There are a few things. One is how do you set up your infrastructure and from an infrastructure hardware perspective we're very distributed globally and so the risk is very much hedged by having data centers in multiple regions, multiple countries, multiple areas. We're also diverse from where we operate. So we have people in Israel, we have people in the US, in Europe, in Asia. And what has happened is that the culture in the company and quite a few other companies that have gone through this in Israel have tightened up and people have compensated for others that are not available. So people worked harder. People in other countries took some of the roles. I think what we needed to do as management was one, build contingency plans. What if we have people not available for us for a longer period of time? How do we actually operate? But more than that, focus. So we looked at everything we do and then we said what is really top priority? What is second priority? And of course top priority is supporting our customers and making sure that they don't even feel that there is any hiccup. And then in our strategy forward, what is top priority? What is second? What is third? And making sure that whatever resources are available, they're directed at what is really most important. And so we've been able to operate during those five years almost in an uninterrupted way. Of course, customers have not seen any impact from this. And we've been able to continuously drive our top priorities. In some cases, there were lower priorities which we would have loved to have done but were not able to but those were the lower priorities. So we were very successful in managing it in the right way.

So I think it changed over time, but as I said, first priority is customers get uninterrupted service. They know what they paid for, what they bought, and they need to get the best level of service regardless of what the company is facing. I think that's an almost absolute truth in any environment. Beyond that, for example in the last year we were working on a release that was a big announcement by the company about a month ago that we were expanding beyond our initial operation and being a native advertising company into doing everything performance. And that means being able to serve any type of brand, agency, a business really in driving outcomes for them. And that expansion required significant efforts. And that was our number one focus over the last year. And we've seen the team really drive that execution and actually deliver more than what we've seen from a velocity perspective than what we've seen in the last few years before that.

So I think it starts with the culture. The first thing we did when COVID started is we said we now need to be much much closer to the employees because things are changing very quickly. There is fear, there is uncertainty. So we set an all hands meeting with everyone in the company and reminder we're about 2,000 people. We were slightly less than that but not dramatically less every week. So every week the top management team would get on a call and update everyone what's going on and accept any question from anyone on any topic and people started asking questions that started with their personal safety with COVID and continued with their employment safety and then with how we're going to progress on priorities and being that close to the employees that frequently I think made that connection much easier. We also built contingency plans of what happens in certain situations and how we deal with infrastructure, how we deal with reduced capacity in development and sales. So we built the plans but I think overcommunication is really a key.

So I meet with quite a few different media companies, publisher CEOs, chief digital officers, and the AI storm is here. It's not coming. It's already here. And seeing how they react to it is interesting. It ranges. There is a full spectrum. I meet with some CEOs who are actively changing their company. And I'm meeting with others that are almost covering their eyes and saying it'll come at some point in time. I'm not doing anything. I think there is both a threat here and an opportunity. The threat is on a few levels. One is consumer behavior. So if you're in media, you know that people used to buy newspapers 15, 20 years ago and it was a very profitable business. Then came online, then came mobile, then came social networks. And each one of those negatively impacted the publishers themselves. And now GenAI is coming. And the question is, how will consumers consume content? How will I get my news? Will I go to NewYorkTimes.com or will I ask my OpenAI assistant or will I get it from my personal robot that will also deliver the news to me? And how do publishers become part of that? So there is a risk there and they need to adapt and leverage what they have which is a unique asset which is real premium content that is researched and created and delivered and find the right distribution channel for it. So that's a risk and really a threat to publishers because the ones that won't adapt I'm not sure will survive. The other is they need to improve their internal performance. So if other publishers around them are able to research a topic by leveraging GenAI and be able to produce certain aspects of the content delivery chain in a more efficient way, they won't be competitive. So those are on the threat side, but then there is a huge opportunity. This opens up a new way to consume content. So, for example, let's imagine that I'm a journalist and I created a great piece of content. I did my research and I wrote an article. Today, that article is a one-fits-all. I wrote an article and now it doesn't matter if I'm young or old, if I'm speaking this language or another, if I generally consume content in video or in text, I get one piece of content. GenAI can take that premium content and actually transform this into hundred different pieces of content. A video version, an audio version, a text version, a children's version, an older age version, a specific community version, all based on the same research. And if you have the right personalization engine, you can now give the right content in the right format to the user and that engagement will be dramatically higher because it'll fit that user. So there's also an amazing opportunity here in becoming much more productive in giving your audience a very personalized approach. But there's a lot to do.

So there's always a risk, but I think that there's a lot of hope because I don't want my kids to consume their news content from TikTok.

So I have a 26, 24, and a 19 year old. They're consuming content. And interestingly enough, they can figure out what is good content, what is fake content. They've adapted and they've learned, but I don't want them to get that content at a level that's not premium. And publishers have a unique asset. They know how to create very good content, deeply researched, deeply analyzed and very engaging and that is an asset. Now the question is how is it used? On the other hand, they're not technology companies and they're facing the big tech of the world which is really trying to eat their lunch by accessing consumers directly, the walled gardens, the social networks. So the key is to enable those premium publishers with the right technology to be able to deliver consumers the right experience. I'll start with a very basic fact. When we go to a publisher's website, we all get the same experience. When we go into Instagram or TikTok, each one of us gets a personalized experience. Why are publishers not personalizing their homepage, their articles, their content towards that? So part of what we're doing at Taboola is exactly that, building the technology to enable publishers to customize, personalize their content per every specific user out there.

