During a Category Management Association (CMA) webinar, the topic of discussion was the "say-do" gap in retail. This gap presents various challenges, but the article examines how AI has revolutionized demand transfer analysis and provides examples of its practical applications.

The "Say-Do" Gap: A Complex Challenge

Understanding shopper behavior has long been a challenge in the retail world. Consumers often say one thing but do another, creating discrepancies between their intentions and actions. This "say-do" gap can significantly affect category strategy and assortment planning decisions for three critical reasons:

1. Complex Variables: Retail behavior is influenced by numerous variables, such as demographics, shopping habits, lifestyle preferences, and even social media interactions. This complexity hampers the accurate prediction of consumer behavior using traditional methods.
   
2. Gaps in Traditional Research: Conventional market research methods, like syndicated and household panel data, primary and secondary research often suffer from biases, including recency bias, omission bias, and the "say-do" gap itself. These issues undermine the accuracy and reliability of collected data.
   
3. Limited Resources: Retailers and CPG manufacturers operate under budget, time, and workforce availability constraints. These limitations can hinder thorough analysis and lead to the underutilization of available data.

AI and the Transformation of Demand Transfer Analysis

HIVERY employs ensemble learning techniques, notably deep learning, which uses neural networks with multiple layers to analyze and interpret data. This approach excels at identifying intricate patterns within complex datasets, making it ideal for understanding consumer behavior at the store level. With such granularity, category management experts, sales leaders, and shopper insights professionals can assess the impact of new innovations, promotions, adds, deletes, or price changes at both the store-specific and cluster levels.

The HIVERY model, as used in HIVERY Curate, also incorporates consumer decision trees (CDT), another vital component of its ensemble learning strategy. Decision trees allow the model to make decisions based on a series of questions or conditions. By incorporating CDTs, HIVERY creates a more holistic view of consumer choices.

The predictive aspect of demand transfer is where ensemble learning particularly shines. By leveraging insights from multiple models, such as deep learning and decision trees, the model can offer more accurate predictions about the interplay of various factors in the retail environment.

HIVERY Use Cases to Reduce Gap and Improve Decision-Making

Three significant use cases demonstrate how AI transforms demand transfer analysis:

1. King Cobra and Coors Light Example: AI algorithms detected unexpected correlations between the removal of Coors Light and an increase in King Cobra sales in specific stores. The model identified nuanced shopper segments in a particular Bentonville, AR store and suggested that if Coors Light in a 4-pack size format were unavailable, the "best next offer" would be King Cobra in a single-serve format. This recommendation seemed counterintuitive initially. However, after receiving feedback from the retailer's buyer, who had experience in the category and stores, the buyer agreed to the change. This use case underscores the model's precision in understanding shopper behavior at the store level. After all, the best predictor of shopper behavior is actual shopper behavior.
   
2. Bifurcation Effect: AI identified a trend where shoppers were trading up or down within a category based on changing preferences and financial constraints. This bifurcation effect was observed across various categories, highlighting AI's ability to pinpoint emerging trends.
   
3. Regional Trends: AI identified a specific Texas sauce brand with regional popularity in one location. However, based on changing population migration patterns, AI recommended expanding the distribution of this sauce to other regions. This example underscores AI's ability to detect regionally influenced behavioral shifts.
   

The Future of AI in Retail Decision-Making

The future of retail decision-making looks bright with the continued evolution of AI technologies. AI will use various data sources such as security cameras, RFID tags, and augmented reality interactions to collect accurate shopper data.

These new data sources and analysis methods will allow retailers to create personalized strategies and improve the efficiency of category resets and assortment processes.

If you're interested in exploring this further, speak with one of our experts about HIVERY Curate and unlock the potential of your growth strategies.

Related resources you might be interested in:

• Breaking the Doom Loop: Leveraging AI to Transform Data Abundance into Actionable Insights for Retail Assortment Decisions
• Discovering 'Sleeper Stores': How AI is Redefining Retail Insights and Shaping Buying Behaviors
• Interactive Assortment Planning: Using AI-Driven Visual Insights to Make Difficult Decisions
• Category Growth Strategies: 6 Use Cases for Scenario Analysis & Effective Assortment Planning