Predictive analytics is a powerful technology that can improve an enterprise’s efficiency and profitability.  However, despite the obvious benefits, introducing predictive analytics does come with some challenges.  One such challenge is the notion that predictive analytics technology represents a “black box” that is difficult to understand and explain.

RSI’s patented technology is not based on static formulas or algorithms. We use a Monte Carlo simulation technique to simulate supply chains for any given item in any given location. We start with a BOH (Beginning On-Hand) inventory quantity and simulate daily changes in inventory levels over a given period of time (e.g., quarter).

During the simulation, daily inventory goes up due to receipts of supply and goes down due to demand or any other valid consumption of inventory.  During the simulation, RSI pays particular attention to two things:

• Average inventory level: the average on-hand inventory quantity throughout the simulation

• Average service level: is this simulated inventory quantity sufficient to meet demand requirements at the desired customer service level? 

RSI’s simulation technology determines the minimum inventory level that has the highest probability of meeting the required service level. We have found that processing a minimum of 2,000 Monte-Carlo-style iterative simulations best drives a high-confidence statistical probability that the RSI target inventory level will meet the required service level for each item at each location.

Traditional inventory optimization formulas and algorithms create inventory deficiencies or surpluses because they cannot account for the required supply chain variables to accurately determine inventory levels.  By not accounting for these critical supply chain variables, these methods essentially do not represent each inventory item’s actual unique variation and constraints.