Markus Vöth, Uta Herbst, Frank Liess

• Improving the predictive validity of conjoint analysis has been an important research objective for many years. Whereas the majority of attempts have been different approaches to preference modelling, data collection or product presentation, only a few scholars have tried to improve predictive validity by individualising conjoint designs. This comes as a surprise because many markets have observed an augmented demand for customised products and highly heterogeneous customers' preferences. Against this background, the authors develop a conjoint variant based on a completely individualised conjoint design. More concretely, the new approach not only individualises the attributes, but also the attribute levels. The results of a comprehensive empirical study yield a significantly higher validity than existing standardised-level conjoint approaches. Consequently, they help marketers to gain deeper insights into their customers' preferences.