Process, OPS Rules, Supply Chain, MethodologyApplying analytics to solve operations and supply chain problems requires domain knowledge, technology and process. We have developed a 9-step methodology to guide you through this process.

  1. Define the problem and plan – First, we determine the supply chain or operations problem the company is facing. Then, we decide on the best approach.
  2. Create a baseline model – model of the company’s network. This involves mapping the locations of suppliers, manufacturers, Distribution Centers and customers. Then we took a look at various variables such as orders, supply, capacities, inventory, transportation etc.
  3. Validate the baseline model –  We validate the model in two phases. Phase 1 deals with comparing the baseline with information about the historical inventory levels and the difference between fill rate based on the inventory from the model. In Phase 2, the model is given historical fill rates and inventory levels and then it determines where there is insufficient inventory. This phase allows you to identify what facilities have problems and where the gaps are in the model.
  4. Quantify key drivers and quick wins – From the model, we can see that several improvements can be made just by optimizing the current model and without making any real changes.
  5. Identify and Create Scenarios –  In this phase, we make a list of scenarios to analyze by interviewing executives at the company and running a workshop to determine the problems they are facing.
  6. Optimize the Scenarios – Run an optimization of the scenario models and find the various costs, drivers and characterteristics of the solutions.
  7. Consolidate and iterate – Once the basic issues are determined, devise several scenarios to see what is driving the problems in the supply chain.
  8. Drive Change –   By determining the true drivers of supply chain behavior and cost, you can treat the root causes.
  9. Create a Repeatable Process – You need to be able to do the same analysis once there are changes in operations either by design or not. You need the capability to build and analyze models on an ongoing basis.
To see how these steps were applied at PepsiCo, please download the full case study below.