With a great pleasure I have read series of articles about conjoint analysis in transactional projects. One question though has remained unaswered: what is the best (possibly scientific) method to choose the ca. 30 factors/levels' combinations (r

With a great pleasure I have read series of articles about conjoint analysis in transactional projects. One question though has remained unaswered: what is the best (possibly scientific) method to choose the ca. 30 factors/levels' combinations (runs)? With few factors on multiple level one will end up with hundreds of runs, while recommendation is to stick up to mx 30. Could anyone recommend to me the selection method?