Finding value in statistical anomalies

The ability to collect and categorize data with greater ease is helping companies around the word become data-driven entities. Customer service surveys and web-based tracking tools are allowing companies in a wide range of industries accumulate high volumes of data for making informed business decisions. In short, collecting data allow executives to draw meaningful conclusions that can lead to increased business value. When collecting data from customers, prospects, partners, or vendors, it's important for executives to develop an accurate picture of its business situation. To do this, executives must not only look at general themes derived from the data, but also look for hidden gems found in statistical anomalies. While the abundance of information with which executives are inundated makes it impossible to examine all data points, it's important for them to take into consideration statistical outliers that deviate from general trends. While many statistical outliers might provide a skewed view of the business situation, it's important for business leaders to understand that certain statistical outliers might provide key insights that are not reflected elsewhere.

Imagine that you’re an executive at a Fortune 500 company that's recognized for outstanding customer service. In fact, your company is considered the best in the business when it comes to customer service. Now imagine that a recent web-based survey that was sent to 10,000 customers revealed that 98.5% of all respondents confirmed that they're “extremely satisfied” with the company’s customer service—the highest designation in the survey. With such a large sample size acknowledging satisfaction in customer service, it appears that the executive team can assure itself the company is delivering solid customer service. The question becomes: do the statistical anomalies that deviate from the overwhelming majority even matter? More importantly, should the executive team consider feedback from a single disgruntled customer that appears to be providing a biased responses? If the goal is industry leadership, the answer is yes.

While the executive team might not be evaluating the core data, it's important for them to become aware of statistical anomalies. This is important because skewed data that might paradoxically conceal a flaw in the organization’s customer service paradigm (regardless of how impressive the service might appear). The reason for going to this level of analysis is simple. Companies looking to establish (or maintain) a leadership position must relentlessly pursue a level of excellence because they're not only competing for business today, they’re competing for business tomorrow. As a result, it’s important to accumulate as much data as possible to (1) construct an accurate picture of the entire situation and (2) determine if statistical outliers provide clues for gaps in the business model. Remember, the dismal of a single statistical outlier due to its size represents an exercise in flawed logic. As the global economy expands, executives must look at all information with an open mind—and adapt before the competition even has a chance.