Data Analytics exploitation for Growth Synergies in Multi-Business Companies

The paper carried out under the supervision of Prof. Tania Cerquitelli and in collaboration with the Company “Mediamente Consulting s.r.l”, intends to propose data-driven strategies for the achievement of synergies between businesses owned by multi-business companies. First of all, a broad overview of multi-business companies and diversification strategies is presented, focusing on diversification methods (entry, mergers, acquisitions, … ), on diversification types (horizontal and vertical, related and unrelated, … ), and on the opportunities and risks that these strategies entail. Furthermore, the managerial and organizational implications of a multi-business context are specified. From this, the concept of cross-business synergies is introduced, and after outlining costs and advantages for the companies that try to reach them, a synergy classification is made starting from the type of resources to be leveraged across businesses (operational, financial, managerial, … ). Among these classes, growth synergy is deepened, deriving from the sharing of operative resources of several businesses under a single corporate management in order to achieve revenues super-additivities. The document, subsequently, shows the importance of data analytics in business decision-making, starting from the explication of the Data Mining process for the extraction of knowledge from data, to then define the Data Warehousing system to support data-driven decision-making. The main Data Mining and Machine Learning techniques are then presented: classification, clustering, association rules, time series and neural networks. Of each one, at least one algorithm is examined in depth. The presented theory is applied to a real case, to a company operating both on the plant nursery business and on the supermarket business. Starting from a first analysis of the business contexts and of the available Data Warehouse data, some general statistics are integrated with two customer classifications (by purchase frequency and by annual amount spent), and with a clustering of customers by product categories purchased. In this way problems of customer loyalty and failure to exploit synergies were detected. In order to unlock growth synergies, and thus obtain revenues super-additivities, a redesign of the store layout and a method that, through association rules, attempts to improve growth synergies, are presented, emphasizing the importance of correct timing to implement the suggested strategies by time series forecasting. There is no resolution of the business case, and therefore no demonstration of ex-post results. Instead, an attempt is made to define a practical model from which the decision makers of the companies, with a much broader vision of the company and the environment, can find, by exploiting data analytics, ideas and opportunities for their own strategic lines designed to improve cross-business synergies.