Many early career researchers working with code, data and analysis assume goodwill is enough to secure authorship – it isn’t.