I have nothing against the practice of randomizing test data. It's a practice like any other and I'm sure some day I might find a valid use case for it. But there are risks that come with test data randomization that shouldn't be ignored