SCMG 305: Please help with making a response to this post. If possible have an open-ended question at the end. Thanks!

When conducting a time-series forecast model, consideration must be made for seasonal and irregular inputs. Depending on the type of product sold, these two factors could adversely affect the forecast model. An example could be living in a location with a climate that interacts with all four seasons. Sales for goods typically sold in the winter months will be much higher than in the summer.

The two methods most commonly utilized is the additive and multiplicative forms. Both with intended purposes, the additive form allows you to create an analysis with minimal deviation associated to seasonality and irregularities. The multiplicative form is the inverse, causing you to work the algorithm to consider the deviations caused by both seasonality and irregularities.

At my job, we do not use doctrinal terms however, the approach to forecasting demand is essentially the same as the literature explains this week. I believe it is very accurate as we sit around 90% gross fill rate for ~13,000 unique items. This is indicative of an aggressive approach to reshuffling our inventory and rolling out obsolete or unused product taking up warehouse space. Our team does weekly scrubs of our inventory and codes them appropriately with consideration of the trend. We currently use a Power BI model to integrate our historical data with our distribution team. This allows us to make stocking decisions based off order to ship time. It is broken into two phases; how long does it take from wholesale to us and how long does it take us to move product from our warehouse to the customer. If the customer can get it from wholesale with a similar customer wait time, we typically do not bother with that item even if it has high demand.

With that being said, I believe our forecasting model is fairly accurate, but it can definitely be much better. The lack of continuity in the military will always be the reason we cannot operate at a high level. Most units rotate Marines ever 2-3 years, not allowing enough time for any individual to make consistent impact and changes based off data collection.