The ARF hosted its annual flagship conference, AUDIENCExSCIENCE 2023, on April 25-26, 2023. The industry’s biggest names and brightest minds came together to share new insights on the impact of changing consumer behavior on brands, insights into TV consumption, campaign measurement and effectiveness, whether all impressions are equal, join-up solutions across multiple media, the validity, reliability and predictive power of Attention measures, targeting diverse audiences, privacy’s effect on advertising and the impact of advertising in new formats. Keynotes were presented by Tim Hwang, author of Subprime Attention Crisis, Robert L. Santos of the U.S. Census Bureau, Brian Wieser of Madison and Wall, LLC and Andrea Zapata of Warner Bros. Discovery.
Sergey Fogelson – Head of Data Science, TelevisaUnivision
Edouardo Vitale – Data Scientist, TelevisaUnivision
Sergey Fogelson and Edouardo Vitale, both from TelevisaUnivision, outlined their motivations for developing a custom lookalike model (LAM) to expand Spanish-language audiences, which were under-represented:
Misidentification: 4 in 10 Hispanics are excluded from 3p datasets.
Waste: 70% of impressions targeted at Hispanics are wasted.
Scale: The true scale of the Hispanic population within a given brand’s 1p dataset is hard to identify without extensive validation.
In order to address this audience underrepresentation, data sources were leveraged to create a household graph incorporating 1P (first-party) viewership, 3P (third-party) viewership and demographics on age, gender, income and education from TelevisaUnivision’s partner. The combination of these sources created a robust household level representation of approximately 17MM Hispanic households in the U.S.
Embeddings were used to create a latent space that allowed for the comparison of user similarities. Mathematically, very similar users have very similar embeddings. Similarities between individuals may be based on what content they viewed, where the content was viewed (zip code) or demographics of the viewer. Additional details of the embedding process as well as the autoencoder architecture steps and validation process were presented.
Developing a Lookalike Model (LAM) to expand Spanish-language audiences, corrected for the underrepresentation of this consumer target.
Expanding an audience with LAM identifies individuals who look and act just like a given target audience. These look-alike models are used to build larger audiences from smaller segments in order to create reach for marketers and advertisers and enable them to transact on an expanded audience.
Use of LAM can overcome the challenges of misidentification, waste and scale. LAM plus the household graph achieves significant increases in overall audience scale.
Heather Coghill – VP, Audience, Warner Bros. Discovery
Daniel Bulgrin – Director, Research Operations & Insights, MediaScience
Heather Coghill (WBD) and Daniel Bulgrin (MediaScience) shared methodologies and results from two in-lab studies that sought to understand how impactful category priming can be without brand mention and if viewers associate brands with adjacent unsuitable content.
Their presentation focused on two types of contextual effects within program context—“excitation transfer” and “brand priming”.
To see if these effects carried over to ad content through excitement or brand recognition in the content, the research team utilized distraction-free viewing stations that enabled neurometrics and facial coding followed by post-exposure surveys. Impact on brand perception was measured with lifts in brand attitude, attention and memory.
Results showed brand priming did change how viewers experienced the ad by lifting brand recognition, with stronger effects in heavier ad loads. The research also concluded that although brands are not harmed by adjacency to perceived unsuitable content, context effects still need to be considered.
Even moderate category primes can push through effects, despite modest impact, in both linear and CTV. Category priming in streaming with limited ads impacted middle and lower funnel metrics, with 31% of viewers noticing a connection between the ad and the program.
Although viewers agreed that low intensity “unsuitable” content was most acceptable for advertisers, there were no adverse effects as intensity levels increased—all levels were deemed suitable for advertisers, with no significant differences in brand recall or purchase intent.
More research is required to understand what is unsuitable for brands. The current guidelines are based on what is thought to be unsuitable—not social science.
Metrics for planning, buying and evaluating buys have been in great flux, especially over the last five years. New channels have emerged, some have changed, and a multiplicity of data sources have sprouted up. To gain a better understanding of the way advertisers are navigating this complex landscape, the Online-Offline Working Group of the ARF Cross-Platform Measurement Council interviewed representatives from major advertisers and put out a report about what they learned. This report provides the advertising industry with a glimpse into how major marketers are approaching audience measurement in all the different environments.
Deloitte, with input from the 4A’s, ANA and CIMM, has issued “The Future of Television: A Transition to a Multi-Currency National TV Market: Industry Perspectives on TV Currency.” The report describes the TV industry as in transition and that the future of the TV currency — or rather currencies — is yet to be determined.
The ARF Universe Study of Device and Account Sharing (DASH) is a nationally projectable enumeration study of consumer behavior in TV and digital media. DASH records in granular detail how US households connect to and consume TV, use digital devices, and interact with and share streaming media and ecommerce accounts. Launched in 2021, this syndicated study is designed to serve as an industry standard truth set for insight and data calibration. The report just released by the ARF highlights findings from the first wave of DASH 2022. Data and findings from the full 2022 data set will be available in January 2023.