Customer Analytics and Fraud Detection - Energy

The Problem

Thanks to electronic counters, a huge amount of energy consumption data (load curves) is now available and mostly unexplored. Hence, the key challenge of many energy companies is nowadays that of extracting from such data previously unknown and useful knowledge in order to gain insigths into their business, get a deeper understanding of how their customers use energy and, thus, to improve both the quality and the efficiency of their services – e.g., through customer tailored contract offers, or energy comsumption prediction aimed at shielding energy companies from power shortage/surplus.

Exeura solution

From the collaboration among Enel, Exeura and Politecnico di Torino originates SHAPE, a business analytics platform for customer profiling and fraud detection.

Success story


Customer Analytics e Fraud Detection

The Need

The transition from supplier-centric to customer-centric framework carried out by Enel after the liberalization of the electricity market needs not only an efficient and effective metering system, but also new business strategies and innovative services for the customer, retailer and distribution network.
Thanks to the “Telegestore”, the Automatic Meter Management system, today Enel is able to measure and collect remotely large amounts of consumption patterns (load curves) recorded on a 15 minute basis from over 30 million customers. The extraction of hidden “knowledge” from this huge amount of data is crucial for deeply understanding customer needs.

The Solution

For this purpose, Enel recently started the Research and Development project “SHAPE” in collaboration with Politecnico di Torino (scientific partner) and Exeura (industrial partner). The aim of the project is the development of a web-based enterprise platform for Business Analytics focused on the daily load curves sourced from the Enel electronic meters. Using ad hoc developed advanced time series analysis techniques, the SHAPE platform enables:

  • Customer Segmentation: identification of typical load profiles, each representing a set of homogeneous customers based on the energy consumption/production curve;
  • Customer Load Prediction: short/medium term forecasting of the customers energy consumption at 15 min resolution;
  • Tariff or Action Effectiveness: geographical and temporal analysis of energy distribution w.r.t. different real or simulated pricing/tariffs plans and aggregations of customers (“what-if” analysis);
  • Fraud Detection: reduction of economic losses due to energy supply anomalies, including identification of possible fraudulent activities.

Thr Benefits

The knowledge of the typical patterns of the aggregated energy consumption/production brings benefits like identification of new business opportunities and improvement of pricing schemes. The economic benefit of understanding customer loads decreases investment costs by reducing the planning margin. Moreover, the acquired knowledge enables energy fraud detection along with a better estimation of economic losses resulting from interruptions of service. The summary information on the consumptions is useful for more effective management of the electricity distribution network in synergy with existing initiatives in the field of smart grids.

Customer: Enel
Market: Energy
Task: Customer profiling and fraud detection