This research aims at using forecasting algorithm that predicts the demand that is to be needed on a monthly basis while factoring in occasional inconsistent patterns, seasonality, and non-stationary and cyclical patterns of the data. The prediction is to predict around 3000 SKUs in 19 end markets and since the data is necessary for marketing enhancement and strategies, the Forecasting accuracy must be high. Since market strategies will be based on those predictions and revenue will be lost in the case of an error. Hence, we need to keep in mind that the model is not overfitted and that it wouldn’t give a reasonable accuracy when tested on another SKU. In this study, I will use encrypted data from the organization as such the name SKUs are in numbers instead of names where the trends are there while the region and SKUS will remain undisclosed as well as the numbers wouldn’t be the same. The algorithms used were FBProphet and SARIMA for the given SKUs. They were able to forecast at a MAPE accuracy of 77% and 87% respectively.
Professional Studies (MS)
Department, Program, or Center
Graduate Programs & Research (Dubai)
Al Orbani, Mohammad, "SKU Time Series Forecasting Methods for FMCGs" (2022). Thesis. Rochester Institute of Technology. Accessed from