Dollar per kg to orbit of the Space Shuttle was pretty respectable - if you count the mass of the orbiter as part of the payload. If you don't, it's absolutely ruinous. (~$5,000/kg with orbiter, $20,000/kg for what it could carry in the payload bay, 2020ish dollars - yes the orbiter outweighed its payload 3:1)

Look, the Saturn 1B - the man-rated, 20,000kg-class, safe predecessor to the Saturn 5 - cost ~$330m 2020 dollars per launch. The Space Shuttle could carry 24,000kg to orbit at a marginal cost per launch of somewhere between 500 and 700 million dollars. (This ignores all the costs attributed to the Space Shuttle program that were incurred even if a launch never occurred)

The Saturn V, of course, could put 140,000kg into LEO for ~1,250 million 2020 dollars.

So for half the price of a shuttle launch, you can put the 80% the mass into orbit on a Saturn 1B. Or for double the price of a shuttle launch, you can put 5.8 times as much mass into orbit on a Saturn V. Or double the shuttle's orbital payload mass - but on a trans-lunar injection trajectory instead.

And this is all comparing the shuttle to the technology of a decade and a half earlier.

Yes, a Shuttle Two - get rid of the reinforced carbon-carbon (hell, maybe even move to an ablative heatshield), ditch the SRBs, ditch the wings, lose the cross-range capability and maybe even move the fuel tank internal - might have been viable. But the closest thing to that (the X-33 program/VentureStar) never got the funding it needed, and even then might have been a bit too ambitious. Time will tell if an affordable SSTO ends up ever happening, but I'd bet on things like SpaceX's Falcon SuperHeavy/Starship (fully reusable TSTO craft) being the real successes.

DuskStar explained well why the Shuttle was not cost effective in terms of payload per launch cost. I want to mention that there was a tiny minority of missions where this isn't a fair comparison: where they were returning satellites from orbit to Earth. In those cases, the shuttle was not just deadweight that needed to pay for itself through reusability; it was a necessary part of returning the satellite.

This ability was used a grand total of just four times.

https://space.stackexchange.com/questions/15094/what-satelli...

It was not remotely worth making the rest of the launch system more expensive by an order of magnitude. Better to pay for a new satellite when they break, or (in the case of returning scientific samples) to design a sample-return function into the satellite.