Cramer - Rao Inequality:
Every unbiased estimator has a variance greater than or equal to a lower bound called the Cramer - Rao lower bound. If the variance of an unbiased estimator achieves the Cramer - Rao lower bound, then that estimator is a minimum variance unbiased estimator, or, simply, efficient estimator (see Efficiency).
To use the Cramer - Rao inequality in a particular situation you have to know the expression for the likelihood function and the likelihood function should satisfy some regularity conditions (otherwise the Cramer - Rao inequality is not applicable).