Pipeline is the method by which instead of making cpu as a single unit,its devided into various functional units with each unit carrying different tasks
for ex, in instruction pipeline (IF,ID,EX,MA,WB) ,these subtasks are called stages.
The goal of pipeline is to acheive CPI=1 ie in every cycle one instruction should get completed.This was not earlier possible because for ex an instruction takes 5 cycles to complete since cpu is single unit, you have to wait for 5 cycles for one instruction to complete..so instructions will complete at 5,10,15,20..
Now with pipeline, as soon as one instruction moves from IF to ID stage,since functional unit responsible for IF is now free ,we can fetch next instruction therefore instructions are entering at pipeline @1 instructions/cycle and leaving @1 instructions/cycle
instructions 'll complete at 5,6,7,8,...
I'll say it'll give an illusion that processor has become faster,whereas it's the same but we are able to get more throughput(no of instructions in unit time) which is what we care about.