Based on my experience dealing with SFPs I highly suspect they looked at their bug tracker and concluded that 13% of the sketch-ass mystery drives were causing 50% of their labor expenditure.

And by "issues" I mean highlighting all the little cases where they had a) coded to spec with no ability to handle out of spec but foreseeable if you're cynical (which the fresh out of school junior engineers who typically wind up handling these things aren't yet) conditions b) failed to code to spec in some arcane way that shouldn't matter if the thing on the other end of the cable isn't questionable.

Of course, the money side of things almost certainly motivated them to see it one way...