Non-negative matrix factorization (NMF) is a useful method to extract features from multivariate data, but an important and sometimes neglected concern is that NMF can result in non-unique solutions. In this talk I will show the non-uniqueness problem for NMF and afterwards introduce our sampling algorithm that is able to find the set of feasible solutions for NMF. Furthermore, I will illustrate our algorithm on some real life cancer genomics data and compare it to the most recent state of the art algorithm.