The Cost of FailureCarefully formulating analytics and business practices can be the best way to avoid financial catastrophe
by Edward S. Robins Continued from Page 1 Because of their complexity, deep dive metrics are often prototypes for technology projects. You must carefully balance these requests so that the result analysis is actionable in its final scope, yet deliverable in the required time frame. Deep dive analyses frequently drive significant business changes because of their ability to quantify formerly invisible cause and effect behaviors. Protect against fraud. Fraud isn't just a problem of external suppliers and customers knowingly cheating. It can also be an internal matter, involving personnel who misrepresent results because of job pressures or who are illegally lining their pockets. The types of fraud along the money trail are extensive and vary by industry. False accounting, for example, can be detectable through Benford's Law, which can indicate if a sequence of numbers has been artificially created. Abnormal purchasing or selling patterns, illegitimate intrusions (allowed or otherwise) that modify accounting records, subtle misuse of corporate property, establishing questionable financial vehicles or customer accounts, and even money-laundering activities can be uncovered by fraud-detection systems. These systems are as diverse as the types of fraud and are only now achieving reasonable levels of sophistication, ranging from simple scoring techniques to industry-specific behavioral analytics algorithms. Many vendors, for example, focus on telephone fraud. Use watchdog and proactive systems. Senior managers don't want to waste precious time chasing poltergeists; they need alarms to sound early. Unfortunately, although base technology is available, no generic product can help protect against fraudulent activities at the operational level for various reasons (such as the existence of differing accounting practices across business units, industries, and national boundaries). In the early 1990s, for example, both Sybase and Bausch & Lomb discovered business practices that seriously overstated sales figures and affected their corporate integrities. Sudden changes in sales patterns should have set alarm bells ringing, but they didn't. There were no watchdogs on duty. Modern BI tools can serve as crude watchdogs, but more sophisticated third-party products are appearing. Examples include Mantas Inc.'s Behavior Detection Platform, which proactively detects suspicious patterns or events in the financial services industry; NetMap Analytics' Link Analysis technology, which finds links among unrelated events for the insurance and retail sectors; Neural Technologies' Minotaur, which provides neural networks to detect money laundering and other fraudulent activities in the utilities, financial, and telecom industries; and ViPS Inc.'s STARS and IBM DecisionEdge, which target healthcare fraud. Align people and IT. You should engage in at least some rules separation, a process in which business rules are made explicit and automated. Essentially, the modeling must move from IT coders, legacy code, and database modelers to the value chain stakeholders involved in governance, operations, analytic, and auditing functions. If it doesn't, implementing anything suggested in this article is likely to be a costly nightmare. Rules implementations are not all-or-nothing exercises because systems can be structured incrementally. As a start, perform a risks audit — a separation of rules based on their priority and ability to provide governance in the context of planned IT upgrades. In some cases, you may need internal or external consulting analysts to separate the rules and identify the kinds of rules needed. Such analysts are essential to help manage and develop systems. Most BI and ERP companies provide rules analyst services. However, most of them are prepackaged solutions that won't give you the control and flexibility you need. Third party rules-engine middleware vendors — Fair, Isaac & Co. and Corticon among others — are good alternatives worth investigating. Manage the cultural impact. More tightly governed systems may entail changes in corporate culture and likely place a higher demand on honesty. Managers may need to be more forgiving of human error and judgment, at least initially. In the high stakes area of drug discovery, for example, reducing drug discovery costs is causing a cultural tsunami: Systems coming online to capture and share the knowledge of individual scientists and doctors (two groups that jealously guard their personal knowledge capital) enable management and peers to judge their individual competency and efficacy, which also threatens the bargaining position of these high-value individuals. If undesirable behaviors are uncovered, sensitive remedial action will need to be taken. Benefits and DrawbacksBesides overall improvements in flexibility, efficiency, and change management, rules systems containing