background image for GxP Lifeline
GxP Lifeline

Quality by Design Part 1: You Can't Design Something You Don't Understand

Just like every other aspect of a product, quality is determined by the decisions you make in the design stages. While spending less time on planning upfront might seem to save you money, the costs associated with poor quality resulting from early design decisions can ultimately equate to 40 percent of your company’s total revenue. Correctly understanding the true nature of quality and addressing it in the design stages rather than trying to bring it in as an afterthought is the central premise of Quality by Design (QbD), and adhering to this principle could be the most cost-saving measure your company will ever take.

In a recent webinar , Craig Gygi, a Lean Six Sigma Sensei, best-selling author on continuous improvement, and experienced designer took to the whiteboard to illustrate the real cost of poor quality and the importance of planning for quality early in the design stages of any kind of product or service.

The Rule of 10s – $1 Issues That Cost You $1,000 When you think of design, you might think of a scientist in a laboratory, or an engineer creating plans and drawings on a computer. Anytime someone is creating something new, or making some kind of improvement, or developing a system, process or environment, design comes into play. Design is the series of decisions people make to plan and create something new, and those decisions determine how a product is going to perform – from how much it’s going to cost, to how much it helps a patient, to its quality, to how long it takes to develop. All of these things are determined in the design process. A typical design process consists of a concept phase, a detailed design phase, a prototype phase and a production phase (or some variation thereof). As the product moves through these phases, it becomes more and more finalized. Gygi explained that as the design progresses over time, the “Rule of 10s” dictates that the cost of fixing issues increases by a factor of 10 for each phase, such that an issue that costs you $1 to fix in the concept phase might cost you $1,000 in the production phase.

Image 1. The Rule of 10s.

“Sometimes, even if you want to make a change, you can’t. It’s just too late, things have gone too far,” Gygi said. “But in the concept stage, virtually anything is possible.

Making carefully considered decisions early in the design stages can help alleviate the need to make late and very costly changes to your product, or worse, to issue recalls, reimbursements or face damaging lawsuits.

To further express the cost of poor quality, Gygi compared it to an iceberg. The visible part of the iceberg represents the problems and associated costs that are readily apparent in a company with quality issues, including rework, scrap, noncompliance and warranty claims. Together, these amount to 5-15 percent of a company’s revenue – that’s $5-15 million for a company with a $100 million revenue. But this is just the tip of the iceberg, Gygi explained. Underneath the surface of these obvious issues lies a much larger problem: things like lost customer loyalty, excess inventory, cost of engineering change orders, extra equipment and extra headcount can claim another 15-25 percent of a company’s revenue. The entire iceberg of quality issues, totaling up to 40 percent of a company’s revenue, can be enough to sink the proverbial Titanic.  “The only way to fix this is in design,” Gygi said. “You can’t fix this after the fact.”

Image 2. The “iceberg” of poor quality issues and their associated costs, totaling up to 40 percent of a company’s revenue.

The Definition of Quality

So how do you address quality in the design stages in practice? In order to design for it, you must first define it, Gygi explained. Regardless of how you define quality, your design objectives should be driven by the needs of your customer. From those, you will typically create a product specification which precisely indicates a target level for each of the design’s critical characteristics, and both upper and lower limits of acceptable variation around the target values. These specifications apply to things like chemical composition, dimension, speed of performing an action, battery life and so forth, depending on the product.

A traditional misconception about quality, Gygi explained, is that if a critical characteristic of your product is outside of its specification – whether just barely or by a wide margin – it will incur a fixed cost (e.g., scrapping the part, issuing a warranty claim, etc.).

Further, it is believed that cost will disappear as soon as the product is within the upper or lower limits of acceptability – even if just barely inside the specs. Gygi likened this traditional view of quality to kicking a field goal in football. “It’s like football goal posts,” Gygi described. “If you kick the field goal and it bounces off the upright and goes in, you get three points. If it goes right down the middle, you still just get three points. But if you miss, zero points. That’s the traditional view of quality.”                                                        It follows, then, that the traditional definition of quality is:

Quality = compliance with specifications In reality, anytime a critical characteristic’s target is missed – even by the smallest of margins – it incurs a cost due to things like extra processing or a slightly different handling procedure, a cost that increases exponentially the farther the characteristic is from the target. Thus, the cost of a product that is just inside the specification and one that is just outside the specification is virtually indistinguishable.

Image 3. The traditional and realistic views of the cost of poor quality.

“The critical thing to understand about quality is that complying with specifications does not give you quality,” Gygi said. “It’s arbitrary.”

Therefore, Gygi suggested that a better definition of quality would be: Quality = on target with minimal variation Ford Motor Company To illustrate the difference between the traditional and improved definitions of quality, Gygi pointed to a real-world example. In the 1980s, Ford and Mazda worked together on a joint venture to produce a transmission that would be built into Ford’s Probe model and Mazda’s MX-6. The transmissions used in each model were based on the same engineering drawings and the same specifications, but some were fabricated, assembled and tested by Ford and some by Mazda.

Curiously, the performance of the Ford-produced transmissions was very different than of those produced by Mazda. Ford experienced poor transmission performance, warranty returns and lost customer loyalty; Mazda saw higher performance, fewer warranty returns, increased customer satisfaction and greater success overall as a business. To investigate the cause of these differences, tests were performed to evaluate the transmissions produced by each company.

The tests revealed excessive vibration in the Ford transmission, while that of the Mazda transmission was significantly lower. The difference, Gygi explained, was the way that each car company defined quality within its production efforts: Ford employed the definition Quality = compliance with specifications, while Mazda used Quality = on target with minimal variation. Once Ford realized the source of its transmission issues, the company was compelled to take remedial action.

“That changed everything for Ford. They changed their policy and you can see the changes to Ford over the years since,” Gygi said.

When you know these truths about quality, the results are the same regardless of domain, from a manufacturing process, medical device, pharmaceutical product, or blood and biologics product.

“The companies that do best, whether it’s a medical device company, a pharmaceutical company or a car company, work to have all their products and services on target with minimal variation,” Gygi concluded.

To learn more about the importance of QbD and including quality in the design of any product or service, view the complimentary webinar “Quality by Design – Part 1.” In future sessions in this series, Gygi will discuss how to turn the theory behind quality into tools and methodologies for designing high quality products and services. 


Beth Pedersen is a technical writer at the MasterControl headquarters in Salt Lake City, Utah. Her technical and marketing writing experience in the enterprise software space includes work for Microsoft, Novell, NetIQ, SUSE and Attachmate. She has a bachelor’s degree in life sciences communication from the University of Wisconsin-Madison and a master’s degree in digital design and communication from the IT University of Copenhagen.

[ { "key": "fid#1", "value": ["GxP Lifeline Blog"] } ]