Improving Forecasting for the Aftermarket through Big Data - A Case Study at Volvo Group Master of

Abstract

The aftermarket brings profitable advantages to manufacturing companies by providing value adding services for the costumers, where accurate forecasting is essential in order to achieve a smooth material flow of spare parts. Concurrent forecasting methods are mainly based on historical demand with mathematical methods that can trace back to the 40s. Today’s business setting where companies handle large amounts of data, also known as big data, provides new and innovative improvement possibilities. Naturally, spare parts demand is difficult to monitory since occurrence of failure is unpredictable. Forecasting based on big data might be a way to achieve high uptime for the costumer by having the spare parts at the right place in the right time, addressing the challenge of having high availability at a low cost. The purpose of this thesis where to investigate opportunities for improving forecast accuracy of spare parts in the aftermarket of automotive companies, by exploiting big data created downstream the supply chain. A case study was conducted at Volvo Group, which is a market leading automotive manufacturer. Two distinct research questions were identified in order to fulfill the purpose of the thesis; 1) which data created in an aftermarket supply chain has potential to increase the accuracy in predicting demand of spare parts, and 2) how can the identified data support planning processes of automotive companies for predicting demand of spare parts in the aftermarket. Volvo Groups aftermarket supply chain was scrutinized in order to provide insight for potential opportunities of big data utilization. In combination with a theoretical framework which provides academic insight into concurrent research the defined research questions were answered. The result of this thesis is a framework which describes two dimensions for succeeding in implementing big data in the planning process of spare parts. The dimensions are presented through matrices which gives an illustrative view on how big data is utilized in the planning process of spare parts. The first dimension describes the level of sophistication to translate big data into a demand. The second dimension describes the level of integration needed towards the department responsible for forecasting. In conclusion, this framework has the potential to guide not only the automotive industry but also other industries that has access to big data which correlates to the probability that a spare part fails.