catastrophe-avoiding procedures can be built to help managers better appreciate the risks in making decisions. Having a risk dashboard, for example, will enable better strategic thinking. It will make more demands on the senior management to understand risk and equally challenge the designers of such systems. The planning should include considerations to inculcate more honesty in the workforce. Studies in this area suggest that corporate cultures with high ethics and honesty are more successful companies. A company that doesn't hold to such standards inevitably breeds a culture of theft and can collapse. Furthermore, investors should know that catastrophic risk management is being performed; a CEO looking to enhance value and confidence may find it a valuable tool for his or her kit bag. The downside is the risk of discovering unpleasant surprises at an awkward moment. In addition, the presence of a cybernetic Big Brother isn't conducive to a pleasant work environment, although such a presence is increasingly common. Staff buy-in through involvement in development is often critical. Interface design must hide complexity to help overcome estrangement and enable users to feel accommodated as changed processes are implemented. If these needs aren't satisfied, management may find itself with unhappy employees and customers. Good project managers, programmers, and interface designers will be needed for the grinding work of getting it done. Bottom LineCatastrophic failure is an underestimated mode of corporate failure, and today, more sophistication should be demanded of management and their tools. Investors should also be aware that management has potential disaster in check. IT can help, but it'll never be the complete solution; standard accounting practices are also no guarantee against even short-term risk of catastrophic failure. Rather, more sophistication is needed on the part of vendors, managers, and investors in understanding this complex problem. In effect, the risk of catastrophic failure can and should be factored in as part of general business practice. Edward S. Robins, Ph.D. [esro@attbi.com] is general manager of Esrotech, a technology-based consulting company, and on the faculty of the College of Arts and Science at Northeastern University in Boston. He has more than 20 years of experience in discrete product development and manufacturing, software development, decision support and analytics, and systems architecture. Case Study: Good News and Bad NewsNewly on the job as R&D manager at a small discrete manufacturing company, I soon found that the system of procurement for each bill of materials had been left to the engineers. Consequently, the sourcing of some materials represented a significant risk, particularly as supplies of one critical component, for which there was at the time only one supplier, would be delayed by several months. Neither the engineer responsible nor the former manager had identified this procurement risk. In addition, the engineers were more preoccupied with current technical problems and gave little thought to raising the sourcing issue with management. The potential impact on cash flow, customer confidence, and manufacturing costs represented a critical risk to the company's financial integrity. Without an internal risk-analytic system in place — at least, one affordable for a small company — it was impossible to ensure this situation would not recur, so my focus shifted toward production policing and away from new product development. Although an engineering solution was eventually found, its weaknesses highlight the need for a rules at three different levels:
In this case, establishing rules of engagement among chief stakeholders, higher level decision-makers, and line engineers would probably have avoided unnecessary and costly work. In addition, a rules engine may have identified the risk adequately enough to enable risk-reducing actions to be taken — perhaps entirely avoiding problems at the product design stage. RESOURCESBergeron, Bryan and R. Kurzweil. The Eternal E-Customer, McGraw-Hill, 2000. Beroggi, Giampiero. Decision Modeling in Policy Management: An Introduction to Analytic Concepts, Kluwer Academic Publishers, 1999. Starbuck, William H. and P. Narayan Pant. "Trying to Help S&Ls: How Organizations With Good Intentions Jointly Enacted Disaster," in Organizational Decision Making, Zur Shapira (Ed.), Cambridge University Press, 1997. Von Halle, Barbara. Business Rules Applied, Wiley, 2001. Business Rules Community: www.brcommunity.com Securities Industry Association's Preliminary Guidance for Deterring Money Laundering Activity: www.sia.com/moneyLaundering/pdf/AMLguidance.pdf Related Articles at IntelligentEnterprise.com: For more information about the role of analytics in fraud detection, see Girish Keshav Palshikar's "The Hidden Truth," Jan. 14, 2002.
|
 |
Most Popular This Week
IE Weekly Newsletter
Subscribe to the newsletter
| |
| |||||||||||||||||||||||||||||||
