Google Research Opportunity

Research Opportunity: Response Modeling with Aggregate & Event-Level Data 

The Wharton AI & Analytics Initiative (WAIAI) is pleased to announce a new dataset provided by Google as part of our exciting two-year academic partnership. This collaboration aims to advance research on Marketing Mix Modeling (MMMs) and Multi-Touch Attribution (MTA) tools to better understand the impact of marketing on sales and conversions, despite privacy-related restrictions.

Opportunities for Researchers

WAIAI and Google seek to encourage research and new thinking on novel methodologies for outcome modeling. The partnership will focus on integrating academic insight with industry-leading technology to explore the following research opportunities:

• Innovate on Aggregate Models: Focus on fitting MMMs using granular data, running MMMs more frequently, and combining aggregate data with partial user-level data
• Innovate on Event-Level Models: Combine aggregated and user-level transactional data, ensure consistency between MTA and MMM models, and use a combined MTA+MMM framework for marketing optimization
• Enhance Aggregated Models for Brand Metrics: Capture long-term effects of ad campaigns using surveys or query volume data, and consider step-wise or nested model structures
• Propose Novel Approaches: Researchers can propose their own innovative approaches to outcome modeling

Dataset Information

Google will share an anonymized dataset focusing on the Google Pixel mobile phone, covering 2-3 years of sales and marketing data in the US. The dataset includes:

• Sales Data: Daily aggregated data including date, Designated Market Area (DMA), sales quantity, and indexed sales dollar
• Marketing Data: Daily aggregated media/campaign data, marketing channels, and marketing spend
• Engagement Data: Google query data for Google Pixel, including branded and generic search queries, and email marketing data
• Customer-level Data: Anonymized user ID, dates of ad impressions and their corresponding media channels, and conversion dates
• External Inputs: Economic data, competitor data, and more, such as population for each DMA

Google will also provide access to Meridian, a highly customizable open-source Marketing Mix Model framework (currently offered in limited availability!)

Deadline for Submissions

Submissions for this research opportunity are currently closed. Please check back on our website for more research opportunities in the future.