Prediction of hybrid (e-commerce and traditional) demand evolution and the assessment of its impact on production system robustness
Increased market pressure, sharp competition and globalisation are some of the main challenges faced nowadays by companies, pushing them to continuously evaluate the suitability of their business model and to look for new opportunities. Electronic commerce and more generally Information and Communication Technology (ICT) can provide promising tools for supporting the company traditional business model or even be the drivers for the development of completely new strategies. This thesis focuses on e-commerce considered from the viewpoint of traditional companies thinking about introducing electronic sales channels. A comprehensive methodology is developed, encompassing the issues related to the various phases of the e-commerce strategy establishment: choice of a suitable e-sales channel, qualitative and quantitative characterisation of e-commerce and hybrid demand, identification of the e-commerce challenges, subsequent identification of the production system critical factors. The assessment of e-commerce introduction impact on production system performances implies a deep understanding of the characteristics of the hybrid demand resulting from a multi-channel sales strategy. It also requires to identify the drivers shaping the evolution of e-commerce adoption and the interactions with the demand stemming from traditional channels. Behavioural analysis approaches are developed to model customer preferences and to identify the characteristics of suitable e-sales channels. This information, together with a description of the salient characteristics of the market environment in which the strategy is deployed, is included in a qualitative Decision Support System (DSS) that allows to identify the e-commerce success factors and to determine how they influence e-sales diffusion. The DSS rules are validated and further refined on the basis of experts' opinions, collected through a Delphi study. The qualitative model provided by the DSS constitutes the basis for elaborating a quantitative model of e-commerce demand, exploiting Diffusion Of Innovation (DOI) equations. The impact of e-sales channel introduction is evaluated over a medium to long term horizon because of the time required by the e-commerce demand to reach its steady state potential. Furthermore, the sales channel cannibalisation is modelled. The dynamic interaction among various complex and non linear phenomena implies that a high level of uncertainty characterises the hybrid demand evolution over the considered horizon. For this reason, a scenario planning approach is developed in order to generate the potential outcomes of a given e-strategy and their probability of occurrence. The evaluation of the production system, characterized by the large number of hybrid demand scenarios to be considered, is based on robustness concepts. The proposed production system robustness evaluation framework requires the definition of a vector of Key Performance Indicators (KPI) and is based on the use of asymmetric loss functions. The proposed robustness index allows to efficiently integrate information concerning variability, skewness and kurtosis. It is particularly suitable when many scenarios have to be evaluated and when the use of highly non linear loss functions can result in a skewed distribution of the loss values. In this way the production system critical factors can be identified in order to determine the characteristics of a robust production system, which ensures satisfying performance even in the turbulent e-commerce landscape.
Keywords: conjoint analysis ; e-commerce success factors ; decision support system ; hybrid demand (traditional + e-commerce) modelling and prediction ; robustness evaluation ; analyse conjointe ; facteur de succès de l'e-commerce ; système d'aide à la décision ; modélisation et prédiction de la demande hybride (traditionnelle + e-commerce) ; évaluation de la robustesseThèse École polytechnique fédérale de Lausanne EPFL, n° 4382 (2009)
Section de génie mécanique
Faculté des sciences et techniques de l'ingénieur
Institut de génie mécanique
Laboratoire de gestion et procédés de production
Record created on 2009-04-06, modified on 2016-08-08