Marketing Mix Model

It seems like everywhere we look, there is an advertisement for some different product or service. The question for most companies is not, how or where are we going to advertise? Rather, the question becomes, how are we going to limit and optimize our marketing efforts? Whether it is advertising in the newspaper, television, radio, internet, or countless other media outlets, marketers must decide which combination of media channels to use in order to promote their company's products in the most efficient and effective way. Determining a combination of media channels is the essence of Marketing Mix Modeling. This paper will focus on the specific inputs and processing tools required in developing a Marketing Mix Model and the results expected from the model.

Marketing Mix Issue/Problem

The first step in developing any model is to identify a specific problem or issue to be solved. An example of a question posed by marketers of a company is how can my company optimize media spending given a marketing budget of $XXXXXX?

Inputs

In order to conduct the appropriate analyses, the company's available data need to be explored and examined to determine their worth in the analysis. An important issue to consider is, in what manner and how often is the data collected? If we are aiming to optimize media spending, then we will need data that is relevant to our goal. For example, we will need data on past sales and advertising methods. The data collection process will, most likely, involve identifying someone within the company or organization with a working knowledge of the data. This person will aid in interpreting the inputs of the model. Unfortunately once the data has been collected and studied, there is not an easy point and click answer to the issue or problem.

Processing Tools

Complex analytical tools and statistical software are used to relate, for example, sales to other variables in the collected data that will help to determine a marketing mix. Modeling techniques include, but are not limited to, multiple regression analysis, logistic regression, neural net analysis, and genetic algorithm analysis.

Results

The desired output of a Marketing Mix Model is, naturally, the "optimal" or "best" solution that will answer the posed question. For example, from the proposed question concerning the optimization of media spending, the desired output would be in the form of a ratio to recommend to the end-user the combination of media sources to use to optimize sales or returns. Unfortunately, such a solution is not always possible. There may not be just one optimal solution telling the marketer exactly where and how to allocate his or her budget resources. On the other hand, the resulting model may be used simply as a basis for making future marketing decisions.

Reference

“Marketing Mix Modeling” www.teasley.net