SAN JUAN, PUERTO RICO – At the Beet Retreat San Juan, Suvadip Choudhury, head of television partnerships at Alliant, delivered a reality check to the TV ad business. Yes, we have more data than ever. No, that does not mean we should blindly chase every shiny behavioral segment that appears in a dashboard.
“The TV industry has done a great job at evolving the way they think about custom audiences,” Choudhury said in this interview with Beet.TV contributor David Kaplan.
He said the shift from basic demographic targeting such as “moms ages 25 to 34” to far more behaviorally informed audiences.
Instead of broad strokes, marketers can now zero in on “moms who are likely to spend on childcare” or “college educated households that are looking to buy a new house.”
With richer signals, he said, “we now have the ability to hone in on the exact things that an advertiser, an agency, or brand might want for targeting their consumers.”
But more signal also means more responsibility. Choudhury urged buyers and sellers to question the raw material behind those segments.
“Where does the seed data come from? How did you collect it? What does it truly signify?” he said, adding that these are the questions required to execute “highly sophisticated audience strategies in the future.”
Translation: not all data is created equal.
You are more than your floss purchases
Kaplan suggested that Alliant sits at the crossroads of transactions and prediction. Choudhury agreed, then cited an example that may haunt dental aisles everywhere.
“Transactional behavior has been the gold standard for E-commerce marketing,” he said.
If you buy floss every month, marketers know you floss. Congratulations.
But, he added, “there are multitudes to everyone.”
Just because someone buys a lot of floss does not mean that floss defines their entire identity.
“The frequency at which I buy tooth floss isn’t the only thing about me,” he said with a smile.
For marketers, the lesson is clear. Transaction data shows what happened. Predictive insight helps explain why and what might happen next.
“The context of why they’re doing the things that they do matter,” he said, because that broader understanding leads to better predictions and smarter audience expansion.
Balancing scale and precision in CTV
When it comes to connected television, the eternal tension between scale and precision still looms. Choudhury framed it as a business decision.
“Balancing scale and precision really comes down to how an advertiser would like their ad dollars to be used,” he said.
If a brand wants a very specific consumer based on real behavior, precision wins. If there is enough transaction data, that precision can sometimes scale on its own. But often it cannot.
That is where predictive modeling steps in. It helps “fill the gap and size up,” he said, allowing marketers to take a precise audience and reach it at meaningful volume.
After all, “there are more to users purchase beyond just the fact that they’re purchasing,” he said, emphasizing that behavior is layered and fluid.
When prediction beats the past
Prediction shines when past behavior is thin, unstable or unreliable.
“Prediction outperforms past behavior when the data that describes the past behavior is unstable or unreliable, or potentially missing,” Choudhury said.
He pointed to home buying as a classic example. Most people do not purchase homes every year, so there is not a mountain of transaction data to mine. In those cases, predictive insight and a broader understanding of the consumer become essential.
Still, he was quick to defend the value of hard transaction data. If a product has high repeat purchases and strong signals, “you will have really great results with transactional data too,” he said.
The message was not either or. It was about knowing when each tool makes sense.
Context still counts
Finally, Kaplan posed the existential TV buying question: Is the future about content adjacency or real world intent signals?
Choudhury resisted the urge to pick a side.
“I think the future of audience buying will continue to have a mix of content adjacency buying as well as audience-driven buying,” he said.
Context can still matter. If you want NHL fans, buying around an NHL game may be perfectly logical. But as viewing fragments across devices, platforms and environments, audience-based strategies become more powerful.
“There definitely still is a place for both,” he said, adding that advertisers need to be more intuitive about what they are trying to accomplish before choosing a path.
In other words, buy the hockey game if you want hockey fans. But if you are chasing life stage shifts, big purchases or changing habits, you may need more than a rink side seat and a floss receipt.
You’re watching coverage from Beet Retreat San Juan 2026, presented by Alliant and TransUnion. For more videos from this series, please visit this page.
