Streaming has shifted the way advertisers buy TV inventory. Rather than duking it out for key positions in the hottest shows, streaming ads are typically disseminated across platforms in pursuit of smaller sets of viewers that possess certain traits or behaviors. In some ways, the tech-ification of TV has separated its commercial breaks from what makes it a popular medium—the emotionally resonant content consumers turn their screens on to watch.
Contextual ad targeting—5 things marketers need to know about streaming’s latest obsession
While buying inventory for specific shows is still reserved for highly-priced sponsorships or custom content, platforms have begun remedying the disconnect by rolling out new contextual targeting tools. These capabilities, involving an AI-powered process of scouring every second of content available on a streamer, allow marketers to not just target people, but moments in which their ads are more likely to hit the right emotional note, or match the action of a scene.
Read more: Media measurement—tracking the latest updates
Given the smattering of new products being pitched to the market, what is contextual targeting? How does it work in streaming? What does it mean for brands? Ad Age has gathered answers to some of the most commonly asked questions about streaming’s hottest capability.
What is contextual targeting?
Contextual targeting is a way for advertisers to match their ad with the right content to, at a base level, ensure mismatched ads aren’t a jarring interruption during content viewing. More broadly, it can pair brands with content that has primed a consumer’s brain for, say, ordering dinner delivery during a mouth-watering cooking show, or in more subtle cases, place ads in content proven to show higher returns on specific goals, such as app downloads or interactive engagement.
This type of matching show with ad has long been manually possible in TV. Cable TV channels, which typically cater to specific interests like food, travel and home improvement or content genres such as drama, comedy and kids, have been high-level guides for where to place ads for specific categories or how to curate different creative for different channels.
Streaming and the development of AI have made this process not only automated, but more exact in placement.
For example, Chipotle recently used Disney’s contextual ad targeting tool Magic Words for its ads during college football games. Rather than tying the spots to specific ad pods, AI analyzed the action in real-time to serve Chipotle’s ads specifically after game-changing moments, leading to double-digit increases in ad perception over regularly placed ads, according to Disney.
“The compelling part, especially coming from a media agency, is the fact that we’re not having to make a change to the creative—that’s not something that we always necessarily control,” said Kara Manatt, executive VP of intelligence solutions at Magna, who also noted it’s beneficial for brands to not have to change or generate additional creative to utilize contextual targeting tools. “To see such a difference in performance without changing creative and being smart about where it runs is the compelling part for us.”
What is AI’s role in contextual targeting?
AI does the grunt work of cataloging content to make contextual targeting happen. Using AI, streamers automate a process that combs through every frame of every show and film in their libraries, creating tags for the action and emotional context of each one.
Some media companies, such as Disney, have created this automation in-house, while others such as Warner Bros. Discovery, partnered with third-party AI companies such as Kerv AI.
At the base level, streamers are able to create a collection of targetable segments based on emotions or in-scene actions for advertisers to target. Disney announced during its Tech and Data Showcase at CES that it built the ability for its AI cataloging to be done live, allowing for real-time contextual targeting in sports and live events.
However, the immense record of frame-level data generated by this AI process has enabled more than one utility for streamers and advertisers. For example, WBD’s Max uses Kerv’s AI metadata to log the products in each scene as well to enable shoppable advertising that can generate collections seen on screen or similar to products being shown. Max also uses the feature to provide more granular brand safety controls, allowing advertisers that may have blocked TV-MA shows or R-rated films from their buys to run ads in breaks timed near appropriate scenes in such programming.
Which streamers are leaning into contextual targeting?
Multiple streaming platforms have built or partnered on capabilities to enable contextual targeting.
In 2024, Disney announced the launch of Magic Words at its Tech and Data Showcase during CES. Using AI-logged metadata, Magic Words matches ad creative with the appropriate mood or scene content. During this year’s CES, Disney announced an expansion of Magic Words to include live sports and events. The addition does the work of contextual targeting in real time.
In November, Warner Bros. Discovery rolled out a similar contextual targeting tool for Max called Moments. The capability launched with 40 predefined contextual targets for advertisers to target scenes related to cooking, real estate, gaming and science.
During its One24 showcase last March, NBCUniversal launched 300 contextual segments for advertisers to target. For example, for an advertiser targeting moments connected to family values, NBCU can use AI to draw connections across unexpected combinations of the “Real Housewives” franchise, “Law and Order” and a WWE match.
At CES 2025, NBCU also announced a tool that does contextual targeting in reverse by grading creative on its cohesiveness with the content it runs during. The AI-powered capability will grade spots on how well they match target shows or movies based on campaign goals, and provide feedback to marketers on what elements are lacking.
“Everyone has realized it’s not that hard right now to find an audience, but to actually engage that audience is getting harder and harder,” Mark Marshall, chairman of global advertising and partnerships at NBCU, previously told Ad Age. “So when you find them, you sure as hell better have the right creative that’s going to knock it down.”
Roku offers similar genre and show category targeting through The Roku Exchange, its platform for ad buys across its owned and partner inventory. And platforms such as Netflix offer a cable-like option to advertise in certain content genres.
Why are brands using contextual targeting?
Contextual targeting provides a number of unique uses for brands. At a base level, contextual targeting can give a boost to brand metrics by flowing content into complementary ads, and minimizing the risk of angering viewers by jump-scaring them with a bright and cheery ad during somber or tense sequences.
Plus, given the contraction of cookies and increased scrutiny of consumer data, contextual targeting is an additional tool for marketers to utilize in reaching new audiences outside their networks.
Related news: 20 ad tech executives predict 2025 trends
And while streaming ad buys often run across an entire platform rather than selecting individual shows like in linear, contextual targeting allows marketers to get closer to the streaming content they want. In this way, contextual targeting can also provide new information on what content ads are running during and the audiences reached within it as transparency from streamers remains a top need from ad buyers to scale investment.
Does contextual targeting work?
Streamers’ AI-powered contextual targeting tools are still in what multiple media buyers called test phases, and the full potential of them is a work in progress.
More from Ad Age: Advertising’s identity-based targeting crisis
Early results from Magna’s tests with Disney, which included clients Chipotle and T-Mobile, found Magic Words was able to increase certain brand metrics by curating specific matches such as a food-themed ad with tailgate scenes or by identifying exciting moments of a match to serve an ad when viewer attention is higher, according to Magna’s Manatt.
In its tests, the agency found that a game’s key moments, whether they favor the team a viewer is rooting for or against, increase possible ad attention by up to 2.6 seconds with peaks 15 and 20 seconds after the moment. Magic Words was able to target these moments as they happened, raising ad recall by 6 points, favorability by 4 points and search intent by two points more than ads not utilizing Magic Words.
But one media buyer said the tools being rolled out by media companies broadly remain a plaything with potential rather than a substantial capability a marketer can build an entire strategy around in their current form. As such, the price of using contextual advertising is unclear as it’s initially implemented into buys for testing, although the buyer noted that they didn’t anticipate it would add significant cost to inventory.
While the buyer expects the effectiveness of contextual ad targeting tools to increase as media companies build them out and the tools mature, they also see it as a tactic to make advertisers more comfortable with using AI in smaller aspects of their media strategies. As the utility of AI becomes more ingrained in the industry, the buyer expects it to fuel more capabilities, with TV ad sellers ready to ease advertisers into new ways of activating in TV.