Contextual targeting is built on a simple premise: match the ad to the content. Know what's playing, serve the right creative, reach the right viewer. In digital display and online video, that model works because the content is identifiable. In CTV, the premise breaks down before it starts.

The problem is not with contextual targeting as a strategy. The problem is that CTV bid requests usually do not contain the content signal that contextual targeting requires.

What Contextual Targeting Actually Needs to Work

Contextual targeting depends on content metadata: genre, topic, program title, rating, content descriptors. In a web environment, that information is typically available at the URL, page, or article level. The ad tech stack reads the page, classifies the content, and matches the appropriate campaign parameters before the bid fires.

In CTV, the equivalent of that content layer is program-level data. Not the app. Not the network. The specific program, genre, rating, and content context of the stream where the impression will run.

That program-level data is almost never in the bid request.

The Reality of CTV Bid Requests

Peer39 analyzes more than 2.5 billion CTV bid requests daily. A significant share of those requests arrive with nothing more than a device ID, an app bundle, and a timestamp. No program title. No genre. No content rating. No description of what is actually playing.

When a contextual targeting system receives a bid request like that, it has nothing to classify. The match cannot happen because the input does not exist. The system either passes on the impression, falls back to audience signals, or applies a category from the app-level metadata alone. None of those outcomes is contextual targeting in any meaningful sense.

Contextual targeting in CTV is only as good as the signal layer underneath it. Without program-level data, there is no context. There is only an app name.

App-Level vs. Program-Level: Why the Distinction Matters

App-level metadata tells you the publisher channel. It might tell you the app is a sports streaming service or a general entertainment platform. That is useful for exclusion purposes but insufficient for contextual matching. A single streaming app can carry children's cartoons, true crime documentaries, and live sports within the same session.

Program-level data tells you what is actually playing at impression time: the specific title, the genre (and sub-genre), the content rating, whether it is live or on-demand, the presence of specific content advisories, and in some cases, the sports teams or topics featured. That is the signal layer contextual targeting requires.

The gap between what app-level can tell you and what program-level can tell you is not a minor refinement. It is the difference between targeting 'streaming' and targeting a specific type of content environment with verified signal.

Where Program-Level Reach Gets Lost

When contextual CTV campaigns underdeliver against reach goals, the cause is almost always signal gap, not inventory gap. The inventory is there. The addressable audience is there. But the program-level signal needed to classify and bid on that inventory is missing across a substantial share of available impressions.

The result is a coverage problem masquerading as a targeting problem. Buyers see limited delivery. They loosen brand safety parameters or broaden category targets to compensate. In doing so, they buy against unverified inventory while the verified, well-labeled inventory they actually wanted remains untargeted.

Program-level authentication solves this by filling the signal layer independently. Rather than relying on publisher-declared metadata in the bid request, Peer39 authenticates program-level signals at scale and applies them pre-bid across more than 2,100 categories. The reach is there. The signal just needs to be brought to it.

What Authenticated Program-Level Data Enables

With authenticated program-level signals available pre-bid, contextual CTV targeting can actually function as designed:

• Genre and sub-genre targeting at the program level, not the app level
• Content rating alignment, including advisories and parental guidance categories
• Live vs. on-demand differentiation, which affects both context and viewer attention
• Keyword-based content exclusions that operate on verified program data, not declared metadata
• Post-campaign reporting that maps directly to the pre-bid categories, so what you targeted is what you measure

That last point is not a minor convenience. Campaigns that target on unverified signals and report on different categories are not measuring their own targeting. They are measuring something else and calling it performance.

The Takeaway

Contextual targeting in CTV does not fail because the strategy is wrong. It fails because the signal layer most CTV environments provide cannot support it. App-level metadata is not context. It is a starting point.

Program-level reach gaps close when the signal layer closes. Buyers who want contextual CTV to perform the way it performs in digital display need to demand the same level of content signal from their CTV inventory that digital display takes for granted.

Without that, contextual CTV targeting is applying a precise strategy to an imprecise input. The output will always disappoint.