I don't think this is just a limitation of the analogy, but it is the key point: for any structure where we know the process that forms it (like waterfalls or snow flakes), we can have some informed opinion about the chances of it happening elsewhere in the universe, even if we have only observed one instance of it happening.

However, for structures where we have no idea of the process by which they are formed (like cells), we can't have any opinion on how likely they are to happen elsewhere, unless we see more examples.