Marketing Analytics – why decisions based on analytic analysis are better?

People overestimate the probabilities of events and often become overconfident. The quality of the information available to them from external sources, such as media, real life experiences as they occur, or heuristic may give a false impression of the events. The problem with overconfidence, especially in the business environment, is that it often leads to the boiled frog syndrome. Bias is a cognitive phenomenon and has its roots in psychology (availability heuristic). Analytical methods for data processing and decision making are used to remove bias from the process of information assessment.

Analytical tools are used by managers help them to get additional insights into problems and change the way they think about their business. With the use of analytical research methods managers usually explore more options and work with ‘what if’ scenarios a lot more frequently than without these methods. Reaching group consensus and supporting group decisions is usually easier with the presence of hard analytical data. The decisions supported by analytical data are also more accurate and consistent.

Marketing analytics require sets of data which are usually collected in the process of marketing research. It is easier to conduct the research itself than to understand the consequences of the information derived from it. Asking the right research question at the very beginning of the research process is therefore vital. In the world of business management, structured and stylised representations of reality in the form of analytical models are easier to deal with and understand than the messy reality itself. No model is true but somewhat useful. It is not models that make decisions but managers.

 How to build a model useful for business decision making:

• Specify relationships between variables and how do they interact. Ask yourself if you need all of the possible variables or only the relevant ones?
• Compare secondary statistical data with primary data (data from your research) and ask yourself if it makes sense?
• Validate your data and check how significant the variables are (r-square, model fit, correct signs, T-test)?
• Check your analytical model against single and multiple objectives (for example: short and long terms objectives).

 References

• Tuk, M., 2012. Marketing Analytics, Marketing Analytics. Imperial College London, unpublished.
• Malhotra, K. N. and Birks, F.D., 2000. Marketing Research. An applied approach. European Edition. London: Pearson

Written by Michael Pawlicki