Top Global Outsourcing
Services Provider

The Challenge

Are your forecasts the best possible in your
contact center environment?

The WFM Global Optimization team is continually pursuing new technologies to increase the accuracy of their forecasts, and the responsiveness and operational efficiency of their Business Process Outsourcing (BPO) Service Delivery operations. After extensive testing, the WFM team concluded that the weighted averages based forecasting technology provided by their WFM system was ineffective and has poor accuracy. This led the WFM Global Optimization team to develop their own spreadsheet-based tools to determine time series patterns to generate their contact volume and handling time forecasts. They have achieved significantly higher forecast accuracy with their process. However, this new way of generating forecasts required a team of 4 full-time forecasters for one client they serve alone. Given that they serve over 65 Fortune 500 clients, using this process for the others was not resource feasible. Thus, they wanted to study the forecast accuracy level they could achieve with the advanced forecasting technologies provided by the AWO Forecaster. The desire, in this case, was to evaluate the opportunities offered by the advanced forecasting technologies pioneered by ac2 Solutions and determine if they can provide accurate forecasts for their clients. Another objective was to determine if they could automate most of the forecasting process without requiring excessive resources.

Opportunities Identified

The AWO Forecaster has the unique capability to analyze data using advanced statistical forecasting technologies to determine intra-day, day-of-week, week-of-year seasonal affect while capturing linear and non-linear trends. It has a build-in Expert System, Best Pick, to calibrate these models optimally and determine the one that has the best ability to generate forecasts. The AWO Forecaster also has models for special event period forecasting and extreme or missing data detection which improves accuracy.

To evaluate the accuracy of the forecasts generated by the AWO Forecaster, the WFM Global Optimization team developed a plan to compare their forecasts with the forecasts generated by the AWO Forecaster over a 3-month period. Forecasts were generated weekly, 4 weeks out into future for one week to mirror the Company’s forecasting process. For benchmarking, the WFM team selected one of their clients that encompassed:

• 1 contact center environment
• 2 contact groups served by 4 contact centers
• Approximately 20,000,000 calls per year

Using the same historical data and mirroring the forecasting process followed by the Company, forecasts for this environment were generated by the AWO Forecaster and the Company’s forecasting team, one week at a time. No information other than common holidays such as the New Years Day, Easter, etc. was provided to the ac2 team about the contact volume drivers. All AWO Forecasts were strictly machine generated with no institutional knowledge the WFM team may have used.

Each week, after providing forecasts for the 5th week, the ac2 team was given one more week of actuals as the Company was receiving. The weekly forecasting and acutuals updating process continued for three months. Forecasts generated by the AWO Forecaster were evaluated by the WFM Global Optimization team for conformance to their forecasting process and accuracy.

The AWO Forecaster showed significant advantages and improvement opportunities over the WFM Global Optimization team’s forecasting process. Major benefits identified included:

• Over 95% forecast accuracy with blind machine forecasting: The AWO Forecaster achieved higher forecast accuracy greater than 95% over the WFM forecasting team without knowing the client’s customers and their contact behavior.
• 99% reduction in forecast analysis cycle: The usual one full week analysis time to generate forecasts was reduced to minutes by applying AWO’s build-in Expert System, Best Pick, and advanced forecasting technologies.
• 100% reduction in full-time resources needed: The usual one full week analysis cycle required 4 full time forecasters. The AWO Forecaster required only a few minutes to generate the forecasts and verify the results for a week. This was achieved by applying AWO’s build-in Expert System, Best Pick, and advanced forecasting technologies to analyze and capture seasonal patterns and trends in the data using advanced statistical techniques without requiring any manual input.

Using the data driven benchmarking approach, the Company developed a road map to improve forecast accuracy for all its clients. Consistent with this road map, the WFM Global Optimization team piloted the AWO Forecaster for 6 months and verified that their team would achieve the same accuracy levels in their operations before applying their institutional knowledge.