Evaluation of Fast Food Promotion Effectiveness Using a Hybrid Approach of Robust Non-Parametric Methods and Monte Carlo Simulation Based on Market Size Segmentation

Abstract

The fast food industry is highly competitive, requiring effective promotional strategies to drive sales and maintain customer loyalty. This study evaluates the effectiveness of fast food promotions using a hybrid approach combining robust non-parametric methods, Random Forest, and Monte Carlo simulation. The analysis focuses on segmenting the market by Market Size (Large, Medium, Small) to identify the most impactful promotional strategies for each segment. Non-parametric A/B testing using the Kruskal-Wallis test revealed significant differences in sales across promotions, with Promotion 1 emerging as the most effective overall. The Random Forest model highlighted LocationID as the most critical factor influencing sales, particularly in Large markets. Monte Carlo simulation further demonstrated that Promotion 1 yields the highest Expected Monetary Value (EMV), making it the optimal choice for long-term sales growth. The findings emphasize the importance of tailoring promotional strategies to specific market segments, considering factors such as location, timing, and store history. This study provides actionable insights for businesses to optimize promotional campaigns, enhance sales performance, and achieve sustainable growth in the fast food industry.