I don't think he's contradicting himself. His scholar opinion is about the form of Plato's work (i.e. the order it should be read), and not about its content. I think he recomended people to avoid the second, as following the first is very helpful for the beginner who sees themself without knowing where to start, even if the suggested order is very opinionated.
Second, he says that the dialogues should be examined as a whole. Well, how would we understand the entire corpus of dialogues without first understanding each individual dialogue on its own, and vice versa? An individual dialogue is easier to understand on its own, that's probably where we should start.