It sounded like the preponderance of the failures were with the control board and the connections to the inside guts of the drives.
Interesting.
Another aspect of disk drives is that they very cost sensitive (mostly because they are high volume) so engineers will spend a lot of effort saving pennies. They will make the circuit board as small as possible, use less gold flashing on connector pins.
This might impact drive reliability.
Still, it is the spinning media that produces a lot of heat mechanical vibrations and the platters and heads have to very close to operate reliability.
Also, a drive has an automatic/smart defect management system — when a defect is found on the media by the controller and software that area is reassigned to a known good location. So defects are not seen by the user. I think what they are referring to in this article to are gross, catastrophic failures.