Information overload is a well-known problem of conjoint choice models when respondents have to evaluate a large number of attributes and/or attribute levels. In this paper we develop an alternative conjoint modelling approach, called pairwise conjoint analysis. It differs from conventional conjoint choice and preference models in that the attributes of choice alternatives or choice contexts are not varied simultaneously, but in pairs. Properties of the design strategy are discussed. The new approach is illustrated by using activity engagement choice as an example.