Aims to maximize expected returns at a given risk level and formalizes the concept of diversification in investment portfolios

to decrease risk. 

It can calculate the risk and reward based on weekly sales value data (i.e. ratio between incremental and total sales) or on attributes data (i.e. awareness, consideration, etc)

Both models generate portfolio optimisation outputs but they are the best complement to each other as MMM adds info at media channel level and MPT creates historical betas.