Programmatic CTV’s next advantage runs on content

Content intelligence has arrived at scale for CTV buying, providing a rich new data source for brands and agencies that goes beyond the audience signal. It enables human-defined targeting and alignment with brand guidelines on CTV buys. But its greatest potential perhaps lies in serving as a powerful optimization signal for performance algorithms. 

To understand that evolution, we have to take a look back. When CTV entered programmatic, the pipes and data ecosystem were already well established and humming with inventory for web, mobile and even a little audio and digital out-of-home (DOOH). While CTV media itself was very new to the ecosystem — well-produced, professional, recognizable talent, live events — the bid request looked similar to web and mobile. That’s by design. Standards create confidence that specifications won’t change as systems scale, encouraging capital allocation. By that measure, real-time bidding has been fantastically successful, unlocking significant investment, channeling advertiser spend to media properties and fueling the growth of an entire industry. 

Because programmatic CTV ad buyers recognized the shape of the impression, they applied the same playbook — leveraging the IP address, attaching audience data and ZIP code-level location to the IP address, adding the bundle ID and applying decisioning logic. Sure, CTV introduced ad pods and a new alphabet soup of deal types. But the more fundamental change was what advertisers bought. As TV ad spend moved from linear to programmatic, packages of shows gave way to individual impressions tied to household-level audience data.

What was lost in the transition was content context: an understanding of what was playing when an ad appeared. At the time, the trade-off seemed reasonable, too. Programmatic had been built around audience-based decisioning, and that was the playbook advertisers knew. 

But ad buyers have not stopped asking for content data. That brings us to today.

Gracenote’s 2026 study of 500 U.S. media planners, TV buyers and programmatic traders quantifies the disconnect. Among programmatic traders, 91% said missing show-level data limits their CTV campaigns. Media planners were equally clear: 86% said they would move more linear TV budget to CTV if show-level targeting and reporting were available. IAB’s 2026 research also found that just 33% of buyers had high confidence in CTV inventory transparency. The CTV shift to programmatic scaled faster than the content intelligence needed to guide it. 

Now, content intelligence is bringing a core strength of linear TV into programmatic CTV: buying against what’s on screen. Persistent, standardized content IDs give titles a consistent identity across systems, while video language models can add scene-level detail. Advertisers can target by genre, talent and live sports — and, where supported, by a scene’s mood and tone. These capabilities are no longer theoretical. They’re here. 

But the game has not stood still. Walled gardens and performance-focused demand-side platforms (DSPs) have entered CTV in a big way, building incredibly efficient machine learning-based growth systems that combine real-time conversion events, audience context and campaign outcomes to guide decisioning on targeting, bidding, pacing and creative.   

IAB Tech Lab’s Agentic Real Time Framework (ARTF) is another optimization accelerant: dedicated containers are an exciting new path to run optimization models on programmatic impressions,  especially CTV. 

These new CTV buying systems ingest massive amounts of data and let the model decide what drives outcomes. Content intelligence will be a powerful new ingredient alongside audience data. As performance algorithms begin to use this data, they can discover patterns that ad planners would never have thought of. Human-defined targeting is finally here, which is great news. But model-defined optimization is where this new data source may have its greatest impact.