How do I quantify the impact of various marketing activities?

Marketing mix models are typically based on time series data, usually at the monthly or weekly level. In its basic format we could have advertising expenditures and sales data. Time series regression with lag variables are used to determine if advertising has an impact on sales. More advanced models could include VAR models. These could be useful when we have more variables available. For example, in one study, we had weekly ad awareness and brand perceptions. In such cases we would also want to capture indirect effects.