Abstract: The Seven-Up Bottling Company Plc, Kaduna Plant has no mathematical model that takes into account seasonal indexes for forecasting sales volume and since their products are seasonal in nature, the classical decomposition of time series data into the various components was used to model the present using the past trends into the future periods in order to have good forecasts for effective management. Sales records of Seven-Up Bottling Company Plc, Kaduna Plant for the past ten years (2007 – 2016) were collected, analysed and then various statistics such as t-test, z-test and Durbin-Watson test were carried out for the purpose of testing hypothesis. A model was then developed using deseasonalised data series that accounted for 95% of the total variability. The results showed that the seasonal indexes for the months of January, February, March, April, May and December were higher than those of the months of June, July, August, September, October and November. The highest seasonal value of 113.8% was recorded in May and the lowest of 75.3% was recorded in August. The model was used to forecast product sales for 2017 with the result showing good agreement between predicted and observed values.
Keywords: Classical decomposition; autocorrelation coefficient; seasonal indexes; trend-cycle; irregular component; cyclical component; trend seasonal factors.
| DOI: 10.17148/IJARCCE.2019.8221