Adding to the good answer of Curd: notice that from a signal (AC) point of view, \$R_C\$ and \$R_L\$ are in parallel, since the supply act like a short for AC. This means that, if \$R_L\$ is your "real load", it must have a much lower impedance than \$R_C\$, which is used only to set the quiescent point of the BJT.
Moreover, in some cases you have a load that needs a DC component: take as a simple example an LED used for lighting, assuming the BJT is used as an amplifier and not as a switch, for instance using it as a current regulator (current sink in this case; inefficient but simple -- it's essentially your first schematic).
Note however that the load is not ground referenced, unless you use a PNP BJT as a current source on the high side of the load (i.e. connected "above"). Some loads may need one of their terminals connected to ground, so this is a distinct disadvantage in this case.
To sum up, there are different design decisions to be made when choosing where to place the load in an amplifier. The fact is that in most textbooks they stick to the very basic case of an AC coupled, class A, small signal amplifier for low frequency (the second schematic you posted). That's just to keep thing simple for the learners.