Excerpted from Harrington, James H., "The Improvement Process"

One of the most important activities is the development of measurement system for setting priorities and gauging progress. Quality is defined as meeting customer expectations. The real job is to quantify these expectations and document them so employees know when they are conforming to expectations.

Many departments and individuals feel their jobs do not lend themselves to measurement. This is not true. And any job that cannot be measured is probably not worth doing and should be eliminated. The problem many people have is they don't understand the difference between activities and output. When you ask your data processing organization what they do, they talk about filling out forms, loading programs and inputting data. What they do is generate organized consolidated reports. At IBM when the data processing group determined that their measure of excellence was the percentage of acceptable reports delivered on time, things immediately began to get better. In just under 24 months the percentage of good reports delivered increased from, 86 to 97.4 percent. Any job that begins with an input, adds value to it, and passes it on to a customer can he measured. To quote one of the closing and dock supervisors at Avon, "I have learned how to measure so we could see how we are performing in the first place. I have not only established requirements, but now I have a measurement system in place."

Most activities have two key measurements - a productivity measure and a quality measure. Productivity is normally measured by dividing total output by total input. Quality, on the other hand, is measured in terms of percentage of good output. Real improvement occurs when both the productivity and quality measurements get better or when one of the measurements improves while the other remains unchanged. For example, an accountant could be processing 25,000 vouchers per month at a 0.5 percent error rate. She then decides to work harder on productivity and processes 35,000 vouchers that are 5 percent defective. The apparent result is improved productivity, but the overall result suffers.

The ideal improvement measure is the sum of all inputs divided into the quantity of output that met customer expectations, Unfortunately these types of indicators are often difficult to calculate. As a result, separate quality and productivity measurements need to be compared simultaneously so the true degree of improvement can be assessed.

At General Dynamics (GD) measurement is at the heart of their quality improvement process (QIP). Oliver C. Boileau, president of GD, reported, "There are as many as 60 QIP parameters and hundreds of projects in productivity improvement projects (PIP). Each general manager is measured on productivity/quality improvement process (P/QIP) results, and he, in turn, measures his staff on P/QIP accomplishments." Table 6-2 shows some of GD's measurement goals"

TABLE 6-2 Typical General Dynamics Measurement and Goals
GD organizational measures and goals
(corporate: common elements)
Organization Measurement parameter
Engineering Avoidable changes per month, %
Program office Deviation/waivers
Materials On-time deliveries to production, %
Accepted purchased items, %
Production Scrap (labor)
Scrap (material)
Rework and repair
First-time yield
Data systems center Software change requests
Quality assurance Inspection escapes
Number of MRB QARs per 1000 direct labor hours
Logistics Service report request time
General manager Overtime