For many purposes, it isn't so much about the school that gave you the degree as it is your reputation in the field, which is a function of many things (including your alma mater). This can include your advisor, your graduate work, the number and quality of peer-reviewed publications, patents or copyrights you've generated, your research area, scholarships or grants you've been awarded, and so on--basically all of the things you'd include in your CV. The ranking of your PhD program influences your reputation, but it is still one of many factors. Many successful grad students who were accepted to a more highly-regarded program choose to attend a "lesser" program because they believe they can do better research with a particular advisor, for example.
Rankings are an average of many things and may not reflect the research you performed or the education you received. Different programs are known for different things. Some physics programs, for example, are better at mathematical or theoretical astrophysics. Others are known for their work in quantum mechanics or optics. Still others have an excellent reputation in biophysics—or condensed matter physics, or plasma physics, or a particular type of engineering, etc. If your program you got your degree from excels in your area of research, that can often look better to those who are familiar with your field than if you got a degree from a higher-ranked program that did not specialize in your area of research.
I suspect most of the rankings have quite a bit of variation and unreliability. Aside from the annual changes, most rankings probably vary by 20%: a program ranked 30th could vary by 6 rankings (between 24 and 36). The programs in the top 10 likely only vary by 1-2 rankings. This estimate likely varies with field, the entity publishing the rankings, and sample size.
I doubt there's really a cutoff point, however. Imagine if there are 100 biology programs in your country. Differences between the programs ranked 45th and 48th probably aren't particularly significant, but you should expect differences between a program ranked 45th and another ranked 80th! Or suppose your country has 500 programs in biology. You should expect a difference between one ranked 350th and another ranked 500th! However, I would expect little difference between one ranked 450th and one ranked 490th.