It indicates that the author is aware that women exist other than as love interests or adjuncts to male characters. This strikes me as a non-trivial plus.
Gender equality is certainly a good thing, and so is having properly written, rounded characters. Im just saying that this particular test, without the addition of some normalizing elements at the very least, strikes me as a remarkably arbitrary way to rate such a thing.
In 99% of stories the MC will have to be either a Man or a Woman, and in either case the test is dramatically scewed towards that gender choice. As mentioned above Wall-e doesnt pass, but thats mostly because there are only three characters with real gender (5 if you count the pair of robots with 2 word vocabularies) and many of them never interact. But if you randomly say Auto is female (which would make sense since Ships are traditionally female) then it passes with flying colors. But that doesnt in any way affect the quality of the work, or the relative gender equality demonstrated. Knnn's observation about Lesbian Porn seems relevant to this point as well. Hell, you could write a story that passes with flying colors, that is nothing but a couple of women mudwrestling over a pair of shoes. You could also write one that fails simply because the two female protagonists are focused on taking down the male villain, and so dont stray to other topics (granted it would probably have to be short).
To be clear, Im not saying that there isnt bias in Fiction (though Id like to think its a little less so in Literature than in Hollywood), or that Female characters arent trivialized at times. Im not entire convinced that there isnt a similar number of flat, useless male characters out there that are relegated to base tropes, but Ill fully admit that I dont usually take much note of that sort of thing, so i may just be uninformed.
Now, if you remove any POV characters from the equation, and evaluate it as a ratio of the conversations between two women that do and do not center around a man, Id think you have a more representative metric.