The Numbers Behind AI-Enabled Contextually Targeted Ads

A recent survey suggests viewers pay less attention, are less interested, and have a lower opinion of brands if they are advertised in an unsuitable context and are not relevant. AI-enabled contextually targeted ads can change that. 

The study -- Driving Viewer Attention and Brand Metrics in CTV Advertising -- shows 42% of the viewers paid more attention to an ad, learned more about the product, and were more interested in the products when served an AI-enabled contextually targeted ads, according to a study published by the Alliance for Video-Level Contextual Advertising and Tobii, which focused on eye tracking.

A national panel of 24 households that regularly watches ad-supported streaming TV was selected to watch more than 1,000 ad experiences. Participants wore eye-tracking glasses to watch controlled and tested content on their smart TV at home.



They were allowed to have their phones and any other items that they would usually have while watching TV. They watched for 90 minutes and then were surveyed.

Each watched three episodes of a popular sitcom rated TV-MA with the same sequence of episodes. The order of exposure to the control and test ads was random and distributed evenly across participants.

AI created the contextual intelligence that analyzed a video frame-by-frame to create standard and custom contextual segments and GARM brand-suitability segments for targeting and risk management have more impact compared with those that do not.

Contextual engines used AI, including computer vision and natural language, to understand the ability to watch, listen, and read video content frame-by-frame. The idea was to create contextual and GARM brand suitability segments that can be targeted by advertisers.

AI-enabled contextual targeting outperformed standard demo and generic pub-declared metadata when driving attention and brand KPIs in CTV.

AI-enabled contextual ads were seen 4X more, while ads in the demo and pub-declared groups were more likely to be missed entirely.

More ads were seen from beginning, so when the show transitioned into an advertising break, viewers were less likely to look away. Viewers of AI-enabled contextual targeting were the least distracted.

Panelists recalled twice as many brands from AI-enabled contextually targeted ads compared with what the study referred to as “pub-declared” and demo groups.

When given a list of brands to select, panelists recalled four times as many brands from AI-enabled contextually targeted ads compared with those pub-declared ads and 50% more than the demo group.

The demo targeting groups comprised of actual CTV ad experiences was the least engaged and inspired. Participants commented on the generic nature of the ads and pharmaceutical products as particularly likely to trigger disengagement.

Viewers surveyed commented on several positive aspects of the AI-enabled contextual ad experience. Actors from the show appear in the ads, with the same comedy style as the show, and subtle connections with themes and topics added to the humor.

The impact on viewers was positive. Viewers described themselves as happy, surprised and interested because of the ad experience, the ad content felt relatable and personable, and several indicated they liked the brand and offerings more.

The study also set out to determine whether brand suitability impacts a CTV viewer’s attention and brand perception. Twenty percent of ad experiences were categorized as contextually misaligned.

Participants expressed discomfort with unsuitable ad placements and generally tuned out those ads due to the disruptive experience. This resulted in reduced brand favorability.

One ad experience from the AI group was described as unsuitable, while the rest were from the demo and pub-declared groups. Some 54% were less interested in the product, and 31% said they like the brand less.

Next story loading loading..