Monotone retargeting (MR) is a novel approach for learning to rank that exploits the fact that there is no need to fit the given preference scores exactly. It is sufficient to fit any scores that induces the correct ordering. MR searches for that order preserving transformation of the target scores that is easier for the regressor to fit. This infinite dimensional search over the the space of all monotonic transformation is reduced to a tractable, prallelizable, convex optimization problem. When paired with Bregman Divergences, Monotone retargeting has favorable statistical and optimization theoretic properties, and excellent empirical performance.