In some sense, there is no “need” for new translations. If a book is available in English, then… surely that’s enough, isn’t it? There are many works that never get translated at all.
If you want to come at it from a cynical angle, the reason is economic. Different publishers commission translations (or just introductions, or just cover designs!) so that they can sell their own versions of popular works. They’re creating slightly different products to offer in the marketplace. They’re either catering to demand or creating it, depending on how you look at it. Either way, some readers indeed prefer shortened versions, or versions with a more accessible style of English, or versions with lots of scholarly footnotes, or cheap versions, or pretty versions, or–simply—versions that match the brand of other books they already own.
Meanwhile, though, on the supply side, specialists want to re-translate novels. Some of them solicit publishers for funding for their own translation projects, or are happy to be commissioned by them, while some publish their work independently, for reasons that seem more intellectual than economic. Re-translators typically (always?) explain that they don’t like some of the choices the previous translators made—after all, there is hardly ever only one correct way to translate a paragraph. They sometimes also say that previous translators had an “unsuitable” overall approach. Translators all try to make optimal choices on the continuum between “accurate/literal” and “accessible/loose”, some purposely leaning more to one side than the other.
The “accessible” translations may be customized (customised) to a particular English-speaking region of the world (typically the US or the UK), and the slangier they are, the faster they become dated. That’s the most obvious source of a practical “need” for new translations: within every generation, some readers want to read the classics in their own spatio-temporal dialect.
Still, there’s a lot of overlap between translations, since they’re working from the same source. The “major” differences between translations may thus appear rather subtle, rare, or sparsely scattered to non-specialists, or may be confined to certain parts or aspects of the book (for example, dialog in a regional dialect or foreign language or poetry) and thus not be apparent on page 1 or a page chosen at random.
In some sense, the question is beside the point. You may as well ask why Monet painted water lilies so many times (250 canvases over a period of 30 years!). Do we “need” so many impressionistic paintings of water lilies?
Translators, like painters, are creators. In a sense, they’re rewriting the same novel the way Monet repainted the same lilies. There will be some similarities and some differences each time. Not everybody has the same favorite Monet painting, because it’s not obvious which one is “the best”. And could there ever be too many to choose from? The more the merrier, I say.