Marketing Mix Modeling – the holy grail of marketing!

Marketing Mix Modeling is use of databases available to a brand in terms of sales data on the one hand and the marketing variables such as advertising, promotions, pricing, season, distribution etc on the other, and finding a statistical relationship between the two. This relationship once established does two things 1. It separates the impact of sales of each of the elements even if they are running concurrently (ah, the magic of software!) and 2. It allows us to optimize each element or at least the important ones which means that we get more out of the marketing spends. The improvement could be as much as 30% or more as a lot of the marketing spend presently may be having no impact whatsoever and so dropping it will not hurt. Or better still pushing it into something that is working will push up sales further.
The difference from what we have done earlier in Indian Marketing circles is that we are modeling actual behavior (sales) and not using mind measures (Brand Tracks) to guess at the behavior and sales numbers tracking that do not break down the impact by element. So MMM helps by quantifying this and therefore is pretty revolutionary for marketing. Secondly, we can validate the model by running it with inputs only on past data and see the “fit”. This means that the model will predict sales and we can then compare it with actual sales. If it is indeed predictive than we know that the model works! No wonder MMM is become the holy grail for marketers world-wide.  Manoj Tadepalli