I wonder whether this is because the subscription method was overpriced compared to the number of articles they actually accessed or if academics are now just thinking twice about whether they really need to access a particular article if it’s not easy and “free”. I’ve certainly downloaded articles I never actually got around to reading.
It’s still wild that universities don’t just en masse refuse to use the for-profit journals. They services they provide could easily be managed and funded by a university consortium. They just need to actually make the leap.