Technology

page 250

1. stage-gate processes

2. quality function deployment (“house of quality”)

3. design for manufacturing

4. failure modes and effects analysis

5. computer-aided design/computer-aided manufacturing

TOOLS FOR IMPROVING THE NEW PRODUCT DEVELOPMENT PROCESS

Some of the most prominent tools used to improve the development process include stage-gate processes, quality function deployment (“house of quality”), design for manufacturing, failure modes and effects analysis, and computer-aided design/computer-aided manufacturing. Using the available tools can greatly expedite the new product development process and maximize the product's fit with customer requirements.

Stage-Gate Processes

go/kill decision points Gates established in the development process where managers must evaluate whether or not to kill the project or allow it to proceed.

As discussed in a previous section, escalating commitment can lead managers to support projects long after their expected value has turned negative, and the cost of pushing bad projects forward can be very high. To help avoid this, many managers and researchers suggest implementing tough go/kill decision points in the product development process. The most widely known development model incorporating such go/kill points is the stage-gate process developed by Robert G. Cooper.25 The stage-gate process provides a blueprint for moving projects through different stages of development. Figure 11.2 shows a typical stage-gate process.

At each stage, a cross-functional team of people (led by a project team leader) undertakes parallel activities designed to drive down the risk of a development project. At each stage of the process, the team is required to gather vital technical, market, and financial information to use in the decision to move the project forward (go), abandon the project (kill), hold, or recycle the project.

FIGURE 11.2 Typical Stage-Gate Process, from Idea to Launch

In Stage 1, the team does a quick investigation and conceptualization of the project. In Stage 2, the team builds a business case that includes a defined product, its business justification, and a detailed plan of action for the next stages. In Stage 3, the team begins the actual design and development of the product, including mapping out the manufacturing process, the market launch, and operating plans. In this stage, the team also defines the test plans utilized in the next stage. In Stage 4, the team conducts the verification and validation process for the proposed new product, and its marketing and production. At Stage 5, the product is ready for launch, and full commercial production and selling commence.26

Preceding each stage is a go/kill gate. These gates are designed to control the quality of the project and to ensure that the project is being executed in an effective and efficient manner. Gates act as the funnels that cull mediocre projects. Each gate has three components: deliverables (these are the results of the previous stage and are the inputs for the gate review), criteria (these are the questions or metrics used to make the go/kill decision), and outputs (these are the results of the gate review process and may include a decision such as go, kill, hold, or recycle; outputs should also include an action plan for the dates and deliverables of the next gate).

Because each stage of a development project typically costs more than the stage preceding it, breaking down the process into stages deconstructs the development investment into a series of incremental commitments. Expenditures increase only as uncertainty decreases. Figure 11.3 shows the escalation costs and cycle time for each stage of a typical development process in a manufacturing industry.

Many companies have adapted the stage-gate process to more specifically meet the needs of their firm or industry. For example, while managers at Exxon were strong advocates of using a stage-gate process to track and manage development projects, they also felt that the standard five-stage system did not adequately address the needs of a company in which basic research was a primary component in generating innovations. Exxon managers created their own extended stage-gate system to include directed basic research. The resulting stage-gate system included two basic research stages (Stages A and B in Figure 11.4) and five applied research and development stages. In Stage A, the company identifies the potential business incentives and competitive advantages of an envisioned technology. The company then develops a basic research plan that establishes specific scientific deliverables, the methods of achieving these deliverables, and the required resources. In Stage B, Exxon's research division begins to execute the plan developed in Stage A, using scientific methods to generate leads for addressing the business opportunity. Stage 1 then identifies the best leads, using “proof-of-principle” assessments to establish whether the leads are feasible.27 Stages 2 through 5 proceed according to a typical stage-gate process.

