Nikolaus Franke (Vienna University of Economics and BA)
Heribert Reisinger (University of Vienna, Austria)

Vienna University of Economics and BA and
University of Vienna, Austria

Cluster analysis is widely used in management and marketing research and practice. One important criticism is that the clustering algorithm is so powerful that it will provide clusters even if no meaningful groups are embedded in the sample (Barney and Hoskisson 1990) – that is even if the elements of the sample are actually too heterogeneous to be grouped the method delivers a small number of seemingly homogeneous clusters on request.
In our study, we ask how much variance is typically left unexplained. The term “dark side” refers to the fact that a considerable fraction of heterogeneous customer need may indeed currently be going unserved by standard commercial products on offer in the marketplace. This constitutes a opportunity for (new) firms employing toolkits for user innovation and design.