case study

Industry: Finance

Account Withdrawal Accuracy

FOCUS: Improve the billing process for Administrative Services Only (ASO) accounts to reduce the number of incorrect and late administrative fee withdrawals.


A health insurance administrator for ASO accounts had a flawed billing process. The result was 25% of the customers had experienced incorrect administrative fees being charged to their account. An additional 8% of the customers were impacted by late withdrawal of fees. A project was launched to

improve internal efficiency (too many reconciliation checkpoints and manual workflows) and decrease customer dissatisfaction.


A survey was conducted to gain the Voice of the Customer and to help focus the project. To capture the current state, a high-level process map (value stream) was completed which included the five functional areas (sales, underwriting, installation, systems, and finance). Process time analysis was
then used to determine where bottlenecks existed. In the Measure Phase a cause-and-effect analysis was performed. The team determined that the distribution of hard copy change-request forms was a factor that had a big impact on inaccurate withdrawals. By performing a Measurement System Analysis the team determined which source of data was the most accurate, repeatable, and reproducible. The process accuracy was performing at 2.72 sigma. In the Analyze Phase hypothesis testing was used to show that the group cancelling policies had experienced a statistically higher rate of billing errors. Subsequently, it was determined that the change-request form and how it was distributed across the organization was a primary factor in billing errors. The Improve Phase witnessed four major actions: (1) the Request-for Change form was redesigned and automated, (2) the process was altered, (3) a Lotus Notes Database was created, and (4) the withdrawal process was automated. The process map was updated to capture the new state and the process was put in control. Both Finance and Underwriting review the process metrics monthly. The Lotus Notes database monitors the response time of the process owners.


The new process resulted in a financial benefit of $304,688 and over the course of 8 months no errors occurred. The cycle time for changes to be implemented went from an average of 2 weeks to 3 days.

*For additional information see iSixSigma Magazine, July/August 2006, pgs 47-55, “Account Withdrawal Accuracy” by Michael Stober.