The company had grown from catalog to online business. Until then, many decisions about the shop were made without test data. A mid-sized company has less traffic than a large corporation. Test the way a corporation does, and you wait weeks for every result.
So we worked differently: small and medium sprints instead of big overhauls, several tests in parallel, testing first where the potential is greatest, and the ROI of the program as the most important target metric. That turned a handful of tests a year into a fixed rhythm.
