The authors propose a Plackett-Burman experimental design to rearrange profiles in blocks in choice-based conjoint analysis as an alternative technique for measuring preferences that accommodate large numbers of options. Although in choice-based conjoint analysis profiles are usually randomly organised in blocks, we propose a manually statistical arrangement because its design takes into account all the factors in the same number and equally distributed, and because it allows us to determine the degree of resolution in advance. Plackett-Burman can be an efficient design if we consider a trade-off between the number of stimuli in each choice set and the number of choice sets used in the assessment process. To illustrate its uses we describe an empirical application measuring preferences for shocking warning messages on cigarette packages described in 11 pictures and we estimate the distribution of these on tobacco products to optimise impact on teenagers.