Chapter 10 CONCLUSIONS Customization is a continuously growing business trend that aims at providing customers with individualized goods and services. In dynamic business environments, it is even a necessary condition in order to maintain competitive advantage and outpace competitors. The discussion presented in this book concentrates on the challenging case, in which manufacturing enterprises strive for customizing physical goods by taking into account the costs efficiency perspective. Mass customization is an oxymoron that joins both perspectives of product customization and mass production efficiency into one concept. In order to lead mass customization to success, there are some necessary conditions to be satisfied. A distinction should be essentially made between the conditions to be satisfied before shifting to mass customization and those to be maintained and further developed during the pursuit of the strategy. Mass customization induces a high level of product variety, which increases the internal complexity in operations and manufacturing related tasks. The effects of complexity generally arise in the form of hidden costs that trigger efficiency problems. Mass customization involves high production program complexity, high configuration complexity for customers and increasing planning and scheduling complexity. However, mass customization not only increases complexity, but it has some potential for reducing complexity. The pursuit of mass customization drives the reduction of complexity at three main levels, which are: the order taking process, product and inventory. The success of mass customization can only be reached if the complexity increasing factors are adequately managed. Another important issue for the achievement of a successful mass customization is an optimal understanding of customers who should be considered as partners in the value creation process. In the specific context
244 Chapter 10 of mass customization, customer needs are frequently misunderstood. It frequently occurs that customers find out that the selected product configuration does not exactly meet their expectations right after delivery or during consumption. The model making the distinction between the subjective and objective customer needs reveals many insights about the nature of customer requirements in mass customization. The subjective customers needs are the individually realized and articulated requirements, whereas the objective needs are the real ones perceived by a fictive neutral perspective. The subjective needs are explicit and lead to sub-optimal customer satisfaction. But the objective needs are implicit and yield optimal customer satisfaction. The customer needs model provides interesting clues for making optimal decisions concerning the variety steering problem. It suggests the elimination of the over engineered variants that neither correspond to the subjective nor to the objective needs. Another important result is that the product variants that only correspond to the subjective customers needs have to be eliminated. These product variants are especially problematical because they confuse customers. Since customers identify their subjective needs, they may believe that a product alternative is optimal. However, there is another product variant in the solution space that fulfills the objective requirements. The result is that customers select product variants that do not exactly meet their requirements. In addition, the model suggests that customers can better recognize their objective needs if the external complexity faced during the product configuration process is considerably reduced. The customers needs model for mass customization reveals that customers have to be adequately supported during the product selection process in order to make them identify their objective needs. In the webbased mass customization, the customer assistance task is supposed to be ensured by online configuration systems. These software tools enable the customer-supplier interaction and should play an important role in the customer needs elicitation process. However, the examination of the morphological box that includes the main dimensions according to which the configuration systems are developed, suggests that the state of the art configurators are very product oriented. They assume that customers are able to make rational decisions in extensive variety contexts, which is not true because large variety confuses customers. In order to mitigate this main drawback, the extension of configuration systems with an advisory system is necessary. This software system takes over the relevant task of capturing the objective needs and assisting customers during the decision-making process. For the development of the advisory system, two main steps were necessary. The result of the first step is a software system that is based on a hybrid approach including a knowledge base and recommender techniques. This
Conclusions 245 basic system takes over the task of human advisors by asking the customers suitable questions in order to capture their preferences and needs. Then, the advisory system proposes a limited number of product suggestions, from which the customer can choose the appropriate alternative. The second step of the development process deals with the improvement of the system performance with respect to its capability of optimally eliciting the objective customers needs. The logic, according to which the system works, remains the same, which means on the basis of customer dialogs in order to make suitable product suggestions. However, the final implementation includes additional technologies such as a CRM system, web mining technology, and Kansei engineering. These technologies are appropriate solutions for an optimal support of the objective needs elicitation. The advisory system only enables the mitigation of the external complexity problem and does not cope with the internal complexity. A comprehensive software system based on agent technology enables one to additionally address the internal complexity problem. For the development of this system, the products to be customized are assumed to be manufactured on the basis of modular and platform strategies. To each module variant and platform, autonomous rational agents are assigned. In order to work effectively, it is necessary to integrate the module and platform agents into a framework that additionally contains a configuration system, advisory system, target costing agent, auction agent, product constraints agent and validation agent. The advisory system captures customer preferences and maps them to product functionalities. Then the target costing agent determines the product platforms and modules that could form product variants, in which the customer would be interested. The selected module and platform agents negotiate with each other in order to form consistent product variants. The Dutch auction is the most suitable mechanism that supports the reaching agreement process. That is why the corresponding agent is called the Dutch auction agent. It determines the auction functions and communicates them to the platform agents who initiate the variety formation process. Because the product constraints agent has direct access to the product logic of the configuration system, it only allows the formation of consistent product variants. Finally, the validation agent selects the product variants that best correspond to customer requirements. These variants that are displayed to the customer are considered to be successful. A reward mechanism recompenses the module agents that participate in the formation of successful product variants. The notion of a module agent s account enables the implementation of the reward mechanism. Each module agent has an account of money that constantly decreases in the course of time. If the module agent is successful, then it can earn money and extend its life cycle. However, if the module agent does not participate in the formation of successful product variants, then it loses
