Matching software can usually be set to match at different levels e.g. word level, sentence level etc and it can be set to ignore quotations or ignore a reference list or not, so there are many alternatives.

So for example if I write the following paragraph:

When the article is identical, it’s not possible you mean that as little as 60% is matching, I assume.

I could get a score of 60% matched to your comment, but it’s saying a different thing – albeit on the same topic, but many of the words match.

It takes judgement to say whether or not it’s plagiarism. I’m not sure that’s the best possible example, but I hope it makes the point.

If you have blocks of identical text then you need to check to see if those are plagiarised or not. e.g. they could be legitimately copied as properly acknowledged examples or properly cited quotations.