Towards a Natural Experiment Leveraging Big Data to Analyse and Predict Users' Behavioural Patterns Within an Online Consumption Setting

The authors develop a model for multi-criteria evaluation of big data within organisations concerned with the impact of an ad exposure on online consumption behaviour. The model has been structured to help organisations make decisions in order to improve the business knowledge and understanding on big data and, specifically, heterogeneous big data. The model accommodates a multilevel structure of data with a modular system that can be used both to automatically analyse data and to produce helpful insights for decision-making. This modular system and its modules, indeed, implement artificial intelligent algorithms such as neural networks and genetic algorithms. To develop the model, therefore, a prototype has been built as proof-of-concept using a marketing automation software that collects data from several sources (public social and editorial media content) and stores them into a large database so as the data can be analysed and used to implement business model innovations. In this regard, the authors are conducting a natural experiment - which has yet to be completed - to show that the model can provide useful insights as well as hints to help decision-makers take further account of the most 'satisficing' decisions among alternative courses of action.