The methodologies that we use to study cancer are ever changing. In recent years researchers have aligned pathological evidences alongside classical evolutionary theory and with this comes both opportunities and challenges. Here I discuss a practical challenge that underlies our recent paper "The evolutionary landscape of colorectal tumorigenesis", which was how to acquire and process samples from pre-malignant lesions. I also note a particularly surprising result, which could only have been gleaned through an evolutionary directed analysis on the resulting data.