Gracenote’s Bill Michels: Scene-Level Targeting Only Works If You Can Do It At Scale

Why It Matters

If metadata and IDs reach sufficient scale and standardization, brands can better control placement relevance and performance on CTV, unlocking larger, more efficient ad spend beyond walled gardens. This shift could reshape planning, measurement and inventory value across streaming ecosystems.

Summary

Gracenote says its long-standing content metadata and universal IDs are enabling advertisers to move beyond app- or IP-level CTV buys toward content-aware buying at scale. Advances in standardized IDs and richer metadata — including scene-level signals — are increasing the ability to target and block inventory by genre, series attributes or in-show moments, though scene targeting only works commercially if it can be standardized and scaled. Contextual signals are also evolving from niche descriptors into real-time optimization inputs that can drive campaign performance on CTV. The net effect is a partial return to content-driven ad buying reminiscent of linear TV, but powered by programmatic scale and standardized taxonomies.

Original Description

LAS VEGAS — Targeting ads to specific scenes within TV content sounds like the ultimate precision play. The challenge is that precision without scale doesn't spend budgets, and most scene types don't appear frequently enough to matter.

“Someone’s in a shower scene, going to sell them soap. How many shower scenes are there? There’s not that many,” Bill Michels, chief product officer of Gracenote, told Beet.TV Editorial Director Lisa Granatstein at CES. “We need things that work at scale and we also need standardized taxonomies against them.”

Gracenote is focusing on within-scene targeting that operates through standardized approaches allowing buyers to spend budgets at scale rather than limiting purchases to experimental trial levels.

Metadata powers brand-content matching

Artificial intelligence helps match brand intent with content. Buyers use conversational interfaces like ChatGPT and Gemini to pursue goals ranging from broad reach to specific audience targeting to precise content placement.

“What we see for the AI and experiences it is providing, it’s helping match the brand or the buyer’s intent with the actual content itself,” Michels said.

This enables optimization in real-time and pre-campaign reporting by coupling data with desired brand outcomes.

TMS ID enables transactions

Gracenote’s TMS ID serves as the canonical reference for content identification across major CTV platforms, enabling systems to transact and unlock data. The unique identifiers map TV programs and movies with rich metadata that augments targeting quality.

“The ability for different systems to transact on that and unlock the data associated with it becomes a critical piece,” Michels said.

CTV buying historically relied on IP addresses or app-level data, but increasing scale now enables specific content targeting that helps brands identify optimal message placement.

Program-level visibility returns

Connected TV is migrating back toward linear television’s content-based buying approach, though focusing on content type, genre, series, and live status attributes rather than specific shows.

“If we go back to the linear days, that’s how ads were always bought. It was just based on the content. I bought The Cosby Show, you bought Monday Night Football,” Michels said. “We’re seeing more attributes come in to help articulate and get more specific about that buy.”

Brands need reach and storytelling at scale, requiring standard taxonomies and sufficient inventory for meaningful spending rather than small trials.

Contextual as optimization signal

Contextual relevance is evolving in two impactful areas: serving as a performance optimization signal and achieving scale for meaningful campaign spend.

“Traditionally ads, especially in walled gardens, are optimized in real time on purchases, app installs. It’s tied to user data,” Michels said. “I expect we’ll see on CTV that the contextual data is the signal. What is working, what is not, what is giving this campaign the performance I need.”

Brands previously couldn’t allocate sufficient CTV budgets to contextual targeting due to limited ID scale, but standardized IDs now enable revisiting these approaches with meaningful campaign amounts.

ID availability expanding

The number of IDs enabling content understanding will increase significantly beyond app-level or channel data that historically limited CTV targeting and blocking capabilities.

“We’ve really not had that in CTV,” Michels said. “We’ve had app level, maybe we’ve had a channel. But we’re really going to start to see the availability of IDs to help target and sort of block against where you want your spend to go.”