Analysis and forecasting of pet food markets using extrapolation methods

Abstract

One of the manifestations of the socialization of the economy is a change in the behavior of consumers, which leads to shifts in the structure of markets, the nomenclature of goods presented on them, and fluctuations in the ratio between supply and demand. A vivid example of such changes is the market of pet products, where thanks to the phenomenon of humanization of the attitude towards animals, it became possible to observe the growth of the market as a whole, and the appearance of new products and services for four-legged consumers. In this study, the main markets for distribution of pet products are analyzed, in particular, world, European and Ukrainian. General trends and factors influencing them are highlighted. The analysis of this market in Ukraine made it possible to identify common features in the development of the domestic pet industry and in the countries of Europe and the whole world, and to prove that the domestic market of pet products is promising. Sales volumes of Ukrainian pet food market show a stable upward trend. From the variety of formalized methods for forecasting the demand in pet food markets, the authors chose extrapolation methods, namely, methods of trend analysis and adaptive methods of exponential smoothing. Using trend models, as well as adaptive methods, point and interval forecasts of demand indicators on the global and European markets were built. Trend models were built for the total value of the global pet food market and the total demand for these products worldwide. Point and interval forecasts were built on the basis of these models with a sufficient degree of adequacy and accuracy. The trend model of the Tornqvist function was applied to build forecasts on the Ukrainian pet food market. To forecast the annual European sales of animal nutrition, the adaptive method of triple exponential smoothing was applied and point and interval forecasts for the next four periods were obtained. In all cases, the selected predictive models demonstrated satisfactory predictive ability and satisfied the adequacy criteria.