The control helps rule out any other factors that might be affecting the results of your experiment, and is also used as a comparison too.
For example, doctors testing drugs use a placebo on some patients, and the real thing on others. They can then tell if patients are just "making it up" as it were, because those with the placebo, obviously, had no actual drug.
So yes, you wouldn't be using the variable (called a negative control I think), say if you were testing how an enzyme catalyses a reaction, you'd carry out a reaction without any of the enzyme (+ repeats, obviously).