2019

ONE GOOD IDEA

Get Specific

Ask the right questions to ensure successful product testing

By Guy Gimson

Whether it’s a product development group or an R&D group, quality control or quality assurance personnel will typically be responsible for determining whether a lot of product conforms to specifications. That’s what classic quality control is.

But who decides what’s in your product specifications? How do you identify and categorize the key product characteristics and attributes? What belongs in the reaction plans if a test result falls outside the allowed limits? Answering those questions is the first step toward the elimination of nonvalue-added product testing.

Getting specific

A product specification exists to define:

  1. A product’s properties. That could mean its attributes (for example, a Model T is black, not white) or its variables (its height at the roof is 5 feet, 6 inches).
  2. The tolerances around the variable properties (for example, plus or minus 0.5 inches) or specification limits.
  3. The test methods used for each of the property tests.
  4. The materials, which include formulas, used to make the product.

Testing variables is, of course, better than testing attributes, and a product specification might contain more than this list—a digital photo is often a wonderful addition—but these are the essentials.

Determining the ‘in’ crowd

The first category of properties to include in the specification is what matters to the customers. Quality function deployment is one of the best tools for determining what exactly that is. As customers use the product, they will send feedback regarding what is really important to them and what isn’t. We’ll refer to what is important to customers as customer-critical properties.

The manufacturing process will also allow you to develop a category of properties that’s important for maintaining consistent production but does not directly affect the customers. These properties will be tested for process control purposes.

There might be properties you need to maintain within limits, or above or below a certain limit, because they demonstrate conformance to an American Society for Testing Materials (ASTM) standard, maintain an Underwriters Laboratory listing or keep you on the right side of a regulatory agency. We’ll refer to this category of properties as regulatory compliance.

Finally, there might be a category of properties fundamental to the product that is only affected by manufacturing process changes or raw material changes. In other words, the properties show little common cause variation but might be subject to assignable cause variation. We’ll refer to these as product-validation properties.

Knowing your limits

The specification limits on customer-critical properties are set by the customers. You’ll accept or reject a lot of the product if these limits are exceeded. The reaction plan should spell out what action to take, regardless of whether the limits are exceeded on the upper or lower end. You might rework, scrap or cull the lot, but it will not be shipped as-is without the express consent of the customer.

Specification limits are not meaningful for process-control properties; only control limits are. Testing is performed to determine whether the process remains in control. The reaction plan is a root cause investigation, regardless of whether the customer-critical product properties remain within specification limits.

For regulatory compliance, specification limits are set by the relevant ASTM standard or something similar. If a lot of product does not meet the limits, the only option available is to remove the standards compliance statement if company policy allows it. Again, the reaction plan is a root cause investigation; selling product as “seconds” is poor business.

Product validation is the toughest category to define, but in some ways it is also the most interesting. As long as product properties remain constant, the process and product are stable. The rule with this category is to test the properties infrequently but mandate full testing in the event of a change in the raw material or process.

In my experience, using this method eliminates between 25 and 50% of product testing. It’s been used in Food and Drug Administration-regulated consumer products and ASTM standards-driven building materials. Give it a try and see how effective it can be.


Guy Gimson is the owner of Quality First LLC in Charleston, SC. He is a senior member of ASQ and Six Sigma Black Belt.



good informative article

Aylin N. M.
CQPA, CQE, CQA
--AYLIN NAFIA MINIKUS, 01-14-2018

Average Rating

Rating

Out of 1 Ratings
Rate this article

Add Comments

View comments
Comments FAQ


Featured advertisers