Too much of a good thing?

Our natural inclinations tell us that the more information we have, the better decisions we can make, according to the author. However, thanks to one grocery store study, the opposite might be true. Read how a jam taste test proves that information overload can actually stymie a fraud examination.

In 2005, Sheena Iyengar, a professor of business at Columbia University, set up a taste-testing table at a gourmet supermarket in California. Customers were met with either a selection of 24 different jams to taste or a selection of six jams to taste depending on when they walked through the store.

When 24 samples were made available, 60 percent of patrons stopped at the table to sample the jam and to receive a $1 coupon. Alternatively, 40 percent of patrons stopped when six samples were presented. Considering more customers stopped when there were more samples, it'd seem likely that having a greater number of jams to taste would lead to a higher sales percentage.

In fact, the opposite was true. Thirty percent of customers purchased a jar from the table with six jams while only 3 percent of customers purchased a jar from the table with 24 samples. (See Too Many Choices: A Problem That Can Paralyze, by Alina Tugend, The New York Times, Feb. 26, 2010.)

Our natural inclination tells us that we can make better decisions when we have more information. But the study above proves the opposite: We actually can have too many choices. Psychologist Barry Schwartz refers to the problem of having too much information as "choice paralysis." His and other studies show that we're less likely to take action or feel satisfied with our decision when we have too many choices and too much information. (See The Surprising Poverty of Too Many Choices, by Kristi Hedges, Forbes, Nov. 26, 2012.)

Because so much information is available, fraud examiners might hesitate (choice paralysis) to use the data they have believing that eventually there'll be even more — and better — information, which would help them make wiser decisions. Having so much information at their disposal, fraud examiners often get stuck trying to decide when it's prudent to delay or act.

Fraud examiners aren't the only people experiencing information overload. In a 2010 survey conducted by LexisNexis, 62 percent of employees claim their work quality suffers because they can't sort through all the information available to them. (See Is Information Overload Killing Your Productivity? by Lindsay Broder, Fox Business, Sept. 25, 2013.)

The information available in our technologically driven world is almost infinite. And fraud is typically referred to as a "paper case" because organizations collect so much data, which leaves fraud examiners lost.

The potential of data analytics

Many are bullish on the potential of using data analytics as a tool for fraud prevention and detection. "Proactive data analytics is one of the principal tools in fraud detection and prevention," according to author Gary Bauer in his Internal Auditor magazine article, Fraud Detection and Data Analytics, June 2016. "The Association of Certified Fraud Examiners … 2014 Report to the Nations found it to be one of the most effective anti-fraud controls," Bauer writes.

With the ability to search data sets and uncover significant correlations, fraud fighters and investigators can potentially detect and monitor fraudulent activity in real time. Companies have invested millions of dollars to gather as much data as possible to develop tools to minimize fraud losses. In some cases, data analytics has proven to significantly reduce the amount of fraud perpetrated against organizations.

For example, data analytics helped forensic accountants capture a fraud scheme at a major call center when the accountants applied an algorithm based on Benford's Law to the first layer of invoices. "The accountants identified a handful of operators — less than a dozen — who had issued fraudulent refunds to themselves, friends and family totaling several hundred thousand dollars," according to Jo Craven McGinty, author of an article in The Wall Street Journal. (See Accountants Increasingly Use Data Analysis To Capture Fraud, by Jo Craven McGinty, The Wall Street Journal, Dec. 5, 2014.)

Yet, data analytics is no silver bullet. PricewaterhouseCoopers (PWC) and Iron Mountain conducted a survey of 1,800 business leaders in North America and Europe concerning their organizations' data analytics efforts. The results were somewhat surprising. The survey found that "overall, 43 percent of companies surveyed ‘obtain little tangible benefit from their information,' while 23 percent ‘derive no benefit whatsoever,' " according to Study Reveals That Most Companies Are Failing At Big Data, by Sarah K. White, CIO magazine, Nov. 10, 2015.

If we were to look further at the cause for these failures, could data analytics hamper fraud detection and prevention by providing too many choices through access to massive amounts of data?

