Whitecap Consulting was appointed by the FCA and CoLC to conduct an independent evaluation of this dataset, engaging with users, assessing its effectiveness, and providing recommendations for improvement. Our findings, now published by the FCA, confirm that this synthetic dataset is a valuable tool in the fight against APP fraud.
The Role of Synthetic Data in Fraud Prevention
One of the key barriers to developing effective fraud prevention technology is access to real-world data. Due to privacy concerns and regulatory constraints, organisations often struggle to obtain datasets that allow them to train, test, and validate fraud detection models.
The APP fraud synthetic dataset addresses this challenge by providing a privacy-safe yet realistic dataset for fraud prevention innovators. Created using agent-based simulations, the dataset replicates real-world consumer interactions and fraud patterns, spanning 20,000 synthetic individuals and 15 million transactions over a two-year period.
The dataset, created by Smart Data Foundry, includes:
- Individual and business identities
- Banking transactions across four synthetic banks
- Telecom data (calls and SMS metadata)
- Over 61,000 attempted fraud events
- Seven types of APP scams, including romance, investment, and bank impersonation fraud
Whitecap’s Evaluation: Key Findings and Recommendations
Our role was to assess the dataset against key success criteria, gathering insights through workshops interviews with dataset users and interviews with the FCA. The evaluation confirmed that the dataset is a powerful resource for testing fraud detection solutions. However, we also identified areas for enhancement, such as:
- Expanding fraud coverage: Incorporating additional scam types and behavioural nuances to reflect evolving fraud tactics.
- Enhancing usability: Improving data structuring and accessibility to ensure smoother integration into fraud detection models.
- Strengthening real-world applicability: Increasing granularity in transaction narratives and telecom data to align more closely with live fraud scenarios.
These insights will inform the next phase of dataset improvements, ensuring that future iterations provide even greater value to fraud prevention specialists and financial institutions.
Driving Innovation in Financial Crime Prevention
The FCA’s Digital Sandbox initiative, under which this dataset was developed, is a crucial step towards fostering collaboration between regulators, fintechs, and fraud prevention experts. The success of the APP fraud dataset highlights the potential for synthetic data to play a transformative role in financial crime detection, enabling faster, safer, and more effective fraud prevention solutions.
As fraud continues to evolve, so too must our approach to combating it. At Whitecap Consulting, we are proud to have played a role in this innovative initiative and look forward to supporting future advancements in fraud prevention, data-driven innovation, and financial sector resilience.
Read more about the FCA’s APP Fraud Dataset initiative here
If you’d like to discuss this blog post or share your own perspective on the issues covered, please get in touch or comment.
Established in 2012, Whitecap Consulting is a regional strategy consultancy headquartered in Leeds, with offices in Manchester, Milton Keynes, Birmingham, Bristol and Newcastle. We typically work with boards, executives and investors of predominantly mid-sized organisations with a turnover of c£10m-£300m, helping clients analyse, develop and implement growth