CHALLENGE:

Procter & Gamble is an American multinational consumer goods corporation. The office of a South East Asia country, where P&G operates, faced a challenge – with the product portfolio presented in outlets across the country, the company followed a network design and inventory policy that could not support current demand and was not efficient enough. This, in turn, could lead to negative financial implications and, in the perspective, have an adverse effect on all regional operations.

P&G Distribution Center The company wanted to review its supply chain network design, as well as optimize inventory policies. Seeing opportunities in digital supply chain optimization, they tasked consultants from the SupChainEra company with network design and auditing. The advisors from SupChainEra would create a computer model of the existing network and test business assumptions in a risk-free environment to create a plan for implementation of best-performing strategies in real life.

SOLUTION:

For building a network optimization model, the consultants applied anyLogistix supply chain software. The combination of network optimization and dynamic simulation capabilities in one package would allow them to measure every aspect of supply chain performance and would provide more accurate and transparent decision making. A powerful set of in-package experiments would save time when designing and testing new policies.

The consulting team started with building a baseline model that incorporated all 20,000 sites where P&G’s goods were manufactured, distributed, or purchased, starting from major production sites to small retail stores. anyLogistix naturally fits for creating large-scale network models, and the consultants benefited from this opportunity.

In the baseline model, goods could be transported from P&G factories to its distribution centers (1st leg, fixed delivery cost), between distribution centers, and from distribution centers to end customers (2st leg, distance-based costs). To make the baseline network design more efficient, the consultants applied supply chain network optimization, available with anyLogistix software. Tracking metrics included:

  • Estimated lead time (ELT) by products and orders
  • Number of vehicles used for transportation
  • Vehicle utilization rates
  • Inventory level per DC

When looking at shipment costs, it turned out that they were much higher when transporting goods between DCs. The consultants analyzed the model output statistics and proposed converting two of the most heavily used DCs into logistics hubs. This way, instead of being shipped from one DC directly to the other, goods would be collected in the hubs and then transported to other DCs. This would entail lower costs and more efficient supply chain network operations.

Using anyLogistix capabilities, the consultants were able to map out and test the optimized network structure, including direct and indirect routes between DCs. It was calculated that P&G would save 20% of the total transportation costs if implementing the network structure.

The next step for P&G in supply chain processes optimization was shifting to MRP strategy. This strategy would allow the company to keep the service level high when shipping goods through the supply chain. To plan the strategy, the SupChainEra consultants applied the known demand of end customers and calculated demand forecast for DCs using machine learning algorithms. The results were later used as inputs in MRP policy settings in the baseline model. This resulted in an average inventory level decrease, but also affected the service level at some distribution centers. In particular, two island-based distribution centers experienced a service level dropdown from 98% to 94%, which was not acceptable. To normalize this metric, consultants applied the anyLogistix safety stock optimization approach, which resulted in a balanced service level and a 40% inventory level reduction.

Finally, the consultants incorporated the designed MRP strategy into the optimized network, ran the simulation, and compared the metrics with the baseline model.

OUTCOME:

The final network design and policies, compared to the baseline design, showed better performance in terms of service level and cost saving. Average inventory level dropped 35% and total costs decreased to 20%. The designed structure was proposed to the executives, received positive feedback, and was advised for implementation.

More case studies