Strategic process flexibility under lifecycle demand

Abstract

The better utilization of production resources has been a concern in the automotive industry over decades. Though costly itself, increasing process flexibility across the entire production network is seen as a particularly efficient approach for managing the mismatch between available production capacities and time-varying product demands. On the demand side model-based support for flexibility strategies has mainly been derived from improved absorption of instabilities brought about by demand uncertainty. In this paper we include a second source of demand dynamics, namely fluctuating demand along the product lifecycles, which has a non-negligible impact on both strategic planning and evaluation of flexibility in a production network. Built on a stochastic programming approach and numerical studies, our analysis suggests that the benefits of flexible configurations might be substantially misjudged if product lifecycles are not considered. Specifically, the benefit of a (semi-)flexible configuration hinges critically on the lifecycle phase of the different products manufactured in the network. However, our results also indicate that prominent flexibility strategies like chaining plants remain robust even when lifecycles are included in the analysis.

Introduction

Growing uncertainty about future demand and declining capacity usage are major challenges for manufacturing strategy in many industries. Particularly in the automotive industry, increasing product differentiation, higher competition due to globalized markets and distinct product lifecycles have led to a sharp decline in capacity usage over the past decades as illustrated in Fig. 1 (PricewaterhouseCoopers, 2006). The left part of the figure shows that global vehicle assembly has not been able to keep pace with the growing capacity since 1990 (measured in million [m] units). As a result both global excess capacity and capacity utilization evolved in an unfavorable manner in the last 15 years as depicted in the right part.

In the capital intensive automotive industry the resulting low capacity utilizations entail severe economic consequences for the car manufacturers. Due to high fixed costs, operating margins closely track capacity utilization (Goldman Sachs, 2004). Not surprisingly, a major strategic concern has been a better resource utilization. The two seemingly obvious leverages to better matching production capacity with demand (and hence, maximizing capacity utilization) are increasing demand and/or downsizing capacities. However, in a dynamic and stochastic demand environment, both of these direct approaches have inherent drawbacks which are related to the difficulties of adjusting technical capacities in the short term. On the demand side, boosting sales to better utilize available production capacities has been a common (and, as it turned out, highly adverse) countermeasure many manufacturers have pursued by offering substantial price discounts and special features without extra charge. The resulting price wars finally harmed the profitability of all players (Holweg and Pil, 2004). On the capacity side, downsizing production capacities entails the risk of lost sales and hence reduced turnover in the light of demand fluctuations. This is because technical capacities, unlike, e.g., manpower or inventories, cannot be adjusted in the short and medium term.

Managers in the automotive industry are aware of these challenges and started to rethink their manufacturing strategies. Specifically, several car manufacturers recognized the ability of flexible plants to cope with demand fluctuations and reduce slack capacities while ensuring high service levels. In this paper, we address particularly such investments in process (or mix) flexibility which allows for jointly manufacturing multiple products in a single plant in order to shift production within the network. Process flexibility is therefore linked to volume flexibility, which describes the ability to adjust the output volume of a product without incurring large costs (Jack and Raturi, 2002). Recent examples from the automotive industry show this new development. For instance, as a response to the perpetual demand fluctuations of its products, the Chrysler Group invests into transforming existing plants into “world-class, flexible manufacturing facilities”. Recently announced investments at Sterling Heights Assembly Plant and Sterling Stamping Plant will make the plants capable of producing multiple products (DaimlerChrysler Times, 2004, DaimlerChrysler Times, 2005). Likewise, Volkswagen regards flexible plants and the resulting capability of shifting production within and between plants as an important component of its risk management systems (Volkswagen, 2002).

The selection of a flexible configuration and the corresponding investments in flexible plants are usually taken for a planning horizon in the range of several years. The planning objective of the strategic level is therefore to find a network configuration that meets the requirements of the operational level in terms of production capacity to satisfy demand. Academic research on this issue commonly addresses the influence of uncertainty in product demand on the benefits and the optimal design of process flexibility. However, beside stochastic influences, the evolution of product demand in the automotive industry is heavily affected by distinct product lifecycles. These have substantially shortened on average in the last 20 years and the large differences between average demand in distinct stages of a lifecycle have become a major concern for manufacturing strategy in the automotive industry (Holweg and Pil, 2004). Typically, lifecycles of different products are not in the same phase of a lifecycle curve. One may think of situations where some products are in their ramp-up phase, while production of others begins to ramp down and successor products will be introduced soon. For car manufacturers, it is essential to understand the interaction between lifecycles and manufacturing flexibility. In this paper we address this issue explicitly.

The paper is organized as follows. After reviewing the literature on process flexibility in Section 2 we develop a formal flexibility planning model that captures both demand uncertainty and lifecycle dynamics in the framework of flexibility provision on a network level (Section 3). Subsequently, we present numerical studies inspired by planning practices in the automotive industry (Section 4). The paper concludes in Section 5 with a short summary and a discussion of our results.