Problem
Diageo is a British multinational alcoholic beverages company. Diageo Russia is one of Russia’s top five wholesale alcoholic beverage distributors, which is traditionally a low margin business where the bottom line is sensitive to the customer service level and high logistics costs.
Diageo sought assistance from the consulting companies Logistics Field Audit (business and supply chain management consulting) and Amalgama (simulation modeling consulting) when they experienced an increase in sales volume but failed to realize larger profits due to logistics costs per unit.
Other mounting concerns for Diageo were customer service level and cost of goods sold, combined with an increased inventory and future plans of development including a new warehouse in Russia and expansion to Urals and Siberia.
To translate these initiatives into action, predictive analytics and logistics planning were necessary. The consultants were also assigned to manage Diageo’s big data with simulation-based decision support to show and prove ways to decrease logistics costs and implement logistics network optimization. Logistics planning and optimization with AnyLogic software provided the ability to achieve these goals.
Solution
The logistics optimization model includes three existing and one prospective factory, three border crossing points, three existing and five perspective warehouses, two customs offices and up to 300 demand points grouped into 45 service groups. In addition, the logistics optimization model contains a replenishment algorithm, an order aggregation algorithm, load bearing algorithms, and delays at border crossing points. Demand and sales- forecasting accuracy for all 280 Diageo products in 6 types of warehouses were also built into the model.
Logic built into the logistics optimization model includes a replenishment algorithm for a segment of the logistics network that begins at the consolidation warehouse Due to Russia’s large size and relatively slow transportation capacities, it allows for five days of lead time to the central distribution center, and five days of lead time to the original distribution center. The replenishment algorithm takes into account a requirement diagram (planned sales diagram), current stock, lead time, and minimum order size, then generates the requirements for replenishment, identifies the coverage gaps, (time periods when stock will be lower than the lowest threshold), and takes action to prevent any gaps in coverage. The capability to observe the model behavior dynamically, a result of the logistics simulation using AnyLogic software, was extremely useful in this case.
Validation of the logistics optimization model was imperative and began by comparing data from the previous year’s SAP ERP system data and gave results with less than a 5% differential.
Outcome
The initial value for Diageo included an increase in sales forecasting accuracy from 60% to 80%, with a pay-off period of less than 2 years. This increase will allow Diageo to reduce their target stock level by 40% which will reduce logistic costs per unit by 7%, even with sales growth plans. The research also showed there was no need for additional warehouse space, as the stock required to maintain the target service level was unreasonably high.
After running the simulation, the logistics optimization model provides the stock level forecast for each product (15 days ahead), a complete cost for each and every delivered product unit and proves to the client what the target state of the supply chain should be.
View Andrey A. Malykhanov’s entire presentation about logistics simulation and optimization using AnyLogic software delivered at the AnyLogic Conference 2013:
