The aim of this paper is to provide an approach to analyze the performance of TV programs and to identify what can be done to improve them.
The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS), the Ng-model, Grey relational analysis (GRA), and principal component analysis (PCA) were applied to evaluate the programs, using audience, share, and duration as the performance criteria.
By comparing TOPSIS to the Ng-model, PCA, and GRA, we verified that SVD and bootstrap SVD TOPSIS provide a good balance between equal-weights TOPSIS and the other models. This is because SVD and bootstrap SVD TOPSIS break down the data to a higher degree, but are less impacted by outliers compared to the long tail models.
To determine which TV programs should be replaced or modified is a complex decision that has not been addressed in the literature. The advantage of using a multi-criteria decision-making (MCDM) approach is that analysts can choose as many criteria as they want to rank TV programs, rather than relying on a single criterion (e.g., audience, share, target rating point).
This work represents the first time that robust MCDM methodology is applied to an audience data set to analyze the performance of TV programs and to identify what can be done to improve them. This study shows the application of a detailed methodology that is useful for the improvement of TV programs and other entertainment industry content.
