Colwick, Kevin
Le, Thanh Ha
Makki, Amr
McKerlie, Chad
Patel, Niralee
Thapa, Aditi
- Starbucks logo branded coffee maker
- It is distributed from five distribution centers in the US
- They want to simplify the supply chain
- New system is programmed to use one of two forecasting methods:
- Simple Moving Average
- Exponential Smoothing
Forecasting Supply Chain Demand: Starbucks Corporation
- Three Weeks and Five Weeks of Data
- Evaluation Methods
- Mean Absolute Deviation
- Mean Absolute Percent Error
- Tracking Signal
- Simple Moving Average Results:
- 3-Week data is less biased
- 28% Error from Actual Demand
- Three Weeks and Five Weeks of Data
- No need to continually carry large amounts of historical data
- Most used of all forecasting techniques
- Exponential Smoothing Results:
- .4 alpha-value is less biased
- 26.3 % Error from Actual Demand
Exponential Smoothing is the Forecasting Method of Choice
- Provides more accurate results
- Easy to use and requires little computation
Starbucks should consolidate to a single distribution center
- Uniform data
- Lower maintenance cost