Call Center Operations
Call Centers are often the principal interface between a business and their customers, particularly in the service and finance industry sectors. Customer experience for a business can be irreparably damaged by inefficient or impersonal call center responses. It is said that after only two frustrated calls, 85% of those customers take their business elsewhere. (A 5% improvement in customer retention can improve profitability by 25 to 100% - Bain & Co.).
Call center performance is affected by many factors, but major factor is the fluctuations in call volume. Traditionally, as daily call volumes increase, more staff are hired. The limitations being the availability of suitable people and the associated training and other direct costs. Our client recognized these difficulties and sought alternative solutions.
Modern Analytics applied sophisticated time series analytical models to predict the likely volume of in-coming calls that would be received each hour of each day for six months in advance by the client's 500-person Call Center. By knowing well in advance the likely level of inbound and outbound call volumes per hour, the client can adjust staffing and call distribution to meet and exceed its customer response expectations without additional hiring.
These highly detailed Call Center predictions, for up to six months in advance, utilized Auto-Regressive Integrated Moving Average (ARIMA), Spectral Analysis and Recurrent Neural Networks time series models to predict daily call volumes and call distributions. The results were impressive in that 95% of the forecasts were within 3% of actual values. The client reduced labor costs, while increaseing response efficiency and customer satisfaction.
