Advertising giant dentsu has dropped a new offering to enable video content transparency and keyword-level precision targeting for CTV streaming video advertising.
The new cookieless tool, Contextual Intelligence (CI) for CTV, runs on digital ad company GumGum’s contextual intelligence platform, Verity, which uses computer vision AI to analyze content at the video level.
IRIS.TV, a video data platform, works with publishers to make their video-level data accessible to partners like GumGum — acting as a “content data clean room,” as IRIS.TV CEO and Co-Founder Field Garthwaite puts it.
A Nielson report from earlier this summer found cable/linear TV in July dipped to below 50% for the first time and that streaming accounted for nearly 40% of total TV usage that month.
“Viewers are streaming more, and while TV advertisers want to invest in streaming, they do not have transparency into where their ads run,” Garthwaite told StreamTV Insider.
For example, Garthwaite referenced research from GumGum that revealed one-fifth of CTV ad breaks in children’s content contained “problematic” ads for casinos, gambling apps, alcohol and other adult-oriented products. But the advancement of AI has helped advertisers gain more knowledge of TV subject matter through the tool's ability to convert a TV show or movie into structured data relevant to targeting and measurement.
“This partnership came about because Dentsu wants to activate these capabilities at scale,” he said. “The GumGum contextual intelligence engine watches, listens and reads content frame-by-frame and assigns contextual and GARM [Global Alliance for Responsible Media] brand suitability segments to videos from IRIS-enabled supply sources.”
When a viewer watches an IRIS-enabled video, the IRIS_ID — a tool used to help publishers standardize content identification and metadata — is passed through the ad server. Ad platforms can then decode the data according to the buying method: direct, private marketplace or pre-bid.
“Our platform integrates and enables these new AI capabilities for publishers, data partners and ad platforms, hence the term ‘IRIS-enabled’ being used by different players in the ecosystem,” Garthwaite elaborated. “We protect our publisher partners by ensuring only select data is available for targeting and measurement by ad platforms to prevent violations of laws like VPPA [Video Privacy Protection Act] and to ensure media is purchased at scale by preventing cherry-picking.” He noted that only videos with IRIS_IDs will be eligible to receive ad investment from this partnership.
“Until the arrival of video-level contextual intelligence, CTV buyers had to target or filter by app, channel, or genre. This causes buyers to increase program restrictions to block certain apps, channels, and genres resulting in a reduced scale that gets further reduced by audience filtering,” he commented. Garthwaite said this collaboration allows buyers to capitalize on the GARM brand suitability framework to have more flexibility in managing risk.
And beyond risk management, the increased transparency enabled by the tool aids in more relevantly placed programming based on viewer sentiments and values.
“Brands can reach consumers in the right mindset and not just hope they reach the right person in a household of a family of four,” he said.