Well. BBC. Their title contradicts their article which says: "And although the exact dynamics will depend on many factors, she has shown it takes around 3.5% of the population actively participating in the protests to ensure serious political change." which is more believable.

I think it's an interesting topic. You probably want to read the original research if you want to get better insight. Naturally with all things human it's not an exact science. Also should we look at the population of China or the population of Hong Kong? Presumably if 10% of the population of China was protesting there's some chance of a political change.