No one’s interested in guesswork.
The recent surge in AI tools means that it’s easier than ever for advertisers to generate multiple versions of an ad for different audiences, but increased versioning is moot if the right ads aren’t reaching the right people. The challenge, of course, is figuring out the best ad for a given person.
CTV ad platform Tatari believes that accurate targeting relies on a wide scope of data and inventory, as well as machine learning that’s constantly training on the latest data.
On Thursday, it announced the launch of Tatari Max, an optimization engine that sits at the ad server level to help brands and publishers decide which TV ads to serve to what audience.
Always learning
Although Tatari Max uses AI optimization to determine the most relevant ad placements for a client, Tatari CEO and Co-Founder Philip Inghelbrecht is adamant that the product is a far cry from agentic AI.
While agentic AI relies on a series of “if, then” rules and automates existing processes, machine learning (which is what Tatari Max relies on) evolves over time as new data emerges, Inghelbrecht said.
By looking at signals, including device IDs, shopping behavior and demographic information, Tatari Max can determine which people – and types of people – respond best to specific ads.
It also calculates what it calls an “affinity score,” which measures “the extent to which an impression is best fit for a particular brand,” based on previous viewers’ engagement, said Inghelbrecht.
Activewear brand Rhoback, an early tester of Tatari Max, hopes the self-learning element of the tool will help it determine which creatives are most relevant to which viewers, said Caroline Nemceff, senior digital marketing analyst at Rhoback.
Rhoback was initially founded as a men’s clothing brand, but this year it’s been focused on expanding its women’s line as well as its collegiate sportswear. Since those sects of the brand are newer, Rhoback has less data on how potential customers will respond to the ads. TV mainly serves as an “awareness play” for the brand to find new customers, said Nemceff, and it’s optimistic that Tatari Max will help it “learn who the correct user is” for a given piece of creative.
It’s too soon to see demographic results on exactly what kinds of people are responding to which creatives, Nemceff said, but early results are promising.
Over a one-month period of testing an ad on Hulu marketing a new women’s activewear collection, Rhoback saw a 32% increase in purchases and a 10% increase in site visits.
Soon after implementing Tatari Max, Rhoback had to pull back on streaming spend and pause the tests (due to external factors) – but, thanks to strong performance results from that early testing, said Nemceff, it plans to return to using Tatari Max later this year. “We are planning to test Max again on ESPN using our collegiate creative, and expand into Hulu during [Black Friday and Cyber Monday] with our sale creative,” she added.
“It’s only going to get better,” she said, “because you’re feeding it more data and more data points.”
You can see everything
Tatari Max is built on Upstream, Tatari’s supply-side stack that integrates with publishers at the ad server level.
Some similar products, like Meta’s Advantage Plus and Google’s Performance Max, only connect to the company’s own inventory. Upstream, on the other hand, plugs into several premium TV publishers, including Paramount, Peacock and Disney, so clients can see all accessible inventory across the TV world.
Plus, Inghelbrecht said, Upstream isn’t just limited to programmatic inventory. Since it sits at the ad server level, it can see data on inventory that’s only available as direct buys as well. Since the AI is receiving so much data, it has more to learn from and can improve rapidly.
The wealth of data is a draw for captioning and live transcription app Rylo, which started working with Tatari about two months ago and began testing Tatari Max at the same time.
Tatari has been more effective for Rylo than other CTV ad platforms, said Annica Lin, the app’s VP of growth, adding that since adopting Tatari’s tech, she’s seen an increase in customers saying that they discovered the brand on TV.
Spreading the wealth (of ads)
Most of Rylo’s users are older individuals, since the app is specifically aimed at people who are deaf or hard of hearing.
The problem with targeting older folks directly, though, is that many of them “are not tech-savvy,” said Lin. Tatari Max, which was doing household targeting, found that the people who actually respond best to Rylo’s ads are often younger household members, like the older individuals’ adult children or caretakers.
“They are the people who actually [are] influenced and then help their parents to download the app,” said Lin.
Tatari Max was able to determine that the most effective ads among that younger demographic were testimonials, in which hard of hearing individuals describe their experience with Rylo. Personal testimony emphasizes the resources that are available to people who are hard of hearing, and often leads to these other household viewers helping older relatives download the app.
Now, said Lin, Rylo thinks more about “how to communicate with the caretaker or the kids,” rather than exclusively targeting potential app users themselves.
