Compilers are, in essence, translators that take input in one language and produce output in another. For example, Eiffel Software's compiler takes Eiffel-language input and produces C. GCC for Intel reads C-language input and produces x86 assembly. The GAS assembler for Intel takes x86 assembly and produces x86 object code. All three of these things are technically compilers.
Regardless of format, the input read by a compiler is called the source and the output is called the target. The latter term is taken from one of its definitions, "intended result."
The majority of compilers are designed to produce assembly or object code for a particular processor or architecture. Because of that, target is often used to refer to the architecture itself rather than the output format.
The target of a compiler does not need to be the same as the architecture where it runs, and in instances where that happens, the program is called a cross-compiler. (For example, GCC can be built to run on x86 systems to compile C into ARM assembly.)
Additionally, there are single compilers capable of producing output for different targets depending on input such as switches on the command line. These are called multi-target compilers.