One-Stop Campaigns: How Generative AI Is Transforming Digital Advertising: Interview with Adam Singolda, Founder and CEO of Taboola

Taboola has been a leading content recommendation company for several years. Could you briefly summarize the main areas of your business?

We see ourselves as the world's largest discovery platform. We have exclusive partnerships with many of the world's top publishers and serve 360 billion content recommendations to over one billion people across the web each month. Taboola helps publishers monetize their sites on all channels and increase traffic. You have probably seen our feed serving recommendations on sites like Süddeutsche Zeitung, CNBC, NBC News, Le Figaro, El País, The Independent and The Weather Channel. With massive scale, unique content consumption data and world-class AI technology, we help thousands of advertisers reach their audiences with compelling native ads in a brand-safe environment when consumers are most open for discovery.

It is fair to say that you have been one of the earliest adopters of GenAI. Taboola has truly embraced the technology. I read that 25% of the ads you now distribute are being created by your AdMaker tools.

Yes, and mostly as self-service. We offer one-stop service campaigns. Most of our revenue comes from advertisers who buy from Taboola directly – not through agencies or programmatic advertising, which is an automatic way of buying. They open a campaign on our dashboard and then on their website and directly start their campaign.

ADAM SINGOLDA

ABOUT ADAM SINGOLDA

Adam Singolda is the founder and CEO of Taboola. He was born in Israel and graduated from the IDF Officers Academy and the Mamram IDF Computer Science Program. Adam has been a keynote speaker at various events, such as the Digital Innovators’ Summit. He has also received several awards, such as “Entrepreneur of the Year” by Geektime and “30 under 30” by TheMarker.

ABOUT TABOOLA

Taboola is the leading content discovery and native advertising platform. Established in 2007, Taboola serves as a bridge between publishers and advertisers, recommending relevant content to users and facilitating engagement. It was founded in 2007 in Israel by Adam Singolda and later moved its headquarters to New York City. Taboola uses deep learning, AI and large datasets to help more than a billion people discover what is interesting and new online. Taboola also offers solutions for brand promotion, content distribution and customer engagement and is one of the pioneers of applied Generative AI in online advertising and publishing. Taboola is publicly listed and valued at over a billion dollars with over 1,800 employees and 24 offices worldwide.

www.taboola.com

THE INTERVIEWER

The interview was conducted by Mark Heitmann in January 2024.

For these self-service campaigns to work, you need to identify the right consumers and create copy that works. Do you see GenAI as an extension of existing activities or as an entirely new operation for Taboola?

There are two fundamental schools of thought about the uses of AI. The core AI is powered by deep learning, which is mainly geared towards matchmaking. That's what Facebook, Instagram or TikTok are really good at. There aren’t many good deep-learning engineers in the world who can match what TikTok and Facebook are doing. Their matchmaking is really powerful. This is old-school AI, the hardest deep learning of all time, and it can be magic. Then you have GenAI, which centers on the generation of content. GenAI is a whole new universe using deep learning and LLMs. Different purposes, engineers, skill sets, and data. They are very different.

To offer one-stop services, Taboola needs to master both. How do you target the right people and create content that works for them?

We try to be data- and execution-driven. We have 600 million people a day and generate about $1.5 billion a year in revenue. We show them articles, videos and products and have less than a second to know who they are and to predict what they like. That's the matchmaking, and it is very hard. And based on customer interactions, we tested whether we could use that unique data to prompt little models and started to create titles and text. We took advantage of our historical information with real clients and used the data to generate relevant creatives with GenAI technologies. We offered it to clients and over time, one in four random advertisers that came in through the self-service used GenAI to create their campaigns.

What was the feedback? What has created the most value so far?

Our customers liked it and kept using it. It was a productivity tool for them; it was easy to use and it saved them time. And at Taboola we have a second benefit, which is important in the world of advertising and journalism: moderation and editorial quality. If you are putting an ad on The Independent, ESPN, BBC or CNBC, you want to make sure that it adheres to a certain policy. With GenAI, it is not only fast and productive, but it also complies automatically and is therefore accepted. It is always journalistically high-quality and advertisers need not fully understand the policy. They just know it is going to work.

Do you see potential for creating ads that exceed traditional ads in terms of effectiveness, clicks, purchases and so on?

I do. I think the models will become a commodity which anyone can use. The main differentiation will come from what I know and you don’t. The more you use conversion data and information about campaigns that worked for certain businesses or advertisers with real users, the bigger your advantage. It will take years for others to catch up. In a world of GenAI, scale is a real differentiation because you need data that no one else has. If we both prompt the same thing, having better information gives me a performance advantage over you.

How do you think Taboola can compete with Google and other tech giants? Various players in the market have the skill set, the deep pockets and different types of data to leverage GenAI. What will be the competitive edge of Taboola?

Apart from machines, the human advantage we have lies in our deep relationships with our B2B partners in the publishing industry. Unlike Facebook or Amazon, we work in the open web and allow advertisers to diversify their channels.

You give Facebook your name, your age, your gender and your hobbies, but you will never show on Facebook or Twitter something that you’re concerned about or something related to your kids or a medical case. However, you will read about this all the time and you are curious. It could be news, travel, a passion, or your hobbies. You watch videos about it because it's fun or of concern. It is this curiosity graph that Taboola has access to. Without knowing their names, we understand the interests of people very well. This is the data we prompt back into our engines to say, well people who do this, what else do they like? And what creatives entice them to give it a chance and maybe even convert? Google has search data, and social networks have social data. Taboola knows a version of you that is perhaps closer to who you really are in the real world.

