Fixing Hubble was one of the most complex operations the Shuttle ever did. Although only the shuttle could do things like repair and retrieval, the large majority of the satellites that ever entered the Shuttle cargo bay were simply launched and released.

More specifically, of the 105 satellites the Shuttle interacted with, only 4 were repaired-and-redeployed (including Hubble) and only 5 were recovered back to Earth.

http://rammb.cira.colostate.edu/dev/hillger/Shuttle-related....

> The shuttle was supposed to be cheaper than multiple manned rockets, but not cheaper than regular satellite launches.

Do you have cites to support this? It's not consistent with my impression of the expectations for the Shuttle when it first started flying. Your other comments here give the strong impression that you are arguing based on your intuitive knowledge of spaceflight costs that no reasonable person could have thought launching dumb satellites on the Shuttle would have been cost effective. However, I think that's hindsight, and in fact many people at the time expected the Shuttle to cheaply deploy basic satellites, and that this was part of its official mandate. Indeed, Wikipedia says

> the plan was for all future U.S. space launches—space station, Air Force, commercial satellites, and scientific research—to use only the space shuttle. Most other expendable boosters would be phased out.

https://en.wikipedia.org/wiki/Space_Shuttle_design_process

EDIT: And from another link in this thread (written at the time):

> NASA plans to terminate all throw-away rocket launches as soon as the shuttle is working.

http://www.iasa-intl.com/folders/shuttle/GoodbyeColumbia.htm...