9. 5. 20249. 5. 2024
Like consumers across the board, viewers of connected TV (CTV) increasingly expect deeper personalization in the ads they see. In fact, recent studies have found that audiences pay nearly four times more attention to CTV ads that are relevant to the content they’re watching.

One way brands and publishers can achieve the customization viewers are looking for is through contextual targeting. While contextual targeting on CTV isn’t a new concept, it could be considerably better. Currently, contextual targeting on CTV is overly broad, and inconsistent data interpretation, fragmentation and lack of standardization make it hard to achieve at scale.

The good news is that positive change in contextual targeting on CTV is coming thanks to generative AI. This technology is allowing CTV advertisers to target ads based on emotion, and as a result to better align content, its context and consumers (the three C’s of contextual advertising) to boost performance.

Let’s look at what makes contextual advertising on CTV challenging, how Gen AI is improving the space and why it matters.

Hurdles with contextual advertising on CTV

Contextual advertising is challenging for advertisers and publishers due to a lack of categorization standards and limited granularity in contextual analysis. Most media buys and bids on CTV rely on the metadata publishers use to categorize their content by genre. The problem, though, is that there is a lack of industry standards around this metadata, and a resulting inconsistency and opacity with content labels. For example, one publisher may call a given show a comedy while another may label it a drama. This makes it difficult to ensure brand alignment and safety, and to execute buys at scale.

What’s more, most contextual methodologies only focus on content at the level of genre or show—that’s a legacy of how targeting was done with linear broadcasts. But genre- or show-level targeting can fail to take in the subtle nuances of TV content and how different scenes resonate with viewers.

Let’s go back to the earlier example. “Dramedy” could be understood as a subgenre of comedy, and in turn be labeled “comedy.” But only focusing on that generic label could lead to a cheerful ad (that would align well with a standard comedy) being paired with a downbeat, dramatic scene, making for a negative viewing experience and a missed opportunity to generate positive attention.

The Gen AI influence

To drive greater performance and to better ensure brand safety, brands and publishers today are exploring the ability to design contextual advertising around the emotions that content generates, not just the genre or show to which it belongs. When brands successfully understand and target ads based on their audiences’ emotions, they achieve the long-touted promise of digital advertising: showing the right ad to the right audience at the right time.

Achieving that level of granularity with contextual targeting requires analyzing multiple datasets, metadata and what’s happening on the screen itself, scene by scene. That’s where Gen AI comes in. Multimodal language models can be trained on hundreds of thousands of pieces of content, processing an amount of information and achieving a level of granularity that used to be impossible.

Subsequently, advertisers can use these models to identify the kinds of emotions that a scene of content generates and make a recommendation for the type of ad creative that will emotionally resonate with viewers. And these models can make such determinations across multiple channels and forms of programming, all in real time.

The result of greater alignment among content, ads and audiences will be a far better CTV ad experience for viewers, which ultimately benefits advertisers and publishers via better performance and more revenue.

Why it matters

As third-party cookies and deterministic identifiers go away, brands and publishers are looking to CTV to reach new viewers and deliver them customized experiences. Contextual advertising is one such method to do so, which allows for customization without individual data collection or dependence on outdated identifiers.

Going forward, brands will need ways to calibrate the impact of advertising, and attention will be a prime metric that doesn’t depend on persistent identity tracking. Contextual CTV ads will be especially effective at generating attention due to both the device (a large screen) and their ability to capitalize on audience emotion. In fact, TV ads have been found to be the ad format to which U.S. adults pay the most attention, according to CivicScience.

Overall, contextual CTV ads represent a more equitable value exchange for the industry, one where audiences get better ads and respect for their privacy, and advertisers and publishers get the business results they expect.

Loading more ...