synthetic data

Analytics & Forecasting 2026

Analytics & Forecasting 2026

  • Analytics & Forecasting

AI is transforming how organizations generate insights and forecast outcomes, but questions remain about reliability and the need for human oversight. On September 14-15 in NYC, industry and academic leaders gathered to discuss integrating AI into analytics workflows and maintaining transparency.

Member Only Access
A Year Inside the Psychology of GenAI: What ARF Research Reveals About How AI Thinks

A Year Inside the Psychology of GenAI: What ARF Research Reveals About How AI Thinks

  • Psychology of GenAI
  • ARF; MSI

After conducting eight studies over the past year, what has the ARF learned about the psychology of generative AI? Across experiments examining bias, identity cues, product recommendations, executive decision-making and AI-assisted research, four patterns repeatedly emerged: GenAI tends to compress human variability into dominant responses, subtle differences in prompting can meaningfully change outputs, different models bring their own constraints and tendencies and AI can imitate the language of expertise more readily than it can exercise expert judgment. Taken together, the studies suggest that GenAI’s greatest strength and its greatest risk may be the same: its ability to produce coherent, convincing answers can make compression, instability and missing judgment difficult to see.

Member Only Access
Efficient Brand Growth Through Neuroscience-Trained AI

Efficient Brand Growth Through Neuroscience-Trained AI

Bill Harvey – CEO, RMT-Bill Harvey Consulting

Michael Platt, Ph.D. – Director, Wharton Neuroscience Initiative

Bill Harvey (RMT-Bill Harvey Consulting) and Michael Platt (Wharton Neuroscience Initiative) presented a context-effects framework to help advertisers deliver sales growth more efficiently. Their methodology combines content coding, machine learning and neuroscience. The approach centers on measuring subconscious motivation by analyzing the content people consume. Thousands of psychological descriptors were reduced through machine learning to a smaller set of “Value Signals” that predict choice behavior. These signals are linked to neurological processes, with EEG synchrony identified as a strong predictor of advertising-driven sales outcomes. Resonance Mapping Technology (RMT) analyzes the Value Signals in ads, matches them to media contexts and targets audiences aligned with those signals. Third-party studies provided validation, demonstrating improved prediction of sales and brand lift when context and resonance were applied. Additionally, AI can now provide RMT encoding of ads and programs. Key takeaways:
  • Context is a critical part of the creative effects, and ads should not be measured without considering context and value signals.
  • Content consumption behavior is the key to understanding motivations.
  • All choice behavior in the brain is the same calculation whether the subject is choosing programs or brands.
  • RMT ad-context resonance is the only statistically significant predictor of EEG synchrony, according to Wharton Neuroscience.
  • Synchrony predicts 91% of the sales effect of an ad. The context effect is very strong and EEGsynchrony depends on the content that preceded the ad.
  • Incorporating RMT context significantly improved Wharton Neuroscience AI Digital Twin model’s ability to predict outcomes, such as the impact of ad spend on stock prices.

Watch the Presentation

Download Presentation

Member Only Access

Using Synthetic Data to Unlock Programming Measurement for A Streaming-First World

Rameez Tase – Co-Founder & President, Antenna

Gilles Duterque – VP, Data Science, Antenna

In this presentation, Rameez Tase and Gilles Duterque introduced Antenna’s approach to understanding how streaming viewership drives business outcomes like subscriber acquisition, retention and churn. Because traditional TV ratings do not capture the fragmented viewing behavior in streaming, they cannot tie ratings or viewing behavior to subscription success. Rameez and Gilles described the process and methodology behind Antenna’s solution, “Subscriber Views,” which addresses the gap by linking viewership data with subscription data, enabling analysis of what content drives sign-ups, keeps users engaged or leads to churn. The presentation emphasized that the goal is not just audience measurement, but actionable insights into user behavior across cohorts and platforms.

Download Presentation

Member Only Access

Tackling Fragmentation Challenges with Robust CTV Solutions

Pete Doe – Chief Research Officer, Nielsen

David Kurzynski – SVP, Data Science, Nielsen

To address the growing challenges of fragmentation in connected TV (CTV) measurement, Nielsen’s Pete Doe and David Kurzynski focused on Nielsen’s approach to solving them through a new methodology called STAR (synthetic total audience ratings). Pete explained how traditional measurement methods rely heavily on panel data and simple scaling, which work for large, popular content but break down for smaller or fragmented audiences. CTV introduces additional complexity due to multiple devices, lack of unified identifiers and limited person-level data. As a solution, Nielsen’s STAR methodology combines large-scale census (first-party) data with panel data using advanced data fusion techniques. It matches and enriches census data with demographic insights from panels, builds a synthetic panel to scale analysis and applies viewer assignment models to estimate person-level viewing. This hybrid approach enables more accurate, granular and cross-platform audience measurement—especially for long-tail content—while also incorporating privacy protections such as clean rooms and differential privacy. Key takeaways:
  • Hybrid data (panel + census) is the future of CTV measurement. Combining large-scale first-party data with panel-based demographic insights provides both scale and accuracy, overcoming the limitations of either approach alone.
  • Traditional methods fail for fragmented, long-tail content. Existing scaling methodologies depend too heavily on panel observations, making them unreliable for lower-rated or niche programming.
  • Person-level, cross-platform measurement is now achievable. Through data fusion, synthetic panels and viewer assignment models, solutions like STAR enables detailed, person-level insights across devices while maintaining privacy safeguards.

Watch the Presentation

Download Presentation

Member Only Access
 Next »