Drove 35% EBITDA Increase for $2B Food ManufacturerSupply Chain & Operations
- SKU-Level Forecasting
- Big Data Analytics
- Operations Modeling
- Production Planning & Scheduling
- Distribution Network Analysis
- Transportation Route Optimization
For a food manufacturer, growth came with added burdens including seemingly unpredictable demand, complicated production, and a sprawling distribution network.
The company sought to reduce costs and increase product freshness by reviewing forecasting, production planning, and distribution operations.
Over 100 million records of customer and trend data were assessed using big data analytics to tease out highly predictive forecasting models. The highly improved forecasts combined with improved production planning yielded lower production costs. Additionally, the optimized distribution network resulted in lower logistics costs and shorter product storage and transit times leading to fresher product received by customers.
The reduced direct operating and logistics costs drove an immediate 35% increase in EBITDA while additional opportunities to further decrease costs and increase revenue were identified in longer-term initiatives
challenge
With hundreds of SKUs, a mix of seasonal and non-seasonal products, as well as consumer and commercial customers, the company struggled to get a firm grasp on demand and production complexity. The distribution network had also grown in an ad-hoc manner over the years, masking substantial inefficiencies and inflated costs.
- Improve the forecasting process:
Big data analytics was deployed to analyze over 100 million records of historical and predictive demand data at a granular level. This quickly revealed mixed but consistent demand patterns where some SKUs had relatively stable demand year-round and others experienced large but predictable spikes throughout the year.
- Improve production planning and determine the right inventory level:
The existing production planning processes was assessed and key cost drivers which were previously unknown were captured. These included time studies on each production line, the creation of a changeover cost matrix, and bottleneck analysis.
- Optimize the distribution network and shipment routes:
The proliferation of distribution centers and longer shipping routes added cost and consumed valuable time where product shelf life was measured in weeks. The team preformed a distribution network analysis to optimize the distribution and logistics operations.
solution
The solution began with developing reliable demand forecasts that could then be used to improve production planning and scheduling, right-size inventory levels, reduce spoilage, and optimize the distribution network.
>Advanced analytics & accurate forecasting::Historical and forward-looking analytics drove a highly reliable SKU-level forecast by calendar day and reduced the mean average forecast error (MAPE) to under 5%. The customized code and techniques were provided to the company’s forecasters to enable them to update the analysis as needed using existing tools.
- Improved production planning and scheduling:
Enhancements to the S&OP process were made to proactively factor in non-production line requirements such as maintenance, R&D and other downtime. Production time and changeover cost studies were conducted and combined with personnel, inventory, spoilage, and logistics costs to develop a production scheduling methodology that better achieved due dates while reducing costs.The team modeled the new methodology to confirm the impacts and the client implemented the changes in a live production test the following week. Despite significant changes to many aspects of plant operations, the live test results followed the model predictions precisely. After two weeks of testing, the new methods were fully adopted as part of the regular production planning process.
- Distribution network analysis and optimization:
The team’s distribution network analysis focused on optimizing distribution center locations, 3rd party logistics costs, and transportation costs, while reducing time in transit to boost remaining shelf life and product freshness upon customer receipt. Some DCs were eliminated, new trucking routes were established, and the total product volume was re-balanced among the remaining centers and routes. Most changes yielded immediate operational and financial benefits although some distribution site closures would take up to a year.
results
Highly accurate forecasts were developed for SKUs where demand was previously perceived to be erratic. The new forecasts enabled development of a new production planning methodology that resulted in more efficient production runs, better personnel management, and improved inventory management. The distribution network was also optimized, reducing logistics costs and time in transit:
- Decreased forecast error (MAPE) to under 5%
Highly accurate calendar day forecasts provided an extremely reliable view of SKU-level demand.
- Reduced costs yielded an immediate 35% increase in EBITDA:
The new production planning process reduced bottlenecks, used fewer and more efficient changeovers, and reduced storage and transportation time.
- Estimated 11% reduction in logistics costs with 8% fewer miles driven
The optimized distribution network operated with fewer DCs, less storage/transit time, and fewer net miles driven.