To take action or not to take action — the choice paradox

At some point, fraud examiners have to determine if the information they possess, for better or worse, is enough. While they might later discover more relevant data, they eventually must take the important psychological step of acting rather than continually waiting.

According to a New York Times article, "Benjamin Scheibehenne, a research scientist at the University of Basel in Switzerland, said it might be too simple to conclude that too many choices are bad, just as it is wrong to assume that more choices are always better. It can depend on what information we're being given as we make those choices, the type of expertise we have to rely on and how much importance we ascribe to each choice." (See Too Many Choices: A Problem That Can Paralyze, by Alina Tugend, The New York Times, Feb. 26, 2010.) Fraud examiners' life experiences, education and skills are all factors when they choose if they have enough information to act.

Some actions won't lead to anticipated results, but that doesn't necessarily mean they took the wrong actions. Fraud examiners who take no actions while continually waiting for more and better information are at equal risk of jeopardizing their cases as those who choose to act on information without necessarily achieving the desired results.

Barry Schwartz, a professor of psychology at Swarthmore College and author of "The Paradox of Choice," says that seeking the perfect choice in big decisions is a recipe for misery. (See The New York Times article.) Just as completely accurate templates for fraud examinations don't exist, we don't have accurate templates for telling fraud examiners when they need to step out of their choice paralyses.

Eliminating choice paralysis

Fraud examiners, faced with an abundance of data, can employ several methods to prevent information overload from stymieing the progress of their fraud examinations.

First, they can limit their choices. A fraud examiner should consider what areas of the business are at greater risks for fraud and what kinds of information could be analyzed to determine whether fraud is being committed. "Begin with outcome-based thinking. Pinpoint one or two business problems that may be solved by better fraud detection," according to Strategies for Detecting and Preventing Fraud With Data Analytics, by Emily Washington, infogix, Dec. 16, 2015.

If you focus only on the relevant areas of potential fraud in organizations then you reduce the amount of information available for analysis. This practice isn't meant to be absolute; rather, using existing data to analyze the areas most susceptible to fraud increases the chances of detection. But if gathered information indicates that other relevant information exists in different areas, you can still gather it. Segregating that particular information from routine analysis reduces choice paralysis by limiting the routinely produced options for review.

Data analytics also can present a larger problem if organizational culture is dominated by silo-based approaches versus enterprise-based approaches. If executives and/or organizational leaders protect or hold data tightly, the incomplete information sets make any data analysis unreliable. "Organizations' data silos and partially integrated information systems cannot look at data holistically, which means they are missing the larger picture, including ways to stamp out fraud," writes Emily Washington in the infogix article.

If the data isn't necessarily reliable, fraud examiners are forced to further assess their choices because they know that the information provided through data analysis might not be accurate.

Because data analytics in its current form is a relatively new technology, organizations are still learning ways to best leverage its competitive advantages. They've expended large amounts of money on building data analytics structure; many corporations have competed with each other to obtain the best data scientists. However, of course, spending the big bucks doesn't necessarily solve the data-analysis problems of a modern corporation. "Businesses might be investing significant money into data capture, but then drop the ball when it comes time to actually use that data," according to Sarah K. White's CIO magazine article.

If an organization wants to best leverage its data analytics programs then it needs to take a multi-stakeholder approach. "It's also not about the data. Rather, it's a way to predict future strategies and support decision-making. That's why the right stakeholders need to be involved," according to 4 Reasons Data Analytics Often Fail, by Tod Newcombe, Governing magazine, May 2016.

If data scientists are allowed to collaborate with fraud examiners as well as other important internal stakeholders, companies can create a more robust fraud detection and prevention program based in data analytics, and choice paralysis could be mitigated.

Real-world choice paralysis

In February 2014, Wells Fargo hired its first chief data officer. Prior to the hiring, Wells Fargo had invested millions of dollars building its data infrastructure and needed someone on an executive level to ensure the company extracted the intended value via data analytics.

