I've tried to explain this idea several times, with little or no success. But I think I've found a better explanation, so I'll try again.
To begin with, "neutrality" is variously defined these days. It's becoming common to define it in terms of what I call "initial neutrality," not the classical "outcome neutrality." So let's look at those two views:
Initial Neutrality. Everyone starts out relatively even in terms of how familiar the auxlang seems to them or how it represents their language compared to other languages.
Outcome Neutrality. Regardless of the auxlang's sources and structure, learners are likely to attain near-native fluency in roughly the same time, and more quickly and easily than they could achieve such fluency in a natlang.
I tend toward Outcome Neutrality: I don't care about an auxlang's sources, just about how quickly and easily I can achieve near-native fluency. I've encountered stories of people who couldn't communicate confidently in any but their native language--and some auxlang. Struggling with a non-native language made them seem inarticulate, but in their own language and auxlang they were eloquent.
The problem is that neutrality (by any definition) varies by individual, and unlike the Initial Neutrality people, I see no way of avoiding this problem: some people will never achieve near-native fluency in certain languages. In Esperanto, such people are called "Eternaj Komencantoj," and they form the next part of the argument.