Examinando por Autor "Arcos Fonseca, Edwin Fabian"
Mostrando 1 - 1 de 1
Resultados por página
Opciones de ordenación
- ÍtemModelo de pronóstico para estimar el comportamiento de ventas en pospago digital, de una empresa del sector telecomunicaciones en Colombia(Fundación Universitaria Los Libertadores. Sede Bogotá., ) Arcos Fonseca, Edwin Fabian; González Veloza, José John FredyThe objective of this work is to propose a statistical model that allows forecasting the monthly sales of postpaid aircraft of a telecommunications company, this is due to the fact that in Colombia in the processes of activation of postpaid aircraft of the digital channel it has time dates of activation managed by the ABD (Data Base Administrator) that causes sales to be effective until approximately 5 business days before the end of the month, due to sim delivery and portability processes, this causes that if you do not have a good forecast at the The middle of the month is when alerts appear that show marketing cost overruns to be able to increase sales, the current model every month we have a gap of more than a thousand even reaching -4765, something that is not correct is annexed table 1. Table 1. Current projection gap. The current cost per discharge is approximately 41,500 and as we observe every month a negative GAP is being presented vs. the goal, there are projections with a lag greater than 2000 almost every month, which causes projected values to be showing close to the goal that there is some calm at the beginning of the month and approximately halfway through the alert is given that we are in the real below. Currently trying to recover what was lost generates an extra cost of almost double and just as from the beginning there is a budget and when a breach is generated it is money, for example in the month of October they stopped receiving $-135,539,000 that they should contribute the area for company goals, this in the future may cause more radical decisions for the area. To carry out the analysis, a comparison was made of the Holt-Winters and Arima models with these models, forecasts can be obtained to be able to make decisions, the data available is only that corresponding to January 2021, the complexity of the case is that As there is no comparison for more than one year, it is not possible to identify trends that occurred in the same month in previous years, in order to know if there are months in which a high or low behavior is normal, although in the same way the models developed allow us to know that sales they were not as high as expected and to be able to have a more timely reaction.