Education technology (edtech) is increasingly prevalent in classrooms, yet 85% of the technologies currently implemented are a bad fit for the school, or are poorly implemented. This is especially problematic for poorly funded schools typically seen in minority communities. To address this limitation, the EdTech Evidence Exchange has collected survey responses to characterize the contexts in which technologies are being implemented. Variable selection via penalized regression and unsupervised methods extracts a subset of the factors that are the most informative in characterizing schools’ contexts. This subset is then used to fit a Gaussian Mixture Model to create soft clusters of schools with similar contexts. A new feature, “Schools Like Mine,” combines soft classification and Euclidean distance to identify and rank schools by similarity. This research will hopefully reduce the likelihood of failed software investments.