At present, enterprise performance management (EPM) systems are widely used in practice, because they facilitate strategic decision-making and contribute to improved information transparency of organizations. However, methodological issues related to managing the development of such systems seem to be insufficiently investigated and elaborated. The purpose of the study is to formulate and justify fundamental principles for managing EPM systems’ development. These principles derive from peculiarities of EPM systems themselves and the features of their development management. In particular, the features of EPM systems are complexity and modular structure, large-scale scope, the long-term nature of planning, monitoring and analysis, use of aggregated information – both financial and non-financial. The features of managing development of EPM systems include the implicit nature of the resulting economic benefits, influence of stochastic factors, as well as availability of “complicated” projects (with uncertain outcomes, ability to reexecute and multiple variants of implementation). As a result, the basic principles of managing development of EPM systems can be formulated. There are the principles of a system, the going concern, business alignment, value for money, program management, alternativeness, feasibility, stochasticity, as well as resources aggregation.
The significance of these principles is explained by the fact that they can be used as a basis for an integrated process of managing the development of EPM systems. These principles are also valuable for formalizing certain elements of the management process, such as assessment of the maturity level of EPM systems, the formation of potential development programs, simulation of implementation of the development programs, as well as decision-making regarding selection of development programs for implementation.
System of logistic regression models is suggested which takes into account variation of financial coefficients in dynamics. Besides it comparative analysis of predictive accuracy is performed for the system of models on the basis of learning and control samples of manufacturing companies.