A control chart pattern recognition methodology for controlling information technology-enabled service (ITeS) process customer complaints

Purpose - The purpose of this paper is to develop a control chart pattern recognition methodology for monitoring the weekly customer complaints of outsourced information technology-enabled service (ITeS) processes. Design/methodology/approach - A two-step methodology is used to classify the processes as having natural or unnatural variation based on past 20 weeks' customer complaints. The step one is to simulate data on various control chart patterns namely natural variation, upward shift, upward trend, etc. Then a deep learning neural network model consisting of two dense layers is developed to classify the patterns as of natural or unnatural variation. Findings - The validation of the methodology on telecom vertical processes has correctly detected unnatural variations in two terminated processes. The implementation of the methodology on banking and financial vertical processes has detected unnatural variation in one of the processes. This helped the company management to take remedial actions, renegotiate the deal and get it renewed for another period. Practical implications - This study provides valuable information on controlling information technology-enabled processes using pattern recognition methodology. The methodology gives a lot of flexibility to managers to monitor multiple processes collectively and avoids the manual plotting and interpretation of control charts. Originality/value - The application of control chart pattern recognition methodology for monitoring service industry processes are rare. This is an application of the methodology for controlling information technology-enabled processes. This study also demonstrates the usefulness of deep learning techniques for process control.