246 Chapter 10 money. When the module agent s account is exhausted, the module agent has to be considered for eventual elimination. By tracking the agents accounts, managers can evaluate the success of the modules and their suitability in fulfilling customer s requirements. It is important to emphasize that whereas the implementation of an advisory system is appropriate when the external complexity is high and the internal complexity is low, the multiagent based system is an adequate solution when both external and internal complexities are high. The multi-agent based approach has a higher complexity because it has to retrieve data from several autonomous data sources that must be interfaced. For the development of the multi-agent based approach, the modularity of products is an important assumption. Product modularity is a relevant requirement in order to put the mass in mass customization. The development of products around a modular design has many managerial and organizational implications to be considered by the mass customizing enterprise. For instance, product modularity has major influences on the intra-firm and inter-firm levels. Though modularity enables the reduction of the costs of variety, the complexity costs that are induced by variety can be only brought down if the product assortment is based on a few modules. But variety in mass customization can be very large to where a high number of product modules may be required. In order to control the variety-induced complexity in mass customization, key metrics systems including measures that can be calculated on the basis of available data in the company, are appropriate managerial tools. The sub-process analysis is a structured method for the identification of the relevant key metrics. The integration of these key metrics into a comprehensive system is advantageous because key metrics systems have a stronger explanatory power than single measures. A time pattern analysis that aims at identifying the long term, middle term and short term key metrics reveals that the establishment of one single key metrics system is impossible because of some incompatibilities. The integration of these metrics into two different systems is necessary in order to avoid the incompatibility problems. The first system incorporates the variety-sensitive key metrics, whereas the second system integrates the measures that are appropriate for the evaluation of complexity in the information sub-process. The variety-sensitive key metrics system consists of the relevant measures to be used for variety steering. It captures the complexity increasing effects of variety. Managers can use this key metrics system for the evaluation of the internal complexity that is induced by the introduction of new product variants. However, the variety-sensitive key metrics system does not consider the customer s perspective. An extension of this system with the appropriate measures that evaluate to which extent the mass customizer meets the objective customers needs is fundamental.
Conclusions 247 Consequently, the extended final key metrics system assesses the internal complexity inside production and manufacturing-related tasks as well as the extent of the external complexity. Making mass customization work efficiently is certainly not an easy task. Throughout the discussion presented in this book, managerial and information technological tools are elaborated in order to adequately cope with the main problems that are encountered while pursuing the mass customization strategy. However, this research has opened many interesting questions, which can be considered as directions for future research. The following is a list of research topics that could be of interest: 1. 2. 3. 4. The investigation of the necessary conditions for the successful pursuit of mass customization or as many researchers prefer to refer to as success factors provided in this book are based on comprehensive literature research. It would be interesting to empirically validate the identified factors through surveys including mass customizers from many industrial fields. The customer needs model that makes the distinction between the objective and subjective customers needs reveals many important research issues. The advisory system can be considered as a first step towards a comprehensive solution for the problem of providing customers with products that exactly correspond to their real needs. The identification of the objective needs is challenging, especially when transactions are carried out over the Internet. It is a multidisciplinary problem that necessitates contributions and competences from many fields such as business administration, artificial intelligence, computational technology, and consumer psychology. Thus, sophisticated methods are necessary in order to understand the real customers needs, through e.g. the exploitation of the customer s navigation behavior. The interaction systems in mass customization to a great extent determine the success or failure of the whole customization process. The development of adequate tools (e.g. based on key metrics) or check lists for the evaluation of the performance of these software systems could be helpful for managers. In effect, with such tools, it is possible to assess the performance of an existing system with the objective to determine enhancement potentials or to compare between different systems before deciding which one to be developed or implemented. Up to now, there are only a few methods for a systematic development of modular product architectures. Therefore, further work is necessary in this field. Another important issue is the development of techniques for the assessment of the optimality of a modular concept.
248 Chapter 10 5. 6. In order to assess the variety induced complexity in mass customization, it is pointed out that key metrics are suitable approaches. An interesting issue could be to develop key metrics for the evaluation of e.g. the entire product assortment or manufacturing system complexity in mass customization. In this context, a good candidate may be the entropic measure. For instance, it can be analyzed as to how the manufacturing system or product assortment entropy changes when new product variants are introduced in the production program. The key metrics systems that are provided in chapter nine show the different correlations existing between the selected measures. However, it would be advantageous to find the mathematical relationships between the main measures, such as e.g. the differentiation index as a function of component commonality or the process commonality as a function of component commonality, etc. so that it would then be possible to quantify the influence of each key metric on another. Summarizing, one could say that the concept of mass customization is promising but challenging at the same time. It introduces a new business era, in which each person can receive an individualized product at an affordable price. Pine has mentioned in his introduction to the interdisciplinary World Congress on Mass Customization and Personalization that as he wrote his book in 1993, mass customization was the new frontier. However, today, mass customization is the new imperative for businesses. Thus, mass customization is not a buzzword but a practice relevant competitive strategy. In order to make mass customization...as important in the century as Mass Production was in the century (Pine 2003), mass customization must be further supported by academia and practice. Innovative ideas are required in order to strengthen the fundaments for a successful pursuit of this business paradigm.