Skip to main content Skip to Footer


Fraud detection for retail: Use analytics​ to identify malpractice fraud

Prevent internal and external fraud, reducing losses resulting from malpractice.


When sales networks do not function, organizations experience losses and accounting troubles. Detecting fraud and reducing losses is a key goal for audit managers of retail companies. Detailed analysis of available information is difficult and expensive. The number of transactions is huge, and the size of the network and the volume of detail that can be recorded (including movement of the drawers of the boxes in the stores) make the job more difficult.

It’s important to quickly identify and stop improper behavior, but it’s also necessary to identify new behavior and risks that were not previously associated with losses.

Fraud Detection identifies risks of fraud or improper behavior resulting in losses. Through a mix of business rules, anomaly detection and predictive analytics that identify risky behavior, the application can assign a fraud risk score to any event.


Specific Services

Fraud Detection uses advanced analytics to support audit managers, area managers and sales points with insights on fraudulent behavior. Fraud analysis data at the highest level of detail helps prevent fraud related to risk indicators to shops, cashiers or individual transactions.

The application contains a predefined set of business rules related to the retail market. Users can add new rules according to their experience, and their business knowledge, to spot new unethical behavior. The application adds anomaly detection, predictive analysis and social network analysis (SNA) to the business rules for better predictive detection.

Fraud Detection can detect suspicious behavior not previously known, identifying the drivers of anomaly. With predictive analytics, users can learn from data to identify patterns, and thus prevent them in future. SNA allows users to analyze networks of relationships in transactions (e.g. shift work) to identify potentially risky situations. The ability to define risk matrices allows the user to choose which components to use, and what weight to give them in the evaluation of transactions and risk assessment.

The application covers the business process end-to-end: Different users can see their areas of interest (e.g., individual business areas, individual stores) and be alerted in the case of anomalies according to alerting rules associated to specific actions.

Key features

  • Business rules at all levels (store, cashier, retail)

  • Risk matrix

  • Clustering and anomaly detection

  • Anomaly indices and risk scores

  • Risk drivers and impact analysis

  • New rules from anomaly detection algorithms

  • Ability to manage events / transaction (“fraud” label assigned by the user)

  • Predictive models

  • Social Network Analysis at multiple levels

  • Ability to manage real-time monitoring and alerting logic

Why Accenture

Products applications are built within the scalable Accenture Insights Platform that provides fast and easy access to an array of industry and functional applications that bolt on to the platform, allowing quicker time to market and more rapid results for clients.

Clients also access our wide range of capabilities rooted in:

  • Industry knowledge. The validity of our advanced analytics outcomes is underpinned by our deep knowledge of the sector.

  • Business. Applications are designed for business users and focused on business results, minimizing advanced analytics complexity. Getting to accurate outcomes does not require statistical, mathematical or IT knowledge—just business know-how.

  • Flexibility. Applications are based on a framework that can be easily integrated in an enterprise operational environment to drive business process action.