Contact Center Forecasting Model for Workforce Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for forecasting and analysis in contact processing centers, such as regression modeling, queuing equation modeling, and discrete-event simulation, are either limited in applicability, inaccurate, or cumbersome, failing to provide reliable forecasts and strategic planning due to simplifying assumptions and inability to account for complex workflows and abandonment rates, leading to inefficiencies in resource allocation and customer satisfaction.
Innovation Solution
A system and method that involves developing a computer model based on performance information from contact centers, allowing for iterative analysis and prediction of resource plans, process analysis, and financial planning, enabling the generation of multiple performance scenarios to evaluate resource levels and process parameters, and providing graphical analysis and reports for strategic decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If queuing equation methods (Erlang equations) are used for forecasting, then the system can provide workforce scheduling and personnel management functions, but the forecast accuracy deteriorates due to simplifying assumptions about caller abandonment and contact handling uniformity
Solution Approach 1:
The patent modifies the Erlang equations by introducing new parameters to account for caller abandonment behavior and differentiated contact handling. Instead of assuming no abandonment and uniform handling, the system incorporates abandonment rates, contact types, and routing strategies as variable parameters, thereby improving forecast accuracy while maintaining the operational framework of workforce scheduling.
Solution Approach 2:
The system transitions from static Erlang equations to dynamic forecasting models that can adapt to changing contact center conditions. The model incorporates time-varying parameters such as fluctuating abandonment rates, varying contact volumes, and dynamic routing strategies, enabling accurate forecasts under changing operational conditions rather than relying on fixed simplifying assumptions.
2Device complexity
If Erlang equations with simplifying assumptions are used, then the computational complexity is reduced, but the adaptability to modern contact center workflows deteriorates
Solution Approach 1:
The patent segments the contact handling process into distinct types (e.g., routine contacts, complex contacts, abandoned contacts) and applies different modeling parameters to each segment. This segmentation allows the system to capture modern workflow complexities such as skill-based routing and differentiated contact handling, while maintaining manageable computational complexity through structured modular modeling.
Solution Approach 2:
The system performs preliminary characterization of contact types and abandonment behaviors during the modeling setup phase. By pre-defining contact categories, routing strategies, and abandonment rate parameters, the system prepares the model structure in advance, enabling it to adapt to modern workflows without requiring complex real-time computations during actual forecasting operations.
3Measurement precision
If regression modeling is used for forecasting, then the method can handle stable workflows with consistent customer behavior, but the applicability deteriorates for contact centers with changing workflows and processes
Solution Approach 1:
The patent creates a dynamic forecasting model that can adapt to changing workflows by incorporating time-varying parameters and flexible contact type definitions. Unlike static regression models, the system can accommodate evolving contact volumes, changing abandonment patterns, and modified routing strategies, maintaining prediction accuracy across different workflow stability conditions through continuous model updates and adaptive parameters.
Data Source
AI summary
A method of predicting expected performance of a processing center system is provided. The method includes receiving performance information from a performance monitoring system associated with the processing center system. A computer model of the processing center system is developed based on the performance information. The method further includes generating predictions based on the computer model, and analyzing the predictions to generate performance scenarios for the processing center system.