Hyper-dense small cell networks have been recognized as one of the most effective ways to enhance spectrum and energy efficiency. However, performance enhancement can be significantly degraded due to inter-cell interference and traffic loading imbalance. Optimal cooperation has been widely used to tackle these problems by exploring and exploiting cooperative gains. Various potential cooperation opportunities exist among heterogeneous small cells. The denser the small cells are, the more cooperation opportunities a network may have. In this article, we first survey different optimal cooperation techniques from the network-wide and user-centric perspectives. To explore and exploit these cooperative diversity gains, we present a utility maximization framework, where a newly-defined utility function is formulated to jointly consider spectrum and energy efficiency. As a use case of the presented framework, we further propose a distributed interference-aware traffic offloading scheme. Numerical results validate the improved spectrum and energy efficiency of the presented win-win scheme.