During many Black Belt training sessions, the instructor teaches the class that it needs to develop statements on critical-to-quality elements (CTQs) that summarized the customer requirements in specific and measurable terms. But how? To facilitate this activity, the instructor introduces, a CTQ flowdown process. Unfortunately, experience suggests that the process capability is poor for this critical step in the Six Sigma methodology (Figure 1).
These requirements are usually grouped by common focus areas using an affinity diagram or similar tool. Project leaders generally struggle with how to extract the “raw” VOC data. The Six Sigma improvement methodologies are a process, and each step is an individual process. Collecting the raw VOC data varies from project leader to project leader, resulting in frustration, poor process capability, and projects that take too much time to complete, leaving a less than delighted customer.
Observations on Usual Collection Process
Here are a few observations that set the stage for improving this critical step in any improvement process using Six Sigma tools. Some VOC data collection processes may be flawed, since often the project leader is handed a draft charter to kick-off the project. Since variation is the enemy, then variation in this step is even more critical as the output from this step is an input to the project charter.
Customers do not necessarily know what they want. They know what they do not want (i.e., high prices, poor quality, late delivery, etc.). The project leader has to determine how the product or service affects customer behavior – not just what the customer says. The Japanese have a term for this – gemba. Gemba describes the true source of information. The gemba is where the product or service becomes of value to the customer, that is, where the product really gets used and delivers real value to the customer. It is in the gemba that a company really sees who their customers are, what their real problems are, how the product will really be used by them.
For example, a mechanical seal supplier provides a seal that requires plant maintenance to ship pump parts offsite to be machined to accommodate the supplier’s mechanical seal dimensions at a premium cost, and extends the cycle time to replace pumps in this chemical plant with more than 500 pumps. See the gemba.
The improved process map for collecting VOC data is shown in Figure 2.
A Guide to Enhanced VOC Data Collection
Suggested VOC questions are summarized under four headings.
Identify Customers by Functional Area:
- What is the organizational structure of the business?
- What are the customer touch points (supply chain, manufacturing, marketing, etc.)?
Develop Business-focused Questions for the Interviewee:
- What is the strategic vision of the business?
- What tactical activities are supporting this vision?
- What is the issue from your prospective?
- What products/services are affected?
- What pain are you feeling?
- Can you measure this pain?
- What is the financial impact of the pain?
- What is the impact on your business objectives?
- Do you have any business objectives addressing the issue?
- How long has the issue existed?
- How do you know?
- How much of your personal time is spent on the issue resolution?
- How much time do your direct reports spend on the issue?
- What is the impact on the business, if we do nothing?
Develop Functional-focused Questions for the Interviewee:
- Are your customers internal or external?
- Who are they?
- Do your customers know about the issue?
- What do/would your customers think of the issue?
- How do you measure the business impact of the issue (customer complaints, product returns, repair costs, etc.)?
- Can you describe the data collection process for these issues?
Group Raw Comments by Category:
- List CTQs for each category that are specific and measurable (For example: On-time delivery is not adequate.)
- Prioritize the CTQs with the project sponsor/process owner using a Pugh matrix or equivalent weighted-factor system.
By asking these questions and analyzing the answers, the quality of the VOC data is vastly improved. The raw VOC data will give the project leader tips on potential project Ys, data collection plan candidates, a sense for the process capability and “entitlement” of the process that is being improved. That, of course, will result in reduced cycle time and increased customer satisfaction level. Expect the VOC data to be in conflict. To further show the benefits of the improved process, the raw VOC data in the table below was obtained on real projects, resulting in separate fixed-cost focus and variable-cost focus improvement projects.
|Examples of Real-Life Issues and Benefits|
Enhanced VOC Project Benefits
|“The catalyst recovery equipment is not reliable and needs to be replaced. The centrifuges were only available 50 percent of the time in 1999.”||Tips on process capability and stakeholder analysis are indicated.|
|“Since January 1996, the profit objective catalyst ratio goal has only been met five times in 44 months.”||Tips on entitlement and performance standards are indicated.|
|“The catalyst ratio goal of 0.010 pounds of catalyst per hundred weight of finished product is not attainable. Catalyst will be costing us $22.18 per pound after September 1999.”||Tips on stakeholder analysis and initial financial validation are indicated.|
|“Based on the plant cost summaries, the centrifuge maintenance costs were $1 million in 1998. When catalyst recovery is bypassed, we are purging 38 pph of catalyst or wasting $834 per hour in variable manufacturing costs.”||Tips on potential project Ys, measurement systems and financial stake are indicated.|
The projects resulted in a catalyst ratio improved process capability from -1.18 to 3.40, resulting in $911,000 annual variable cost savings and $545,000 annual fixed cost savings, with the projects completed less than 12 months.
Not All Interviewees Will Embrace Process
A final word of caution: Not everyone interviewed will embrace the enhanced VOC data collection process. On one project, the project leader had to interview the “freight bill error clerk.” (Yes, that was the person’s job title.) The business had created a job description to handle the defects associated with the freight billing process. The project leader had to use these techniques to extract the raw VOC data from someone whose job assignment might be impacted by the improved process.
Project leaders need a little good luck with their gemba hunting.