FORMATION OF THE ANALYTICAL SUBSYSTEM OF THE MARKETING INFORMATION SYSTEM OF AGRICULTURAL ENTERPRISES

Abstract

The purpose of the article is to substantiate the analytical subsystem of the marketing information system of an agricultural enterprise as a decision-support layer that converts prepared data into interpretable evidence, feasible courses of action and feedback for management. The methodological basis combines systems analysis, comparison, classification, synthesis and structural-logical modelling. The study clarifies the boundary of the subsystem: its input is a decision-oriented set of verified and integrated data, while its output is not an automatically imposed decision but an analytical result with assumptions, uncertainty and alternative actions available to the responsible manager. Descriptive, diagnostic, predictive and prescriptive analytics are matched with the principal marketing tasks of an agricultural enterprise, including demand forecasting, pricing, buyer segmentation, market selection, channel assessment, competitive analysis and performance control. The article proposes a structural-logical model that links a marketing problem, an analytical request, prepared data, method selection, validation, interpretation, alternative actions, a managerial decision and assessment of the implemented result. Its distinguishing feature is the combination of two feedback loops. The methodological loop returns an unsatisfactory result to the data, assumptions or method-selection stage, whereas the managerial loop uses actual performance to revise the problem, decision criteria and analytical model. Machine learning is positioned as one group of methods within this architecture rather than as an autonomous source of managerial value. Its use is considered justified when the decision is economically material, the data cover relevant seasons and market conditions, performance is compared with a transparent baseline, and the result can be explained, monitored and integrated into the existing workflow. The scientific contribution consists in a decision-oriented representation of the subsystem that combines the type of managerial problem, the level of analytics, validation requirements, human responsibility and post-decision evaluation. The practical significance lies in using the model as a blueprint for the incremental design of analytical support in agricultural enterprises without requiring an immediate transition to complex artificial intelligence. It is concluded that the maturity of the subsystem should be assessed not by the sophistication of software, but by the traceable contribution of analytical results to the quality, timeliness and economic justification of marketing decisions.