State-of-the-art managed runtimes apply aggressive optimizations often based on the speculative assumption that programs have low variability. However, the behaviour of programs may evolve over time and literature shows that programs are likely to experience different execution “phases”, i.e. intervals of time displaying an homogeneous behaviour. Taking such phases into account may further improve performance when applied to phase-sensitive optimizations such as lookup caches, that may contain entries that are phase-specific. In this paper, we introduce Phase-based splitting, an experimental compiler optimization that utilizes phase insight to guide monomorphization based on splitting. Preliminary results show speedups ranging from 10 to 20% on average, peaking up to 47.6% at phase granularity.