Yeah. So there are various steps in advertising from building the strategy all the way to building the creatives and driving the distribution. And I think all of those are going to be impacted by AI. Some of those are already impacted. At Taboola for example, we created AI-driven technology to help an advertiser create a campaign. So we have an AI assistant we call Abby that interacts with the advertiser and asks them specific questions. What are you trying to achieve? What's your marketing goal? What exactly are your cost structures that you want to achieve? What is the overall messaging? And it builds a full campaign including the creative, including everything you need for you. And by the way most of that technology is built by our local team here in Taiwan which we're very proud of. So there are already changes happening in how advertising works and I think it will be even more than that. As an example I can now do a quick analysis of my competition and see what a marketing strategy they're using by just looking at their campaigns. Is this a fear strategy? Is this a discount strategy? What exactly are they doing? And I can adapt my strategy according to what my competition is doing. All of that is very easily done with AI.

There are a few reasons for this. Anywhere ranging from our contact with Abby who was the representative of Taiwan to Israel which is really helpful and has assisted us a lot to the name of one of our developers' daughters to actually an acronym of an advertiser's best buddy. So it's really... it's already now back into speculation mode. What really started it?

We'll add that one too.

I think how consumers will engage with AI is still very much an open question. I do think we see assistants and agents becoming much more mainstream in the next year or two. I already consume content by asking ChatGPT or other LLMs, what's the news? The level of news I'm getting today is still subpar, but I think it's the right engagement methodology. So, we will be investing more in enabling publishers to drive their content that way. But how exactly engagement of consumers will be in five years, I think is really too early to tell.

So that's a great question because the promise of productivity is an amazing one. When I saw a very early version of ChatGPT, I said this is going to change the world dramatically and we need to start working on it ASAP. So I created a small team. Actually our previous leader for our global R&D organization decided to leave his job of managing 400 to 500 people and start focusing on just how to use GenAI. And he was that leader of that team and we started playing around with how can we become more productive and it's been a long ride. We today have different internal projects that have been automated and have driven significant productivity. Some of it is in our for example content review. We used to review manually every piece of content on our network to ensure that it's the right quality and adheres to our I don't know 300-page book policy of what is allowed on the network. We started building generative AI capabilities into this. Initially traditional AI, now generative AI that will review the content based on the policy and accept or reject or send for further review and now it's serving a huge part of the overall flow which is helping us in being more productive. We tried the test in seeing can we build a legal engine that answers our sales team's questions. We looked at how to be more productive in R&D. What tools can make R&D more productive? I can say that there are a few findings here from our own usage. One, there is a huge opportunity and it needs to be developed. I personally believe that companies that will not focus on this will be less productive, become less competitive and just disappear. So I think it's not a question of if you do, it's how quickly and how aggressively you pursue this. We've also learned that it's a different development methodology. So for example in the past if I had an internal project I have internal users that have some need for automation for productivity I would bring someone who is like a product manager that would research the customer's use and needs and write a spec and then there would be a development team that would take this spec and cut it into small pieces chunks that are digestible and start developing it and after a few months or a few weeks depending on the scope bring a first version and start iterating here and the consumer or the user would only see it when it's built here. The main person, the person in the middle in the center of all of this is actually the consumer. The tech team can help but that consumer can write the right prompts, the right business knowledge and be the center of a team. So what we've built is small teams with the user consumer customer in the middle with a very small development team that builds the infrastructure and they're building very rapidly internal projects to drive this.

I think we're not there yet and I think that there are a lot of different interactions that need to be managed. I think the governments are very much behind like they usually are in building the right regulation but I don't think it's something that is going to stop anytime soon. So I'm an optimist so I'm hopeful but there's also concern here.

Mhm. So the question is what exactly is intelligence here? Is having all the knowledge being able to answer questions a measure of intelligence? And if so, intelligence by AI is soon going to surpass what we do as humans. So does that mean we're at risk? I leave the question open.

So the Taiwan office has two different elements to it. One is a business office which is our local sales team. We see a big opportunity in the local market and we have a business team here, sales, account management both to the publisher side and to the advertiser side working very closely hand-in-hand with that business. The other part of that is the development side. So our Taiwan team is one of four global R&D teams offices that we have and is the only one in Asia and it's a core center for our artificial intelligence development. So, we're very proud of the talent we have here. And one of the unique capabilities that we have in the local market is having that AI expertise that can actually serve in close relationship with our media partners, with our advertising partners and being able to deliver the right solution and adapt and customize to local needs. So it's both part of the global R&D effort but also serves the local business here in Taiwan.

Yes.

So I've been impressed working with our Taiwan team for a while now. Visiting the country, I have to say that it just amplifies that a 100 times. But I see a raw talent that is quite amazing. I see AI expertise. I see a culture of delivery and hard work and a culture that adapts well with the culture we have at the company of trying to deliver impact. It's not about developing. It's not about just purely working. It's about bringing the customers value and driving impact and we see our teams here amazingly focused and dedicated in doing that. So, I'm impressed. I also see differences in cultures. We have an Israeli based culture which we mentioned earlier which has some chutzpah of being a bit more aggressive in a good way, heated discussions and the culture here is more reserved and getting the two teams to work together and to actually engage is amazing because once they do the results are outstanding. So to encourage the chutzpah to localize here maybe not to that extent but at least be accepting and be more open while the Israelis need to listen more and be more attentive. So it's a balance between both cultures to try to work with the other culture. But by doing so, I think that both the strengths of both cultures actually appears and it's a great mix together.

And thank you for hosting me.