The traditional forecast driven model, starts with the forecast. This drives the strategic S&OP, the tactical Master Production Schedule (MPS) and the operational Materials Requirement Planning (MRP) or Distribution Requirement Planning processes (DRP). This is the primary reason why companies are constantly trying to improve the demand signal. It drives everything. Once you have reconciled your forecast with capacity, you apply the item attributes and policies and run a synchronized MRP plan (either daily or weekly). The MRP pushes inventory from suppliers through manufacturing and into your distribution networks.
To compound this, there is the further complication of actual demand. Customer orders that are independent of the planning system drive further adjustments to production schedules. You manage the plan by chasing supplier orders that are late, based on the propagated forecast driven dates.
The result is a Bi-modal stock model, too many materials with too little stock and too many many with too much stock, but not enough of the correct materials required. This is exacerbated by increased work in process (WIP) due to unplanned changeovers. As we all know, MRP is designed to drive all stock levels to zero, so in order to protect us safety stock is added at the raw materials and finished goods levels. The result is stock levels that are even more inflated. Safety stock isn’t designed to be used, it sits directly below the zero stock level within the MRP calculation. What we can and will build!
We live with the conundrum of trying reduce stock levels and improve customer service, when it is proven that there is a direct correlation between the size of the safety stock level and customer service.
In the Demand Driven (DD) operating model we still require a forecast but this forecast doesn’t drive the operating model and requires no optimization. We use it to create an Average Daily Usage (ADU). Using the items supply and demand variability as well as its attributes, we size strategically positioned buffers (Where BEFORE How Much and When!).
We are now able to decouple our operating model preventing variability multiplying across it. The buffers allow us to use actual demand to drive the operating model instead of the forecast, giving us an accurate demand signal. The next step is to run a decoupled DDMRP plan that synchronises the supply chain by pulling inventory from manufacturing and in turn suppliers, based on real demand. The execution of the plan is no longer due date based, but based on what we need to ensure we fill our actual demand. The defined capability to build what we can and will sell!
The DDMRP operating model together with the Projected Daily Usage (PDU) generate the projections required in the tactical range and in demand driven sales & operations planning (DDS&OP), not the forecast. We don’t have a Master Production schedule, but rather Master Production settings. The input for these come from the DDS&OP. Another major difference is the addition of an execution step. We have complete visibility of the on hand buffer status allowing us to prioritise by buffer status. This creates a more stable production environment that will ensure we optimise FLOW and ultimately Return on Investment (ROI).
Smart metrics; Signal integrity, Model velocity, decoupling point integrity and outlier analysis, measure performance and give us feedback to drive continuous improvement.
We’ve used MRP For many years and to good effect. The honest among us will admit to using excel spread sheets to supplement and massage information. DDMRP offers an alternative, a methodology that not only functions in complex adaptive systems, but thrives.
We must ask ourselves, Why keep our replenishment decision tied to inaccurate forecasts? Why put more effort into something we know is wrong when we can base them on actual sales orders?