The document discusses two topics: 1) analysis of air vs ocean shipment methods based on product cost, demand, lead times, and shipping costs. The optimal choice balances lead time and cost. 2) A case study of a footwear manufacturer introducing new products. Historical demand data is used to predict demand 12 months prior using clustering and classification techniques. This decreases lost sales. Network planning aims to optimize locations, warehouses, sourcing, and distribution channels to minimize costs while meeting service levels.
The document discusses two topics: 1) analysis of air vs ocean shipment methods based on product cost, demand, lead times, and shipping costs. The optimal choice balances lead time and cost. 2) A case study of a footwear manufacturer introducing new products. Historical demand data is used to predict demand 12 months prior using clustering and classification techniques. This decreases lost sales. Network planning aims to optimize locations, warehouses, sourcing, and distribution channels to minimize costs while meeting service levels.
The document discusses two topics: 1) analysis of air vs ocean shipment methods based on product cost, demand, lead times, and shipping costs. The optimal choice balances lead time and cost. 2) A case study of a footwear manufacturer introducing new products. Historical demand data is used to predict demand 12 months prior using clustering and classification techniques. This decreases lost sales. Network planning aims to optimize locations, warehouses, sourcing, and distribution channels to minimize costs while meeting service levels.
Air Vs Ocean Shipment Analysis New Product Introduction-Case Study
Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone SUMMARY of DIGITAL SCM : Lecture 17&18 Air Vs Ocean Shipment Analysis New Product Introduction-Case Study Data to be considered: • Footwear Manufacturer, Short Life Cycle & low demand accuracy ✓ Product Cost • Goal: Predict demand for new products twelve months prior to ✓ Mean Weekly Demand product in market. ✓ Review Period • Scope: About 2000 different SKUs ✓ Ocean & Air Lead times Step 1: Historical Data (Data from all the products sold) ✓ Ocean & Air Shipping Cost Step 2: Plotting the demand curve for the data • At higher level it is the trade off between lead time and Step 3: Identifying number of curves which have same behavior of cost demand curves and use machine learning techniques like clustering(K- mean) Base Stock Level Policy Step 4: Use Classification tree and map the new product from the Average Demand during an interval of r days, L being the lead clustered data. time: (r+L)*(AVG) (AVG: Average daily demand) Using Forecast error to check the accuracy Safety Stock: z* STD*SQRT(r+L) Implementation Results: Decreased lost sales Base Stock: (r+L)(AVG)+z*STD*SQRT(r+L) Network Planning • Key issues: Optimal number of location, size of warehouses, sourcing strategy, best distribution channels • Objective is to balance service level against purchasing cost, inventory handling costs, facility costs, transportation costs • Minimum annual cost configuration of the distribution network that satisfies product demand at specified customer service levels. • Major tools: Mapping, Data, Engine • Comparison between cost minimisation and service maximisation networks • Cost differences observed on Aggregation of Customer as well as with Aggregation of Product Category Impact of Aggregating Customers Strategy for Product Aggregation • Customer zone balances the loss of accuracy due to over • Placing all SKUs in a source group aggregation and needless complexity • Aggregating the SKUs by similar logistics characteristics (Weight, • Efficiency of aggregation is affected by number of Volume, Holding Cost) aggregated points and distribution of customers in each zone