Development of a Forecasting Model for Original Equipment Manufacturer (OEM) Components - A Design Science Study at Tetra Pak

Abstract: Title: Development of a Forecasting Model for Original Equipment Manufacturer (OEM) Components. A Design Science Study at Tetra Pak. Authors: Faruk Kodzaga & Giacomo Daniele Supervisor: Jan Olhager, Department of Industrial Management and Logistics Background: The importance of accurate demand forecast is set due to the drivers of short lead-times, just-in-time-deliveries and cost effectiveness linked to demand. Two common methods for forecasting can be used in the context of manufacturing firms, qualitative and quantitative forecasting model. Inaccurate forecasts can result in disruptions of activities throughout the phase of planning, ordering and replenishing of products with high costs. Collaborative forecasting can improve overall supply chain performance, thus increasing overall responsiveness and product availability assurance while achieving optimized inventory. Forecasting plays therefore a vital role for an enterprise to achieve success on the market. Purpose: Develop a forecasting model for Tetra Pak’s OEM Components department based on historical sales data, installation project size and components category. Research Questions: (1) How should the solution be designed in order to fulfill the key properties? (2) What factors except from historical data and installation projects should be included? (3) How do we establish a more secure business environment with the help of the forecasting model? Methodology: The paper is based on a design science study contributing to theory and practice through purposeful design and evaluation. The study also aims to develop theoretical knowledge contributing to solving an improvement problem. Initial As-Is analysis was conducted in order to analyze the performance of the current forecasting model and to receive valuable information from stakeholders at Tetra Pak and Supplier X. The model was initially built using the information regarding the project opportunities pipeline. This first attempt did not deliver the required results but provided valuable analysis and data for the company. Finally, a model based the well-known forecasting method of exponential smoothing, was applied to develop the new forecasting model for Tetra Pak. Conclusion: The new model represents a standardized and reliable method to forecast OEM Components. Improvements have been established when comparing the old model to the new one. Key properties such as Easy-To-Use, Scalable, Reliable and Flexible are represented in the new model. The model is not taking project opportunity pipelines into consideration, but instead is based on time series data. The study has also illustrated the importance of combining qualitative adjustment to the quantitative data obtained from the model such that external factors can be taken into consideration. Keywords: Forecasting Model, OEM, Opportunity Pipeline, Forecasting Accuracy