Swahili dictionaries have been on the market for over a century. New publications often take into account works that have been already published and base the list of headwords on older dictionaries. While this is widely accepted practice, nowadays also new, computer-based activities may no longer be ignored. Analysis presented in this paper aims at detecting the differences and similarities among Swahili bilingual published dictionaries on the macrostructural level, and at identifying possible trends in Swahili dictionary compilation. It also takes into account corpus-based methodology and points to its superiority over intuition-based methods in dictionary compilation.