Synthetic Data: Pitfalls and Promises

Introduction

Synthetic data is one of the most discussed - and often least clearly understood - topics in market research today. Much of the current conversation is driven by vendor claims, future-focused speculation, or second-hand commentary. This workshop takes a different approach. Grounded in real project experience and live testing, it examines how synthetic data is actually being used in market research today - where it delivers value, where it introduces risk, and how research buyers should evaluate it responsibly. Participants will review real examples comparing synthetic and human data, explore practical use cases across modeling, personas, and emerging synthetic sampling, and leave with clear frameworks for deciding when - and when not - to use synthetic methods.

Who will benefit:

Research & insights professionals, analytics leads, brand and product leaders
(Non-technical orientation; no data science background assumed)

Learning Method:

Live online workshop (interactive, discussion-led)
20–25 participants by design

Overall Learning Objectives

By the end of this workshop, participants will be able to:

• Clearly define different types of synthetic data and distinguish between modeling, personas/agents, and fully synthetic survey sources
• Understand what synthetic data can and cannot do today for market research and insights – applications, examples, use cases
• Apply a buyer-oriented evaluation lens to assess vendor claims responsibly
• Identify appropriate limitations and guardrails for synthetic data in real research workflows
• Confidently communicate synthetic data use, risks, and validation to internal stakeholders and clients

Key Topics

• What synthetic data modeling actually provides
• How modeling differs from:
• Weighting
• Traditional imputation practices/approaches
• Personas or agents
• What kinds of data can and cannot be modeled reliably
• Where modeling works best:
• Quota shortfalls
• Hard-to-reach populations
• Poor-quality or fraudulent data replacement
• Core limitations and potential failure areas

Evidence & Examples

• Review of real-world independent modeling tests (publications list to include several independent tests to reference – Quest, NewMR, Dig Insights, etc.)
• Side-by-side comparisons of modeled vs. original data
• Where variance appears — and why that matters

Interactive

• Examples of modeling results – several test cases showing how modeling delivers, attendees to choose which case(s) to discuss, addressing questions and concerns
• Discussion: “Would you trust this output — and for what decisions?”
• Quick poll: attendee opinions about reliance on modeled data for common business scenarios
  

Trainers:

Scott Worthge, Founder, mktresearchguy llc

Scott's market research career includes more than 40 years as an insights practitioner, combined with part-time university-level teaching for more than 20 years, notably for UC Berkeley and Michigan State's Masters in Market research program in the U.S.  Much of his day-to-day work involves evaluating new technology and practices for data collection, testing promising methodological changes that potentially will deliver better data quality and results, especially for quantitative research projects.  Recent areas of focus for his investigations and real-world tests have been synthetic data modelling, voice and video data collection effectiveness, and multicultural trends in the U.S. affecting "representativeness".

Scott is a frequent speaker at conferences, webinars and industry association meetings.  His schedule takes him to more than a dozen locations worldwide annually, with organizations such as ESOMAR, IIEX, the Insights Association and the Market Research Society.  In addition, he has conducted subscribed webinars for ESOMAR and IIEX with record-setting attendance.

Yogesh Chavda, Founder, Y2S Consulting

Yogesh Chavda’s 25+ years as a marketing executive are highlighted by pivotal roles and impactful initiatives at organizations that dominate their field and have left a cultural impact. In addition to stints at Spotify, Pinterest, Amway, Kimberly Clark, and WS Audiology, Yogesh’s background includes a 16-year run at Procter & Gamble, where he held various senior-level positions across 6 countries.

Yogesh is the founder of Y2S Consulting, where he’s brought his passion for empowering and facilitating content on brands impacted by stigma and applications of AI in Marketing. He is recognized as an international expert in the area of agentic AI and its applications to complex marketing uses. Some of his clients have included Apple, Whirlpool, 3M, Parallel, Margaritaville, Lowe’s, Foster Grant, and Mars. 

Yogesh also writes about AI in Marketing, in his weekly newsletter – In Search of Normal on LinkedIn. And he hosts a Podcast called the Next Frontier In Insights where he interviews thought leaders and innovators at the forefront of Marketing.