Forecasting technology for the Future

Every day, the world produces enough data to fill 150,000,000 iPhones. That amount is increasing rapidly, too: The total volume of data worldwide equaled 4.4 zettabytes in 2013, according to Northeastern University. By 2020, that number will rocket to 44 zettabytes.

In conjunction with this growing reservoir of information, analytics and machine learning are reaching new heights. The technology can barely keep pace -- a recent Harvard Business Review article declared that “most industries are nowhere close to realizing the potential of analytics” – but thought leaders are accomplishing unimagined feats based on these new insights. The newest technologies are allowing analysts to identify patterns that produce accurate predictive and prescriptive feedback and automatically respond to change in real time.

Evolving forecasting tools

Forecasts and analytics are nothing new to contact centers – most have been tracking performance and call traffic for decades. New capabilities, however, are redefining the predictive sciences.

These improvements aren’t limited to the big names, either; the most robust contact center forecasting tools on the market have far-ranging impact for organizations of all shapes and sizes. Advanced statistical methods can help users realize consistent customer service, improve retention and lower costs through market-leading capabilities. These tools are increasingly accessible regardless of the size of the organization, allowing small and mid-sized companies to apply the same advanced statistical methods that contact centers two, ten or twenty times their size do.

With forecasting, managers make projections over a shorter range of time, such as those that cover a span of hours, days, weeks or months. Predictions, in contrast, belong in long-term planning -- often years in advance. Forecasting is most effective with large caches of rich data containing patterns that are likely to repeat. Next-generation technologies include the most advanced time-series mathematical methods. When backed by strong data, today’s market-leading forecasting technologies can automate ACD planning and skills-based routing, generate an unlimited number of “what-if” scenarios, automatically detect and smooth extreme outliers and missing values, and optimize scheduling and call distribution minutes, days, weeks or months in advance. The efficacy of these models is extraordinary, and predictive technologies are revolutionizing our world well beyond the contact center.

Predictive methodologies across industries

Among the advanced statistical methods used to predict seasonal changes in call volume is the Box-Jenkins autoregressive integrated moving average (ARIMA) method, which seeks out the best fit of a time-series model based on past data and performance. This same modeling solution is being used to combat crime in India after a recent demonetization project shined a spotlight on that nation’s currency counterfeiting crisis: the government saw a 21% increase in counterfeiting in 2014-2015 over the preceding year. Strategists in Gujarat identified an additional method of preventing fake money when they used a Box-Jenkins ARIMA model to process that state’s crime data pertaining to the counterfeiting of currency to forecast future crimes. The researchers’ findings have the potential to help police departments’ performance by guiding strategic deployment efforts and efficient investigation direction. It’s also one of the methods used by NICE EVOLVE WFM’s Forecaster.

Contact center tools can further refine their forecasts with exponential smoothing strategies that weight data points according to age, with weights decaying exponentially as the observations get older. Variations of this method have been used to predict extreme weather events, project electricity load forecasting and plan for seasonal product cycles. Holt-Winters forecasting, which uses exponential smoothing, has been applied to tourist attractions and seasonal travel events. This method was used to reduce error in forecasts of overnight stays of tourists in Republic of Slovenia. Such findings can be invaluable to the transportation and tourism industries, as well as government agencies charged with managing seasonal influxes of visitors.

The future of forecasting

Forecasting technologies are constantly improving, both in and out of the contact center, and new formulas and growing data sources will continue to increase accuracy. Whether the insight they offer is used to prevent currency manipulation or balance skills-based routing, they are redefining industries and ushering in a new age of data analytics. Unlike many new technologies, these tools are often accessible to small and mid-sized businesses. The movement is resulting in more accurate management strategies that offer a leg up for early adopters – regardless of size.

Paul Chance is a senior product marketing manager for NICE EVOLVE WFM, the leading software solution used by small and mid-sized contact centers to plan and manage the workforce anywhere from the cloud. For more information, visit