The managers of a lockbox operation in the upper Midwest were not at all pleased. The corporate clients for whom they processed millions of dollars in payments were very intolerant of errors. Yet in the first months of a new quality improvement effort, the error rate per 100,000 transactions doubled from 15 to 30. Truth be told, the increase in errors was perception, not reality. Errors had always been worse than believed. And now improved data collection was showing that the process performance was below expectations (see figure below).

Process Out of Control Limits
Process Out of Control Limits

It was fortunate that actions launched in the Improve phase of DMAIC (Define, Measure, Analyze, Improve, Control) project quickly brought the rate back to 20. Unfortunately, even that rate was far from the target of 7.5 errors per 100,000 transactions – an industry benchmark for best-in-class performance.

The Six Sigma project team working on the problem was not finished, however. Though many teams skip through Control doing the bare minimum to document results, this team’s Black Belt encouraged them to give Control their full attention. The team’s work was guided by two questions: “What can we do to maintain what we’ve already put in place?” and “What can we do to build continuous improvement into our processes?” The actions the team took had surprising – and welcome – results. Here’s a quick look at what the team did.

Maintaining Gains

When the team had first started in the Measure phase, it was surprised to learn just how much worse the error rate was than anyone thought. When the team reached Control, it was determined to find ways to help everyone in the department monitor exactly what was going on. The team developed and implemented standard operating procedures (SOPs) for collecting and displaying data on error rates (including a new set of standard control charts to be maintained by department supervisors).

When reviewing the results from the full-scale implementation of the new procedures it had developed in Improve, the team noticed something else. There was a big gap between the error rates of top performers and poor performers. The gap was there before any changes were made and continued afterwards, so obviously the team’s original plans had not addressed that problem. To address this gap, the team undertook a major effort to raise everyone in the department to the level of top performers. Their countermeasures included:

  • Documenting best practices among the top performers
  • Revising the SOPs based on that knowledge
  • Training everyone on the new SOPs, with special attention paid to the poor performers
  • Collecting data on an ongoing basis to see if the error rate improved further

The team also had discovered that one reason for delays in processing was the irregular-but-frequent breakdown of the machines and printers. So a third major thrust of the Control phase was to develop SOPs for preventive maintenance, identifying who would be responsible for those procedures and training them.

Doing the Groundwork for Continuous Improvement

One approach the team took that proved highly successful was working with department supervisors to develop SOPs that would make the corrective actions an integral part of how work was done. So the team not only,

  • Initiated new methods for collecting and displaying data, it provided a process for developing data collection measures.
  • Helped improve the performance of people who had been performing poorly, it established triggers that in the future would alert supervisors when someone exceeded acceptable error rates (an “early warning” system) and recommended actions on how to deal with the situation.
  • Captured current best practices, it identified how to know when a new best practice appeared and how to take advantage of it.

In short, the team developed a set of “super SOPs” that helped establish new management practices in the division.

All of the new SOPs were built into a documented control plan that listed each major step in the process, its quality goal (specification), measurement method for tracking performance against the spec, and recommended corrective actions if and when a problem appeared. (See excerpt of the plan in table below.)

Control Plan
Process/ Process Steps Characteristic/ Parameter CTQ Specification/ Requirement Measure/ Method Sample/ Size Frequency Who/ Measures Where/ Recorded Decision/ Ruler/ Corrective Action
PO delivers mail Properly delivered mail C-I Zero pounds of improperly delivered mail Pounds of mail to be returned 100% Daily Mailroom Mail log and PO scorecard Return mail and review scorecard quarterly with PO
Process incoming special packages Special package SLAs C-E No SLAs missed due to unprocessed special packages Compare to agreed service level times 100% Daily Coordinator Special tracking log Monitor specials through processing and expedite work
Fine sorting mail Sort mail C-I No mis-sorted mail in coops Manual inspection Ad hoc Daily per shift Inspector In-line edit log Mail placed in proper coop and mis-sorted charged to offender
Distribution of work to modules Unprocessed mail C-E No mail in coops 100% Daily per shift Supervisor Zero deposit report Unprocessed work found: expedited. If none found, supervisor signs report.

Management’s Role

Project team members alone could not tell coworkers that data on their individual performance was going to be posted in the work area. Or tell their own supervisor about changes they wanted to make to his or her job description. The kinds of changes that often accompany Control directly impinge on how people do their jobs. If approached poorly, the result is ruffled feathers at best, or outright hostility at worst. That is why it is critical that local management be involved in Control measures, minimally through vocal support and ideally through participation.

In this company, for example, the management team took the control plan from this Six Sigma project team very seriously. Supervisors were involved in developing the new SOPs. Management played a critical role in communicating the need for these measures, and establishing the expectation that “this is how work is going to be done here from now on.”

Results Do Not Go Unnoticed

In the months during and following this team’s Control work, the error rate in the lockbox operation reached an historic low point – just 8 errors per 100,000 transactions. The group has continued to maintain that low rate. Better still, the results of the team’s work did not go unnoticed. A large corporate customer looking for lockbox services picked this bank over a competitor primarily because of the control plan. The documentation and rigorous use of continuous improvement demonstrated to the customer that this bank cared about quality and was constantly working to improve.

About the Author