From agency holdcos to Big Tech firms, the pivot towards AI has resulted in layoffs across the industry. But rather than AI replacing roles at these businesses, Tal Jacobson, CEO and Director of Perion, says these layoffs are “nothing more” than cost-cutting measures carried out in order to invest in AI infrastructure, as currently the tech “can’t fully replace people at companies that are scaled.”
In this interview, Jacobson discusses the ongoing impact of AI, Perion’s evolving ad tech offering, and the Taylor Swift effect at last year’s Super Bowl.
What types of ad tech products does Perion offer, and what does your customer base look like?
We provide technology to help advertisers optimise their spend across the major channels, such as CTV, DOOH, social, web, pretty much everything. And the unique thing about us is we provide one platform to access every single channel. So you can understand what your media investment actually does without going through 40 separate platforms.
We have clients in 40 countries across Europe, APAC, US, Canada and Latin America. We’ve always had a good, healthy mix of independent agencies, holding companies and brand direct customers. We haven’t seen any major shifts, but are maybe working with more independent agencies now.
What do advertisers take away from comparing their media investment across multiple channels?
I think as long as you’re working in siloed environments, it’s extremely hard to understand what would work better, especially if you have a business that is not fully online. If you have a physical store, or an airline, or a car dealership, then it’s extremely hard to understand what drove the customer to buy from you. That’s also the case with online businesses, but at least they can see conversions online. And most platforms provide different KPIs, so it’s hard to compare them. One might provide viewability, the other might provide CTRs, so you’re looking at oranges and apples.
With us, you provide the KPI based on what you’re trying to achieve; I want people to download my app, or I want people to come to my store and use a coupon. So let’s spread the budget across those channels, and then within a week or so, see what works better, and then we’ll give you a recommendation as to what you should do next. Maybe you should do less YouTube and more Meta, or whatever it is. But we don’t have one recommodation for everybody. It really depends on the specific campaign with a specific goal. Meta, YouTube and all these platforms are great for specific things, but they’re not all great for everything. So different KPIs would need different types of creative, different types of audiences, different types of inventory.
Perion’s AI recommendations are currently still reviewed by humans – will this continue to be the case or will AI be able to make those decisions independently?
Currently I think our AI recommendations are fairly good, since you’re going to base your investment decisions on that recommendation, and that’s real money. We need people to make sure that everything is extremely accurate. It’s not like ChatGPT that might say, “You know what, you’re right, I was wrong, here’s a different answer.” Once you’ve spent that money, you’ve spent that money. So I think for companies that advertise with us, spending real money, you always want to make sure that somebody’s overseeing what’s going on.
Maybe in a few years, when we feel extremely comfortable with how the analysis is going, we’re going to let it run wild. But to be honest, when you’re running on Meta or TikTok or YouTube, and using their advertising platforms, AI is doing that, and no human is actually looking at it. The difference is when you’re running across everything, you want to make sure that you’re still looking at things the right way.
The industry has seen significant layoffs as agencies and tech companies embrace AI – how would you characterise AI’s impact on jobs?
I hope that most people understand by now that AI is great, but it can’t fully replace people at companies that are scaled. I’m not talking about a startup with two people, it can absolutely do a job of 12 people. But for companies that are fully scaled, it’s harder to let AI replace people. You can actually do more with existing people, but just replacing them, that’s harder.
I think the reason why companies have massive layoffs is they need to invest money in the future. They need to invest money in infrastructure, in AI, for the future, not for now. But it doesn’t look good on a P&L. So you need to reduce cost, and that means reducing people. But it doesn’t mean that AI is going to do their job now, it’s just rebalancing the balance sheet. It’s nothing more than that.
Advertisers are seeing increasing opportunities in live sports on CTV – what kinds of formats do you offer for live CTV, and where are you seeing demand for live sports inventory?
The important part of live CTV is you don’t want to ruin the live experience itself, so we have a picture-in-picture format that allows us to keep the game on. We also have technology that identifies moments when we know people are not going to be upset about us claiming the screen.
Surprisingly last year at the Super Bowl, maybe for the first time, we saw a lot of demand from lifestyle and makeup brands. I think it happened because Taylor Swift was dating one of the Super Bowl players, which brought a lot of women to the game!
Perion also offers targeting technology for out-of-home advertising – what are the opportunities to use OOH in conjunction with CTV, and how do you measure those results?
The retargeting opportunities are huge. We can tell that people who usually go through certain tube stations probably work in nearby buildings, and then we can target specific ads around those buildings, and then use CTV to target areas where people live. We have anonymous data so we know that people that work in specific areas tend to live in those areas. So if they saw the OOH next to their office, that correlation means they would also see CTV in those locations.
We can also use audio to target Spotify ads in that specific area, so it’s pretty cool when you listen to music, you can hear an ad, and you can see that ad across the street as well. It feels very immersive. And if you also have a Costco or Walmart next to you, and they’re advertising a specific promotion, it feels very targeted. We can then see if there was an uplift in that specific promotion in those specific areas, and we can make that correlation. If we created that promotion for a barbecue grill, we can see how many people actually bought it in areas where we ran the ads versus without, and then we see a lift.
Follow VideoWeek on LinkedIn.
