> If they don't know what's wrong, how did they know what to fix?

Pure speculation, but for example if they did design a process to enrich their pasta with iron, tested it initially, but then never tested it again (who cares if there's iron in pasta...), and then it stopped working for some reason (or possibly never worked reliably in the first place). And then instead of trying to find out what made this process not work, they just came up with a new process that did work.

As I said, it's speculation.

Another option is of course "I perfectly know, but I cannot tell you, because if I did, we'd be in legal trouble".