by Derek Brink

Big Data, the fraud fighter

Nov 04, 20124 mins
IT Strategy

For industries with a full-time focus on fighting fraud, such as banking, insurance and healthcare, greater visibility and timely intelligence is invaluable.

Key takeaways from Aberdeen Group’s recent analyst insight on Fighting Fraud with Big Data Visibility and Intelligence include the following:

Fraud is costly Direct financial losses from fraud can vary significantly from one industry to another or in different parts of the world, but the annual cost of fraud is substantial.

In the Association of Certified Fraud Examiners (ACFE) 2012 Global Fraud Study, an average of 5 per cent of annual revenue, with a median loss per incident of $140K (£87.5K).

Fighting fraud is complex External sources, including tips, notification by law enforcement, external audits, accidental discovery and confessions are responsible for a higher percentage of detection and higher median losses, than are internal sources and methods, including management review, internal audit, account reconciliation, document examinations, and IT controls. Time to detection ranges from 12 to 36 months.

Current methods are inefficient Even in a backwards-looking forensic mode, current IT controls were found to be the source of the fewest incidents detected. This is a situation that cries out for a technology-based solution.

Fraud frequency doesn’t tally with impact Doing the simple mathematics of frequency multiplied by impact doesn’t help much with focus.

Financial statement fraud, for example, is the least common (7.6 per cent) but has the highest median loss ($1,000K; £625.8K), while asset misappropriation has the highest frequency (86.7 per cent) but the lowest median loss ($120K; £75K).

The choice of death-by-severe-trauma or death-by-a-thousand-cuts leads to pretty much the same result.

Success in fighting fraud pays off Every penny of fraud loss recovered or avoided goes straight to the organisation’s bottom line.

In the example of the Health Care Fraud and Abuse Control program (HCFAC), under the US Department of Health and Human Services and the US Department of Justice, every $1.00 expended in fighting fraud returned an impressive $7.20 in judgments and settlements.

New strategies for fighting fraud are emerging Rapid changes in information technology infrastructure are increasing the difficulty of maintaining high levels of preparedness simultaneously against the full range of threats.

In response, organisations are adopting enhanced strategies for fighting fraud: from 100 per cent success at prevention, to greater visibility, faster detection and incident response; from post-incident forensics, to proactively figuring out what’s happening using big data and predictive analytics.

Solution providers are leveraging Dig Data for predictive analytics The problem is not that there is too little information, but too much and most of it in disparate stovepipes and silos.

Next-generation solutions for predictive analytics are solving the Big Data challenge by providing enterprises with the visibility and intelligence they need to move from post-incident forensics to a more proactive and predictive approach to fighting fraud.

The top priority is currently the top line In Aberdeen’s Q1 2012 predictive analytics survey results, top line-oriented sales and marketing issues dominated the primary uses for predictive analytics, with specific drivers for current investments including:

– Tougher competitive environment (38 per cent) – Falling customer retention (29 per cent) – Increasing cost of customer acquisition (24 per cent) – Decreasing revenue (23 per cent) – Difficulty of forecasting demand (22 per cent) – Changing customer demographics (22 per cent) – Proliferation of channels (22 per cent)

In contrast, roughly 1 in 6 (16 per cent) respondents in Aberdeen’s study indicated the current use of predictive analytics for the detection and prevention of fraud. The top priority is currently the top line, but as we have seen fighting fraud can also yield significant bottom-line results.

Crawl, Walk, Run is a proven, pragmatic approach In the beginning, initiatives to leverage big data and predictive analytics for fighting fraud can sometimes get bogged down in debate over the optimal approach.

Both of the following are examples of the pragmatic crawl-walk-run approach that is characteristic of the companies who are the most successful in their enterprise-wide initiatives, as seen consistently in Aberdeen’s research: – Taking more time to integrate all data sources for a single application-process-workflow – Integrating fewer data sources that apply more broadly, and making rapid progress in the ability to analyze, understand and take meaningful action

A free, on demand webcast on this topic, featuring the author, and an expert from IBM, is available at For more information on this or other related research topics, please visit

Derek E. Brink is vice president and research fellow at Aberdeen Group

Pic: jonny2lovecc2.0