Optimizer for insertion of adverts in printed media using mean-variance model

• Grant number: 02/12864-5
• Support Opportunities: Research Grants - Innovative Research in Small Business - PIPE
• Start date: April 01, 2003
• End date: August 31, 2006
• Field of knowledge: Applied Social Sciences - Communications - Public Relations and Propaganda
• Principal Investigator: Pedro Jesus Fernandez
• Grantee: Pedro Jesus Fernandez
• Company: Ipsos Novaction Brasil Ltda (Ipsos)
• City: São Paulo

Abstract

The objective of this project is the development of an optimization system for printed media based on the model of generalized mean-variance (known also as the generalized Markowitz Model), a celebrated paradigm in the financial market, but as yet not exploited in the media market. A mean-variance model manifest itself in the form of a problem of quadratic programming in which the function-objective represents a compromise between the expected return (profitability) and variance of return (risk). In this approach the diversification of steps occurs naturally, without the need for artificial restrictions. Linear restrictions, such as disposable budget, can also be inserted in the model. In phase 1 of this project, the proposed methodology was implemented in a prototype validated by means of real data on the sales of magazines. Two fundamental characteristics of the model were verified. In the first place, it allows linear restrictions to be imposed and dealt with, which lends the model great flexibility. In the second place, it demonstrates its ability to speed up the process of creating efficient media plans. The central aim of phase 2 will be to develop an integrated and user-friendly system for the optimization of insertions in the printed media based on the meanvariance model. (AU)