I'll give you a concrete example: "as shot" white balance metadata. Nearly all vendors store this data in the same way. It is a set of 2 or 3 numbers which tells you what WB the camera picked for the capture (assuming you have WB in the camera set to Auto WB). The raw converter reads this piece of metadata so it can provide the same WB starting point when you load the picture. There is nothing secretive, special, inventive, or proprietary about this piece of metadata, since all vendors do it the same way.
The only question is
where in the file this piece of metadata is stored. In the DNG format, there is a tag called AsShotNeutral that holds this information. Cameras that support DNG (e.g., from Casio, Leica, Pentax, and Ricoh) write the WB numbers into this tag. DNG reading software (whether from Adobe or other companies, including free software like dcraw) will simply look at this tag to find the values. So when a new camera that supports DNG hits the market, this software doesn't have to make any changes to read this metadata. It's done in a standardized way. It's like the mailman dropping off the mail in your mailbox, and you always know where to find your mail (in your mailbox!).
Now let's compare to non-DNG formats. In non-DNG formats, the location of this WB tag varies. Sometimes, a camera vendor will change the location from one model to the next. That means for ACR/Lr, I need to go digging into the file to figure out where they've changed the location. It's as if the mailman dropped by to deliver the mail, but each time he drops the mail off in a difference place next to your house. So you need to walk around your house each day to figure out where he's hidden the mail.
Now, that may not seem like a lot of effort, but you multiply the extra few minutes by a lot of cameras, and the way it's done differently across vendors, and it adds up. And keep in mind that effort is spent doing something that is fairly photographically meaningless (finding the as shot WB) instead of doing something that is photographically meaningful -- namely, improving demosaic algorithms, adding new local adjustment tools, improving runtime performance, building new lens profiles, etc. In other words, time spent chasing tags in proprietary formats means time not spent doing things that will actually make photographs look better, improve workflows, speed up the app, fix bugs, etc.
