Extreme Reach Study Finds Increase in Percent of White People Seen and Male Voices Heard in Video Ad Creative in 2022

‘Diversity in Ad Creative,’ an AI-driven Report, Analyzes 1 Million Ads From 2019 to 2022

Video ad creative features more white actors in 2022 than in 2021 or 2020, and male actors are far more common in ads than female actors, according to a landmark study unveiled by Extreme Reach (ER), the global leader in creative logistics.

Marketers wield enormous power in effecting cultural norms by how they portray people and situations in brand stories. Since the dawn of advertising, who is cast to “play” consumers and who is heard speaking sets standards around beauty, gender roles, who makes purchase decisions, and more. It is against this backdrop that Extreme Reach analyzed 1 million ads deployed in North America, on linear TV and digital platforms, from January 2019 to October 2022, assessing each ad for its composition by ethnicity, race, gender, and age.

The result is a first-of-its-kind, large-scale, benchmark report, Diversity in Ad Creative, that leverages artificial intelligence and machine learning, supported by human quality check, to scan and analyze ads across a range of categories.

Recommended AI News: Syniti Match and Syniti Replicate Now Available on SAP Store

“ER’s platform plays a central role in moving the vast majority of ad creative through the marketing supply chain to any linear or digital screen which uniquely positions us to provide our clients with expansive insights previously unavailable in one comprehensive view,“ said Tim Conley, CEO of Extreme Reach. “In addition to tracking the critical details of global rights management, when and where ads run, which assets are used and which never air, the use of artificial intelligence and machine learning now enable us to understand diversity and accessibility metrics that are increasingly important to brands.”

The research found that video advertising on linear and digital outlets is whiter in 2022 than in the past two years, and roughly equivalent to 2019 levels. White actors have accounted for 73% of the people seen in video ad creative in 2022, compared to 66% in 2021, 67% in 2020, and 74% in 2019. Meanwhile, 94% of ads contain at least one white actor, the highest rate of occurrence in the four years analyzed.

At the same time, Black, Asian, and Hispanic representation in video ad creative declined this year, with Hispanic representation reaching its lowest mark in four years.

“Diversity is such an important topic, yet there has been little quantifiable information available at scale related to the people featured in advertising creative,” said Melinda McLaughlin, CMO of Extreme Reach. “We set out to give brands and agencies the information needed to make the decisions that matter to their business. This report provides a high level view of diversity in ads that establishes the baseline for the industry, from which brands can devise their own unique roadmap.”

Recommended AI News: SilverSky Raises the Bar in MDR Services: with Continual Expert-Led Cyber Range Services

ER’s study also reveals that video advertising is dominated by men, who make up 65% of people seen in ads in 2022 and 73% of the voices heard, up from 64% and 65%, respectively, in 2021. In terms of age, while the composition of ages represented has shifted slightly older, the 20-39-year-old segment dominates 2022 ad creative, at nearly three times that of the U.S. population (78% vs. 27%).

Additional findings include:
-All large advertising sectors covered in the report skew male with regard to who is seen and heard on-screen.
-Three large ad verticals, Retail, Food & Beverage and Sports, include female voices at a higher percentage than they are seen.
-The Restaurant ad vertical is the most diverse.
-11 million people in the U.S. alone are hard of hearing or functionally deaf, yet this year only 1 in 3 creative assets include captions.
-12 million people in the U.S. have low vision or identify as blind. Less than 1 percent of all ads include an audio description.

Overview of ER’s Proprietary Methodology

ER employed a robust, multi-point process to analyze every occurrence of every face seen throughout each video ad in addition to isolating the audio speech track to identify each unique voice. ER’s research leverages 8 well-known existing solutions for different aspects of the analysis and then creates advanced models and continuously-fed machine learning that generates data more specific to advertising content. The company credits the following solutions, which in combination provide the multiple observations and predictions imperative for triangulating data points for high confidence results: AWS Rekognition DetectFace API, FairFace Project, FaceNet, DeepFace, Deezer’s Spleeter Library, Amazon Transcribe, and DBSCAN. Additional methodology detail is provided in the report.

Recommended AI News: Attain Partners with OpenX to Enable Supply Side Targeting For Advertisers

[To share your insights with us, please write to sghosh@martechseries.com]