Why CTV buying looks neat on a plan and messy on screen

New Delhi: Connected TV advertising has no shortage of data. What it lacks is clarity. Advertisers can see reach numbers, impression counts, and completion rates, but often have little insight into what was actually on screen when an ad appeared. 

In a market where content choice is exploding and attention is increasingly fragile, that missing layer is becoming impossible to ignore.

That layer is metadata.

Metadata has long powered content discovery. What is changing now is that the same data is being prepared to define how CTV media can be bought, sold, and measured.

Discovery is the pressure point

The urgency behind this shift is coming from the viewer's side. According to the Gracenote 2025 State of Play report, nearly half of streaming viewers globally say it has become harder to find something to watch. 

Viewers now spend about 14 minutes searching before choosing content, and many say poor discovery could lead them to cancel a service.

Phanimohan Kalagara, CTO of Gracenote described discovery as the biggest unsolved problem in the ecosystem. “The biggest problem that we're solving is the challenge of discovery of content,” he said. “Search actually is such a critical KPI, not only from a monetising angle, but from literally like for every platform.”

What makes discovery relevant to advertising is that it reveals intent. What people search for, how they navigate content, and what holds their attention are signals that go beyond demographics. These signals are now being captured in metadata.

“I think what we get out of metadata, for example with any media associated with it, is the characteristics of the media that it is playing,” Kalagara said.

The question advertisers are now asking

Today, most CTV buying still relies on broad signals, such as the app, channel, or genre, as stand-ins for context. While useful at a high level, these signals do not describe what is actually happening on screen at the moment an ad appears.

Advertisers are increasingly aware of this limitation. Brand safety concerns have expanded into broader questions about suitability, tone, and relevance.

From an advertiser’s perspective, Kalagara said there are two key concerns. “One is viewership, who is watching, how many are watching, and what is my ROI,” he said, noting that audience measurement remains well established. “Our parent company Nielsen continues to work in that area.”

The second concern, he added, is increasingly critical: “Which content am I associating my advertisement with?”

Metadata is what makes that second question answerable. It describes content using attributes such as mood, topics, themes, and scene-level changes.

“If you consider the mood of the content and what is actually playing out, metadata can provide the fidelity needed to measure how effective an advertisement will be,” Kalagara said.

Metadata captures shifts in tone and topic within content, not just labels at the programme level. “There can be shifts in mood within a show or film,” he said. “From topics to moods, metadata provides advertisers with rich context that can be extremely valuable.”

From impressions to environments

The significance of metadata is not limited to better targeting ideas. It changes how CTV inventory itself can be defined.

Traditionally, inventory has been packaged by volume, impressions within an app or programme. Metadata introduces the possibility of packaging inventory by context: moments of high emotional intensity, specific themes, or suitable environments for particular brand messages.

“From mood to object and brand detection, metadata enhances the precision of ad placement,” Kalagara said.

This does not mean advertisers are already buying media this way at scale. Rather, metadata makes it technically possible for platforms and sellers to start offering inventory based on context rather than just placement.

Metadata also enables exclusion, allowing advertisers to avoid environments that do not align with brand values. “Context can indicate, for example, that certain content may not be appropriate for specific advertisements,” Kalagara said.

Where contextual buying will actually scale

One of the clearest signs of where the market is heading is the movement of metadata into programmatic CTV infrastructure.

Gracenote is partnering with major programmatic platforms to enable content-based advertising on CTV. It is live with Index Exchange and Magnite, with more DSPs and SSPs expected to come online over this year. The current focus is on enabling metadata for targeting, reporting, and measurement, rather than fully mature contextual buying strategies.

This preparation matters because it places content intelligence inside the systems where CTV transactions already occur.

By standardising how content is identified and described, metadata creates the conditions for contextual buying to scale when advertisers are ready.

Fragmentation makes context non-negotiable

The TV landscape is becoming more fragmented, not less. FAST channels are growing, streaming services are multiplying, and discovery is getting harder even as total viewing time stays relatively stable.

In this environment, advertisers are under pressure to make smarter decisions with fewer assumptions. Metadata offers a way to bring structure to CTV buying and selling by defining inventory through context, not just scale.

Metadata may not yet define how most CTV media is bought and sold, but the foundations are now being laid. As discovery data moves deeper into advertising infrastructure, it is setting the stage for a future where context is no longer an afterthought in the CTV ecosystem.