Page 111 - Six Sigma Demystified

92 Six SigMa DemystifieD

Recall the earlier recommendations for metrics: The metric should be
customer- focused, integrated with the business strategy, and developed collab-
oratively. It should indicate performance over time in a direct fashion so that it
can be processed immediately. As discussed earlier, immediate feedback is most
useful for process control, particularly when employees are empowered to
respond directly to the information.
The usefulness of the data, however, is not merely a function of their avail-
ability. A reliable, repeatable, and reproducible measurement system is needed,
as described in the “Measurement System Analysis” section of this chapter. The
measurements must have sufficient resolution to detect changes in the process
effectively and differentiate between real changes and process noise, as dis-
cussed in the following sections.

ctQ metrics

Many production processes are evaluated by their yield or, similarly, their scrap
rate. Yield is calculated for production applications by dividing the amount of
product finishing the process by the amount of product that started the process.
For example, if 1,000 units started production, and of these, 980 units com-
pleted production successfully, then the yield is calculated as 980/1,000 = 98
percent. The scrap rate refers to the units that could not be sold (in this case,
2 percent).
Note that once the customer requirements have been determined, any pro-
cess can be defined in terms of its yield. For example, if 380 clients surveyed in
a random sampling of 400 clients of a customer call center indicated that the
service level was satisfactory or better, then the yield of the process can be

calculated as 380/400 = 95 percent.
A problem with the yield metric is that it does not provide enough detail on
the nature of the errors. As an example, consider the following three processes,
each with a 95 percent yield:

• Process A: Of 4,000 units started and 3,800 completed, 200 defective
units each had a single defect.

• Process B: Of 4,000 units started and 3,800 completed, 200 defective units
had a total of 600 defects.

• Process C: Of 4,000 units started and 3,500 completed with no defects,
300 units were reworked for 420 defects, and 200 units were scrapped for
580 defects.