Stochastic models have been used extensively for modelling phenotypic evolution. Studying the evolution of phenotypes offers an opportunity to provide insight into many important macroevolution questions. However, it is challenging to analyze large phenotypic data because observations are highly correlated due to the fact that species are related to each other according to an evolutionary tree. This correlation makes generic likelihood computation methods which use matrix inversion unable to scale. It also invalidates the consistency property of Maximum likelihood estimator in many scenarios. In this talk, I will present recent developments which cut directly to the heart of these challenges.