Pricing in prepaid telecom industry

Abstract

In India, telecom service providers prefer to operate in the prepaid sector. This segment is defined by the heterogeneity of the consumers that creates a big challenge in pricing of the prepaid products. Therefore, it was imperative that we first study the evolution of the tariff plans in the prepaid telecom products and then propose a method that could be used by service providers to effectively price their products in order to extract maximum surplus from the varied consumer segments. Conjoint analysis is best suited to understand consumer behavior when it comes to choosing a particular set among many options. In order to design the conjoint experiment, we decided to use choice based conjoint since it simulates real life trade-offs and due to its lexicographic nature of choice card generation, helps make the survey very easy to administer. The attributes we used for the choice based conjoint experiment were brand, STD calling rates and Local calling rates. These were chosen over the other attributes that were identified in our exploratory research since these were found to be the most significant in our pilot survey. Once these attributes were chosen, we defined the levels for each attribute. The three major brands Vodafone, Airtel and BSNL that are predominant in the IIM Bangalore campus were used for the study since the IIMB student population was used for sampling. STD calling rates and Local calling rates were pegged at 5 levels, the levels being decided through indepth interviews that we conducted within the IIM student community. The generation of choice cards and subsequent multivariate data analysis was conducted using SAS. Here we specify 25 cards as the design size. Once these cards are used to administer the survey among 30 respondents from the IIMB student community, the Logit model (logistic regression) is used to regress the data. We then use the results of the regression to demonstrate a few ways in which this model could be used to analyze the pricing situation in the telecom industry and also give pointers as to how the model could be used for screening of the heterogeneous consumer base. We demonstrate how the model could be used to predict market shares (choice probabilities), how to understand whether consumers are sensitive to STD call rate changes as opposed to Local call rate changes, to understand how the choice probabilities of brands would change in response to change in an observable attribute like STD calling rate or Local calling rate by the same brand or a competing brand. Finally we demonstrate how a marketer can use this model to get the most critical input, in terms of screening the heterogeneous market. We demonstrate 2 different choice behaviors by developing the above stated models for 2 different sets of students, ones with prior work experience and one without prior work experience. Using these models, we can understand the sensitivities of these segments to the various attributes, and thus make pricing decisions based on that. While we have demonstrated the utility of the model comprehensively, we would like to point out that the above model is ridden with limitations like being capable of representing only systematic taste variation, and the problem of aggregate bias. In conclusion, we would like to point out that this model could be directly used by any player in the prepaid telecom sector with a touch of refinement based on the need of the study. Further, Software Programs can be used to understand better ways of segmenting the sample data and thus understand the market better.