Plenty of learners "listen to HSK audio every day" while their listening score stands still. The problem is not volume — it is passive listening: sound washes over you, your brain nods along, nothing sticks. Shadowing fixes exactly that: you must reproduce the audio almost in real time, so your brain is never allowed to coast.
Why shadowing works unusually well for Mandarin
Mandarin has tones — and tones are precisely what non-native ears register as "approximately right". When you shadow, you must reproduce the tone you just heard; if you get it wrong, you hear it instantly. Within weeks your ear starts separating 妈/麻/马/骂 inside fast connected speech — something 100 hours of passive listening often fails to deliver.