Right now, you are helping people create advertisements with GenAI. What else can be done with similar technologies?

A few pieces are still missing. There is potential in AI-powered landing pages. Right now, the modeling is deterministic in the sense that all users land on the same page. But people react differently to different forms of landing pages. One person might be very visual and love big images but hate videos. Somebody else prefers text and just wants to read and move on. The same content can be created very differently on a landing page, and time of the day, device, screen size etc. could also be considered. Creating individualized landing pages could be an opportunity to increase conversions, which today, at least, we don’t offer.

As technologies evolve, very different services will be possible. Where do you expect to see the most traction beyond content creation?

We do not yet have what I call a buddy or a strategic advisor powered by GenAI. Imagine having an account manager who is awake 24 hours a day to help you with your own strategy and with creating the right targeting, objectives etc. That's an opportunity because there is also an emotional element in doing business. When you have the perfect account manager, you can feel relaxed and build trust. Even if you fail, you trust someone you know. This emotional element is currently missing, but I think it can be created.

Do you think machines can go the full distance in terms of consulting and customer success? As technology becomes more complex, might we need more human explanations and more assisted services?

We live in a very complicated world. One of the most exciting things about GenAI is the opportunity to simplify complicated tasks into something people can master without fully understanding how they are being done in the backend. If you have a flower shop and you want to get people to your store, implementing a campaign on the open web would be truly complicated. If you contact an agency, they will ask you weird question about things you might not even know existed. There would be so many steps and you just want to sell flowers. Most clients do not fail because their flowers are not beautiful but because it is too complicated to reach customers. Too many services, too many account managers, too many dashboards, and too many terminologies to get familiar with. I think GenAI will simplify that. You will need fewer services. My idea is that you just define your business objective and maybe provide some unique data. Then with GenAI, a variety of touchpoints can be created automatically.

If everybody ends up using the same tools optimized for the same objectives, don’t we risk that marketing communication may become highly homogenous and interchangeable? Do you see a risk that GenAI might erode differentiation in marketing communication?

It is a risk if you do not provide something unique about you. There are a variety of companies offering different models for consumer and business use, including OpenAI. Using them will be as easy, almost like making coffee. It will be affordable and available to anyone. The main differentiation will be what you put into the system, something that makes you unique. GenAI will mainly be different based on the data you prompt. If you don’t have unique data, you will become obsolete and irrelevant.

What you know about customers is one side of the training data. The other side is the creative content needed to train GenAI models. How can we make sure that people keep writing and producing images and videos when everything is available through GenAI?

I think the AI environment is going to create many new jobs and that GenAI won’t replace editorial content. In the world we live in, it is required more than ever. We appreciate an editorial team that cares about what they write, a publication that has writers who spend their time learning and educating themselves on a topic. You want to believe their good intentions and follow their thoughts and ideas. Compare that to the world of TikTok. It's entertaining and can spark curiosity. But God forbid that our children end up making important life decisions about science, politics or health care based exclusively on such social networks. I still believe that humans should select and create editorial content.

What about advertising? If you can create effective well-performing ads with the click of a button, why should humans bother producing novel ad content themselves?

The way I imagine it, creative content will be created by humans and then supplemented by GenAI to make them better or faster. It will increase productivity and maybe even effectiveness. It might also help humans develop really new ideas. I think on the front end, in terms of titles and images, humans do not necessarily have an advantage, but when it comes to full communication strategies or more complex content such as videos, humans are not as easily replaced.

Do you think consumers will continue to value a human touch in advertising?

Human relationships and authenticity matter. Consumers, and especially younger generations, want to have an emotional connection with the brand or product they buy. We see it all the time with direct-to-consumer brands. For instance, people want to see the founder's story. Mid-funnel direct consumer advertising is a huge portion of the advertising market and includes subscription services that create relationships. If you want to buy a bed, you want a furniture brand you can relate to, not just a good bed.

Advertising involves a whole ecosystem of creative agencies, market researchers and marketing professionals. How might GenAI alter these collaborations? Will we see more in-house content creation or perhaps a greater need for partners that monitor and leverage technology in the best possible way?

Smaller businesses won’t necessarily hire an agency because they can’t afford it. Google, Facebook and companies of that scale have about 10 million advertisers, and probably 70 percent or more use self-service. Clients will use GenAI almost automatically without even realizing that they do. It will be rare to find someone who does not use it. Then we have the segment of bigger brands and bigger enterprise accounts, and they will still use agencies and marketers for multiple reasons. They need a creative strategy, trials to lead the strategy and gut feelings, and they will use more touchpoints.

There has been much debate about GenAI and the business model of creative agencies. Is GenAI more of an evolution or a revolution for agencies?

My guess is that there will be both an evolution and a revolution on the agency side. There will be those who embrace technology to get better and those who reject it and may become obsolete. We will see more innovation, more jobs and more funding. It's all about embracing change. And the change will not be driven by technology alone, but by culture. The biggest innovation to enable GenAI and AI is culture. You need to embrace it and feel comfortable being uncomfortable.

A very nice closing statement. Those with the right culture stand much to gain, be it an agency, a marketer or a publisher. Thanks, Adam, for taking the time. It has been fabulous talking to you.