Assessing the impact of seasonal changes on advertising revenues in the age of digitalization and sustainability: A print-visual media business example

This study mainly aimed to investigate the impact of seasonal changes on media advertising revenues (AR) in R environment through a visual-print media company example. So, the study primarily aimed to examine whether visual, print and internet ARs were affected by seasonal change. The study also aimed to indicate the trends of visual/ print/ and internet ARs over time in the light of seasonality. Furthermore, it was also aimed to make AR estimates for the next year with the best performing methods covering ARIMA and ETS. Methodologically, R programming language and its libraries were used together with ARIMA and ETS time series models. The data set consisted of a monthly ARs series consisting of 180 rows covering the period between January/2007-December/2021 and having a sufficient number of observations for the application of the models. Data were analyzed with R and next year’s forecasts were done with ARIMA and ETS. The study revealed that all AR types were affected by seasonal changes in the specified period and included seasonality, i.e. specially in the summer months. Additionally, it was observed that AR types have different trends despite containing seasonality. The study found that visual ARs denoted a slight downward trend including seasonality during the period. It was also indicated that print ARs reflected a very rapid and drastic downward trend covering seasonality. The study also revealed that internet ARs showed a very rapid and sharp upward trend compared to the other two AR types despite having seasonality. Furthermore, taking into account the calculated error metrics, it was found that the ARIMA model showed the best estimation performance in both i.e. visual and internet ARs. On the other hand, the ETS model indicated the best estimation performance in print ARs. Accordingly, estimations of the visual/print and internet ARs for the next year were made with the best performing methods. It was pointed out that estimated ARs included seasonality in the summer months and revealed similar results to the previous trends based on historical data.