Client Objective
The client, a multinational consumer goods company specializing in health, hygiene, and home products, aimed to enhance the accuracy of their short-term demand forecasting, specifically at the SKU (Stock Keeping Unit) level. Their primary goal was to improve their ability to predict the number of units required to fulfill orders from a leading online seller and thereby avoid compliance fines due to overstock or understock situations.
Client Introduction
Challenge
The client faced several challenges in their supply chain, primarily centered around short-term demand forecasting:
- Accuracy Issues: Traditional demand forecasting methods based on time-series algorithms often resulted in inaccurate predictions, particularly when dealing with products that had yet to establish stable demand patterns.
- Compliance Fines: Inaccurate demand forecasting led to compliance fines, as the client frequently faced situations of overstock or understock, causing both financial losses and potential damage to their relationship with the leading online seller.
What We Did
To address the client's challenges and enhance their short-term demand forecasting, a comprehensive approach was adopted. Instead of relying solely on historical data and time-series algorithms, a combination of advanced data analytics and predictive modeling techniques were employed. The services included:
- Diverse Data Sources: Utilizing various data sources beyond historical demand, including POS data, inventory status, customer sentiment, footfall data, and competitors' strategies. Additionally, non-traditional sources such as social media, new product information, weather data, economic indicators, internet search trends, and IoT data were integrated.
- Data Integration: Advanced data integration and data warehousing techniques were implemented to harmonize and process data from multiple sources in real-time.
- Machine Learning Models: Machine learning models, such as neural networks and decision trees, were employed to analyze the integrated data and generate predictions.
- Predictive Modeling: Advanced predictive modeling allowed for more accurate short-term demand forecasting, specifically for the next 8 to 12 weeks.
Outcome
The project delivered a range of significant outcomes:
- 50% Improvement in Forecasting Accuracy: The client's ability to predict short-term demand at the SKU level improved by 50%, reducing the occurrence of overstock and understock situations and the associated compliance fines.
- Cost Savings: With more accurate demand forecasting, the client optimized their inventory management, reducing unnecessary costs related to excess inventory and emergency restocking.
- Improved Customer Satisfaction: Accurate demand sensing enhanced the client's ability to meet customer demands promptly and efficiently, resulting in improved customer satisfaction and retention.
- Competitive Advantage: The adoption of advanced data analytics and predictive modeling techniques gave the client a competitive edge in the consumer goods industry, enabling them to respond quickly to changing market dynamics and consumer behavior.
Conclusion