Wells Fargo executive management believed the investment in data analytics would not only help the company but better ensure compliance with external and internal banking regulations. The hiring of Charles Thomas, the former chief data analytics officer at USAA, was hailed as Wells Fargo's transformational change to the new era of big data. (See How The Chief Data Officer at Wells Fargo Views Big Data, by Kristin Potts, The Platfora blog, March 13, 2015.)

Wells Fargo serves close to 100 million customers; each creates their own set of data. The total amount of information available to analyze can become overwhelming.

The company immediately ran into difficulty when it began working with data analytics in 2014. "All we were doing is collecting all the data we could get our hands on and integrating it and trying to be able to come up with the stories about our users and our users' behavior," said Christine Birtel, former head of customer insights & analytics at Wells Fargo in the Tableau article, Storytelling with big data at Wells Fargo.

Since their initial efforts, Wells Fargo and other businesses have learned ways to leverage their data for useful purposes. But as we've seen, no system, including Wells Fargo's, is perfect.

On Sept. 20, 2016, Wells Fargo CEO John Stumpf was called to testify before the U.S. Congress to explain how as many as 5,300 Wells Fargo employees were able to create more than two million fraudulent accounts in bank customers' names without the customers' knowledge. (Stumpf subsequently resigned on Oct. 12, 2016.)

According to some former Wells Fargo employees, the fraudulent accounts were created within a corporate culture that required employees to sell at least eight Wells Fargo products (second bank accounts, credit cards, investment accounts, etc.) to their customers each month. The program was internally referred to as the "Eight is Great" program. (See Former Wells Fargo Employees Describe Toxic Sales Culture, Even At HQ, by Chris Arnold, NPR, Oct. 4, 2016.)

In an interview in American Banker on Nov. 3, 2014, Chief Data Officer Thomas discussed how data analytics could enhance the "Eight is Great" cross-selling program. "His [Thomas'] team's first priority is delivering more complete and useful cross-selling information to employees on the front line, to help them make better-informed product recommendations," according to the article. (See Monetizing Big Data: A Q & A With Wells Fargo's Data Chief, by Penny Crossman, American Banker, Nov. 3, 2014.)

Thomas believed that data analytics would help Wells Fargo employees decide which products its customers needed and didn't need. "If you bring together all this data, does that mean you're going to try to sell me more stuff? My answer, we should be trying to sell less stuff. In other words, relevance and timeliness are really critical," Thomas said.

Thomas might have had good intentions when it came to using data analytics with a competent team that could harvest and interpret that data. However, the sheer volume and the choices this data provided might have had an impact on why the company didn't detect the fraud via analytical methods. When Wells Fargo decided to assimilate as much data as possible, it inadvertently created data sets that weren't relevant. Without proper guidance and/or fraud experience, Wells Fargo's data analysts succumbed to choice paralysis when the amount of information detracted from their ability to focus on the data that would've discovered the fraudulent accounts.

The value of controlling the number of choices available

Fraud examiners are always trying to perform the best fraud examination possible by discovering as many pieces of evidence as they can. The information available at fraud examiners' fingertips is infinite. However, if they aren't careful, the endless information can consume fraud examinations and lead to choice paralysis.

Data analytics, however, has proven to be a valuable tool. "Victim organizations with this control (data analytics programs) experienced losses of 60% lesser value and schemes 50% shorter in duration than organizations that did not," according to the CIO article.

Still, having too much information is just as debilitating as not having enough. As in all cases, fraud examiners should identify the investigative objectives and determine the best routes to obtain the evidence to support or refute the allegations.

If fraud examiners can learn anything from a taste test involving jam, maybe it should be that data analytics is a great tool that offers a number of choices, but the options are only as effective as the fraud examiner who knows how to manage them.

Bret Hood, CFE, is the director of 21st Century Learning & Consulting and a retired FBI supervisory special agent for the FBI Academy's Leadership & Communications Unit. He's a federal court-recognized expert in fraud and money laundering. His email address is: 21puzzles@gmail.com.

Learn more about overcoming choice paralysis when using data analytics in the ACFE's Fraud Talk podcast interview with Bret Hood.