Manufacturing Cost of Quality

In order to agree on the most important thing to work on first, we must have agreement on a company-wide method of evaluating the impact of a problem. The Cost of Un-Quality prioritizes improvement opportunities from segments of our company that are not traditionally measured in the same units.

Many improvement activities will require involvement of resources from many functional areas. The participants need confidence that their own problems, which are important to them, will be addressed in some fair way. It is important that time be spent getting consensus on the methodology used to choose priorities.

Two kinds of problems will emerge:

• An Event occurs occasionally and requires only containment actions. That is, we must immediately minimize its effect on operations and customers. Events are usually resolved before the measurements can point them out, but if we don't measure them, we might miss a pattern.

• A Pattern is evidence of problems that recur and require Problem-Solving. The object of the Cost of Un-Quality is to rank these recurring problems in order of their business impact.

To investigate a pattern, we assign resources (allocate Manpower, Money and Machines) based on the rankings, and look for the root-cause of the problem. Once the root-cause is identified, we move to problem-solving.

As we initiate solutions, we need to identify what changes we should expect in existing measurements. At the business level, some will clearly not change at all, but any improvement activity can cause some local measurements to improve, while others we exj2e_cl to see decline.

As an example, running smaller batches and increasing the number of set-ups is a typical strategy for inventory reduction. This clearly violates EOQ logic and quickly produces a negative effect on traditional efficiency measures. However, almost every­one who has done this, has more slowly achieved predictable and massive reductions in inventory and lead times, and gained improvements in quality and flexibility, as well.

In order to validate changes to our processes and make them permanent, we must understand the relationships between the activities and their results. We need to:

• Develop breakthrough measurements—What does success look like?

• Set milestones and timelines—What does progress look like?

• Match frequency and level of detail with measuring capabil­ity—Who can and should report? How often? To whom?

• Publish or post progress and trends—How are we doing?

Single-Minute Exchange of Dies, or SMED, was both a local set-up goal and a local measure of success. Implied were progress measures of time reductions, quantity of reductions achieved, and how much there was left to do. The global measures affected were lead times, inventory levels, delivery performance, lower total cost, and increased sales.

Some of these measures did not go in the intended direction at the beginning. Though most SMED activities are relatively low cost (and should be done immediately), some are quite expensive and require more analysis of costs and benefits both for the short-term, and into the future. Patience, demonstrated by milestones and timelines, allows a fair chance for changes to achieve positive results.

Measuring set-up time seems simple (start—finish), but to focus improvements, we first separate internal (machine needed) from external (preparation and parallel) activities. Any set-up reduction team will break these down further and find ways to measure the details. Warning: People not on the team will resist outside measurement, so bring in everyone who affects a set-up (tool room, blueprints, material movement, planning, etc.).

Seeing progress toward a goal improves morale, garners support, and increases total participation in all improvement activities. Widespread quality assurance attitudes instill internal and external

confidence that quality is constantly improving. Any quality breakdown is examined more closely because it is now not acceptable.

Caution: Some measurements may cause embarrassment. If area results, not names are posted, area participants will generate significant peer pressure on poor performers. They already knew who wasn't doing the job, and by-area posting gives them incentives for group correction. Using specific names causes a defensive posture from most people, even if they, themselves are good performers.