Integrating Conjoint Analysis and Engineering Design

The authors describe an approach to integrating conjoint analysis and engineering design. Developed in collaboration with General Motors, the proposed approach can be used for designing products comprising several interconnected subsystems. The proposed approach has four key steps: (i) representation, (ii) linking, (iii) constraining, and (iv) costing. Representation concerns the depiction of a product in ways that are meaningful to consumers and to engineers. Linking refers to the mapping of engineering attributes onto the conjoint attributes. Constraining concerns the restriction of designs to those that are feasible from an engineering standpoint; these constraints are propagated, using the linking procedures, to the conjoint designs. Costing refers to the methods used to estimate fixed and variables costs associated with module designs. The linking procedure then maps these costs onto the product concepts evaluated by consumers, and allows the implementation of a profit-maximizing objective in the selection of a product or product line. The proposed procedure allows an assessment of consumer and market response to engineering choices. It enables product designers to optimize the tradeoff between increased customization and increased standardization via the use of common platforms and modules. It allows for the evaluation of the economics of product variety, and the cannibalization in sales due the use of common platforms in product-line designs. It makes possible the imposition of technological and cost constraints both in conjoint simulations and in algorithms for identifying optimal products and product lines. Finally, it permits feasibility assessments of alternative technologies, helps prioritize R&D projects, and facilitates the evaluation of cost-cutting efforts on the market performance of products. We illustrate aspects of the methodology using as a running example the design of a line of mid-sized cars by General Motors. We also summarize the key learnings that have accumulated in over a decade of use of these methods by the company.