There are much more efficient systems for only carrying stuff into space and dropping it there. The Space Shuttle example makes no sense if this is what you care about.
I don't know about the Dragon, but, like before, if the Dragon's weight is the thing you really want out of the atmosphere, then yes, but if you only care about the supplies inside, then no.
>And even then, it's apples to oranges, because Space Shuttle was a manned spacecraft which brings millions other requirements into place.
What requirements? But yes, it's fair to break out "human" vs "inanimate" kg of payload. Do SpaceX products for carrying humans perform poorly when comparing to alternate ways of launching human payloads? Do SpaceX products for carrying inanimate matter perform poorly when compared to the same?
When you talk about launch costs, you care about the cost to get “the stuff you really want in orbit”.
The formal term for that stuff is “payload”. So cost per kg of payload is the relevant figure of merit.
The mass of the space shuttle is not payload. It is a means of delivering payload. As such, you do not get credit or establish launch efficiency by launching it.
It is not inefficiency to launch without carrying the space shuttle; just the opposite.