Yes the media is failing. Floppy disk have a finite usage/storage time. Say the disk is 20 years old and only been read about 20 or so times, that would equate to a 1 year disk read about 100 or more times IIRC....
As for just sitting: there is a lot of electromagnetic energy today that slowly helps ruin floppies: lighting strikes, cell phones and towers, cordless phones, etc and pollutants in the air too...

added: I have fairly good luck with Sony, Maxell and TDK 3.5" floppies and also 3M and Memorex branded 5.25". They all came with lifetime warranties (which means not much) but seem to last longer than generic ones (Office Depot, etc).
Am sure others have similar stories about what brands they have used that were good or bad 😀

I've heard that Sony floppy disks aren't actually that good. At least the ones from the late 90s. But then again, were there any from the late 90s that weren't terrible?

I just opened a sealed pack of BASF disks I've had saved. They are are stamped "assembled in U.S.A" and have a label that says "135 tpi certified". What does that mean? It does not have a date printed on the packaging, so I don't know in what year they were made.

135 Tracks Per Inch, certified for High Density 1.44Mb formatted PC (or 2MB max) IIRC

Why does it need to be certified? It either is or it isn't. Is this important information or just marketing speak?

Unless the disks are physically degraded or degrading (i.e. oxide layer dettaching from the substrate, heat damage, mold growth, physical damage, etc), I would have imagined that practically anything to do with magnetization/demagnetization should be fixable by either formating or running through a bulk eraser .

Then again, I am no expert in these matters.

Kahenraz wrote on 2022-07-10, 01:53:

Why does it need to be certified? It either is or it isn't. Is this important information or just marketing speak?

It probably means that it successfully passed whatever testing post-manufacture testing process was in place . Without more details as to testing methodology, it could mean anything from an automated test write/read process for every disk, to having that test run on random samples or to having some dude eyeball/smell the disks to make sure they seemed "OK".

aren't there disks with magnetic shielding?

It could be many things. Drive heads caught a particle (very dangerous and annoying), disk grew mold on a certain location, a magnet passed by and ruined a few clusters... I'd start first with a visual inspection of the surface on both sides under a strong light.. this will show any obvious liquid/dirt/kinks/scratches to the surface. If it's all clear, then a few full erase passes are usually enough to bring the disk back to life (around 3 passes from my experience).. after that a full unconditional format usually succeeds with 0 bad sectors.

Is an erase pass something different than running format a couple of times?

Yes it's different. It basically sectors the disk to 1 sector each track (so 1 sector is full 360°), thus wiping the typical 15-sector metadata areas too. I do this in ImageDisk (a really nice imaging tool under DOS).