Medical Device Link.

Originally Published MDDI January 2004

Product Development Insight

Product development teams should explore many design concept options to determine which best suits user requirements.

David Warburton

Perhaps the most crucial phase of the new-product development process is the concept selection phase. During this period, the array of design ideas are narrowed down to one concept that the design team predicts will best meet user requirements. This phase offers the team the most freedom to choose one design path over another. However, little objective data are available to evaluate the alternative paths, so the design process can be influenced as much by office politics as by engineering considerations.

This phase can be when brilliant, award-winning products are conceived, or when fundamental flaws are designed into the product. These are the kinds of flaws that delay the schedule by months, yet are never satisfactorily resolved, because they are inherent in the concept itself.

The most frustrating types of product flaws are those that could have easily been avoided had a little more thought gone into the original concept. I once asked an engineer to design a mechanism to open a loading door. The concept he chose involved an articulating linkage that tended to stall the drive motor if everything in the mechanism wasn’t set up just right. Subsequent design changes intended to prevent the motor from stalling only added cost and complexity without improving reliability. By the time the engineer was done refining this design, it was obvious even to nonengineers that there had to be a simpler way to open a loading door.

Much of a product’s success is dependent on the concept phase, yet the output from this phase can vary from exceptional to barely functional. Some product development teams seem to consistently produce good designs. What do these teams do differently?

The performance of the development team ultimately depends on the design skills of the individual engineers. For example, the engineer who designed the door linkage came up with one workable idea for the product. His initial review of the design suggested that it would meet the most important requirements. Consequently, he pushed forward and began detailed design of parts for the linkage using computer-aided design (CAD). When he built and tested the prototype, the motor-stall problem surfaced. He responded to the test failures with further refinements to the initial design because he was, at that point, psychologically committed to seeing this design succeed. Unfortunately, no amount of refinement could alter the basic physics of the design. The engineer finally had to start over with an entirely new concept very late in the process, delaying the schedule for the whole project.

Had I asked a different engineer to the design the same linkage, that engineer might have approached the problem differently. The second engineer might have spent time sketching one idea on a whiteboard and making crude paper and cardboard prototypes of another, before finally choosing a design concept to pursue. Although this engineer might have appeared to take much longer beginning the project, the final design would likely have had fewer problems. The team would have avoided starting over with an entirely new concept. 

The first engineer fell into the common design trap of locking into a single design concept too early without fully exploring all available options. The second engineer avoided the pitfall by exploring as many options as possible in the early phases of the design process. Using this approach, an engineer has a much better chance of selecting the optimal design. Yet the question remains: how does an engineer choose the best design?

Table I. In Pugh analysis, user requirements are weighted by importance (Click to enlarge).

There are a number of formal techniques for concept selection, but this article focuses on the methodology most commonly called the Pugh concept selection matrix. In the 1980s Stuart Pugh developed the methodology at the University of Strathclyde in Glasgow, Scotland. It provides a technique for choosing among design alternatives. 

Here are the principal steps:

• Identify relevant user requirements; develop engineering specifications for those requirements.
• Develop weights for each of the requirements.
• Generate several viable design concepts.
• Rank the concepts using Pugh analysis.
• Synthesize the best elements of each initial concept into a final optimal concept.
• Iterate until a clearly superior concept emerges.

Generally, the top-level user requirements and critical design specifications exist before the project enters the concept phase. As the project advances into the concept phase, the engineering team typically begins to divide the product into subassemblies. During this process, the team draws up more-detailed specifications for each element. For example, a team developing a portable dialysis machine would need specifications that address the user requirements shown in Table I.

The numbers after each concept represent the weighting factors assigned by the team, 5 being an important requirement and 1 being a requirement of minor importance.

Once the requirements are identified and target specifications are applied to them, concept brainstorming can begin. This can be the most stimulating part of the whole design project. For larger projects, I will often set up one or more team brainstorming sessions in order to get as many points of view as possible. For smaller projects, I ask the project’s engineer to generate three to five ideas. 

I typically ask the engineers not to use their 3-D CAD package for concept creation; instead, I ask them either to sketch ideas in a notebook or to make crude prototypes. This exercise breaks the cognitive bias that often results from the up-front effort required to create a design on 3-D CAD. In the time the engineers take to create the necessary models to illustrate an idea in a 3-D CAD package, they have already invested so much mental energy into that design that they are often compelled to just finish it. Allowing the engineers to work with 3-D CAD early in the concept phase results in the tools driving the creative process instead of aiding it.

In the case of the dialysis machine, the fictional team of engineers came up with the following concepts:

• Backpack.
• Four-wheel wagon with steerable front wheels, similar to a child’s wagon.
• Two-wheel hand truck.
• Suitcase.
• Motorized, three-wheeled cart with adjustable-height tiller.

Table II. The Pugh matrix uses a weighted system to help engineers evaluate and compare their design concepts. To calculate the score, multiply the ranking of each design concept by the assigned weight of each user requirement (Click to enlarge).

One design idea should be chosen as a baseline, to which the other concepts will be compared. The team may consider the baseline design to be the one determined most likely to succeed. Alternatively, the baseline design may be an existing product. Existing products are often used if the purpose of the development project is to improve upon that product. A competitor’s product may also be selected.

With the baseline selected, the team created the Pugh matrix, with the requirements in rows and the concepts in columns. They ranked the five concepts against the baseline concept and came up the results shown in Table II. The ranking was simple. If the concept met a requirement better than the baseline design, then it received a “+1.” Concepts that were viewed as equivalent received a “0,” while a concept that did not meet a requirement as well as the baseline design received a “–1.” Finer graduations may also be used.

As is often the case the first time the concepts are ranked, no concept was clearly superior. When this scenario occurs, the solution requires engineers to reexamine both the requirements and the concepts. First the team reviewed the user requirements to see whether all the relevant requirements had been identified and ranked appropriately. Second, the team considered whether successful elements of one concept could be applied to other concepts. This iterative evaluation process is one of the great strengths of the Pugh method of concept selection. Laying out all of the concepts into a single table and then ranking them forces the team to carefully evaluate and discuss each concept against the same set of detailed requirements.

When the dialysis team reevaluated the requirements, they determined that an additional requirement was relevant to the exercise. They added a system requirement for a 24-hour battery life, which was now relevant because the motorized cart would draw power from the batteries.

When the team examined each concept for successful design elements, they determined that the adjustable handle in the motorized cart could be applied to both the wagon and the hand truck concept. With this change, all of the designs were essentially equal in their ability to be used by both short and tall individuals. Table III shows the resulting revised matrix.

Table II. The Pugh matrix uses a weighted system to help engineers evaluate and compare their design concepts. To calculate the score, multiply the ranking of each design concept by the assigned weight of each user requirement (Click to enlarge).

In this fictional example, the results of the exercise do not indicate a clear leader. However, three of the five concepts have been decisively eliminated. The next step in the process would be to refine the two top concepts further, and then to compare those two concepts again, either with a new matrix or with actual prototype testing. 

Use of the Pugh technique accomplished several objectives.

• It compelled the design team to review the user requirements in detail and to understand how the requirements apply to the design.
• It compelled the team to look beyond the obvious first concept and to fully explore a wider range of concepts.
• It provided an objective way to evaluate those concepts.

The Pugh concept provides a concise, auditable document for the product’s design history file. This document demonstrates that the team used a process to achieve a result, and those results can be reviewed and understood. Pugh analysis also provides a clear format for presenting information and an evaluation and ranking process. Finally, it provides an easily understandable deliverable for the design history file to prove that the process has been completed.  

Copyright ©2004 Medical Device & Diagnostic Industry