Statistical modeling in a wider frame: A marketing primer

Issue title: Statistical Modeling in Marketing and Advertising Research

Article type: Research Article

Authors: Mandel, Igor

Affiliations: Telmar Inc., 470 Park Avenue, New York, NY, USA. E-mail: imandel@telmar.com

Abstract: General limitations of traditional statistical modeling are considered and its insufficient relevance in explaining complex phenomena of social life is shown. The extended scheme of modeling – Compound Analytical Platform (CAP) – is briefly considered, with main elements such as classifier of typical statistical situations, statistical analysis and sociophysics as a way to reduce complex descriptions, simulation (including agent based models) as a key mechanism oriented on searching for real causes of the processes. These principles are applied to analyze one of the most popular tools in marketing – marketing mix models (MMM). Critics of the typical model's versions from marketing and statistical points of view are presented and way to answer business questions beyond MMM described based on the proposed analytical platform. It is shown, that even in marketing, where statistical models play very important role and look reasonable, there are many methodological and technical difficulties to be overcome only in wider frame of modeling.

Keywords: Statistics, statistical modeling, sociophysics, mediaphysics, simulation, agent based models, sociology, marketing mix models, regression

DOI: 10.3233/MAS-2009-0125

Journal: Model Assisted Statistics and Applications, vol. 4, no. 3, pp. 215-237, 2009