Fraud has become a chief concern for governments and companies. In fact it is estimated that losses from fraud in organizations can be as much as 5% to 9% of their annual profit. To better understand the different fraud management frameworks, it is necessary to have an understanding of what fraud is, its components and the various forms it can take.
In the business world, fraud is associated with an action that goes against truth and integrity, damaging the organization against which it is perpetrated. Fraud can compromise a company, whether it is committed externally by clients, suppliers and other parties, or internally by employees, managers or shareholders.
Some characteristics of the current environment and the opportunities it has to offer are as follows:
The value added by these management mechanisms is reflected in economic terms (according to an ACFE study, losses from fraud at the global level fell by 54% thanks to the adoption of proactive data monitoring measures), and also in reputational and compliance terms. Both these two aspects are particularly relevant given the current regulatory environment, which encourages companies to invest in and implement fraud management methods.
The purpose of this document is to share some insights on the concept of fraud, as well as on the key elements used to manage fraud and the opportunities for optimization that arise as a result of technological advances such as Big Data and Analytics. These are based on the availability and analysis of large data volumes as well as the implementation of profiling and segmentation methodologies.
With a particular focus on the Energy industry, this document describes fraud events specifically for this industry which, due to their representativeness and the fact that they drag on the resources of companies, require specific treatment and for which detection techniques and the integration of these techniques into the management process are even more relevant.
In addition, this document will show how the modeling, profiling and segmentation methods complement the implementation of a methodology for quantifying the economic usefulness of actions, a methodology that discriminates the quality of the segmentation performed for fraud detection purposes (the effect of the segmentation models), from the appropriateness of implementing the actions (or the effect of the detection campaigns themselves), with the aim of separately assessing the cost effectiveness of investing in modeling techniques and investing in theft detection inspections. In this sense investment in fraud management is considered as just another company investment.
These techniques are supported by modeling platforms that combine mass data processing components with statistical software and tools for both access control and the management of roles, incompatibilities, etc.
Finally, this publication includes some examples to illustrate the implementation of energy theft detection probability modeling techniques (a specific case of external fraud). These models are based on the characterization of the point of supply using variables that identify the factors underlying fraud, such as the physical characteristics of meters, commercial and socio demographic characteristics of customers or users, usage and behavior history in relation to theft, other transactions with the customer, customer engagement, claims, the result of inspections, etc.
What is shown is therefore the added value of data, of information on costumers and transactions (hourly consumption, customer data, access to systems, etc.), in quantifying the probability that fraud events will occur and the use of this calculation to optimize both preventative action (e.g. segregation of duties and system access control) and mitigating action (e.g. implementation of inspection campaigns and segmentation of profiles according to their propensity to theft). Thus, action can be prioritized under an economic and profitability rationale based on the estimated probability that a theft event will occur or the possibility that fraud may be committed in the commercial cycle, as well as the materiality of the potential impact (energy defrauded, amounts stolen, etc.).
In fact, according to data made available by one of Europe’s main electricity distribution companies, following their use of data collected from intelligent meters, the percentage of fraud cases affecting the company that were detected went from 5% to 50%.
Last white papers
© GMS Management Solutions, S.L., 2020. All rights reserved. The information contained on this publication is of a general nature and does not constitute a professional opinion or an advisory service. The data used in this publication come from public sources. GMS Management Solutions, SL assumes no liability for the veracity or accuracy of such data.