FIGURE 11.3 Escalation of Development Time and Costs by Stage

Source: From Frederick D. Buggie, “Set the ‘Fuzzy Front End’ in Concrete,” Research Technology Management, vol. 45, no. 4, July–August 2002. Reprinted with permission of Industrial Research Institute.

page 252

FIGURE 11.4 Exxon Research and Engineering's Stage-Gate System

According to studies by the Product Development and Management Association, nearly 60 percent of firms (including IBM, Procter & Gamble, 3M, General Motors, and Corning) use some type of stage-gate process to manage their new product development process. Corning has made the process mandatory for all information system development projects, and Corning managers believe that the process enables them to better estimate the potential payback of any project under consideration. They also report that the stage-gate process has reduced development time, allows them to identify projects that should be killed, and increases the ratio of internally developed products that result in commercial projects.28

Quality Function Deployment (QFD)—The House of Quality

QFD was developed in Japan as a comprehensive process for improving the communication and coordination among engineering, marketing, and manufacturing personnel.29 It achieves this by taking managers through a problem-solving process in a very structured fashion. The organizing framework for QFD is the “house of quality” (see Figure 11.5). The house of quality is a matrix that maps customer requirements against product attributes. This matrix is completed in a series of steps.

page 253

1. The team must first identify customer requirements. In Figure 11.5, market research has identified five attributes that customers value most in a car door: it is easy to open and close, it stays open on a hill, it does not leak in the rain, it isolates the occupant from road noise, and it protects the passengers in the event of crashes.

FIGURE 11.5 Quality Function Deployment House of Quality for a Car Door

1. The team weights the customer requirements in terms of their relative importance from a customer's perspective. This information might be obtained from focus group sessions or direct interaction with the customers. The weights are typically entered as percentages, so that the complete list totals 100 percent.

2. The team identifies the engineering attributes that drive the performance of the product—in this case the car door. In Figure 11.5, four attributes are highlighted: the weight of the door, the stiffness of the door hinge (a stiff hinge helps the door stay open on a hill), the tightness of the door seal, and the tightness of the window seal.

3. The team enters the correlations between the different engineering attributes to assess the degree to which one characteristic may positively or negatively affect another. The correlations are entered into the matrix that creates the peaked roof of the house. In this case, the negative sign between door weight and hinge stiffness indicates that a heavy door reduces the stiffness of the hinge.

4. page 254The team fills in the body of the central matrix. Each cell in the matrix indicates the relationship between an engineering attribute and a customer requirement. A number (in this example, one, three, or nine) is placed in the cell located at the intersection of each row (customer requirements) with each column (engineering attributes), which represents the strength of relationship between them. A value of one indicates a weak relationship, a three indicates a moderate relationship and a nine indicates a strong relationship. The cell is left blank if there is no relationship. The ease of opening the door, for example, is strongly related to the weight of the door and moderately related to the stiffness of the door hinge, but is not related to the tightness of the door seal or window seal.

5. The team multiplies the customer importance rating of a feature by its relationship to an engineering attribute (one, three, or nine). These numbers are then summed for each column, yielding a total for the relative importance of each engineering attribute. For example, the stiffness of the hinge influences how easy the door is to open, and whether the door stays open on a hill. Thus to calculate the relative importance of the stiffness of the hinge, the team multiplies the customer importance rating of how easy the door is to open by its relationship to the stiffness of the hinge (15 × 3 = 45), then multiplies the customer importance rating of the door staying open on a hill by its relationship to the stiffness of the hinge (10 × 9 = 90), and then adds these together for the total relative importance of the hinge stiffness (45 + 90 = 135). These scores indicate that the tightness of the door and window seals is the most important engineering attribute, followed by the weight of the door.

6. The team evaluates the competition. A scale of one to seven is used (one indicating a requirement is not addressed, and seven indicating a requirement is completely satisfied) to evaluate the competing products (in this case A and B) on each of the customer requirements. These scores go in the right-hand “room” of the house of quality.

7. Using the relative importance ratings established for each engineering attribute and the scores for competing products (from step 7), the team determines target values for each of the design requirements (for example, the door's optimal weight in pounds).

8. A product design is then created based on the design targets from step 8. The team then evaluates the new design that was created. The team assesses the degree to which each of the customer requirements has been met, entering a one to seven in the far right column of the house of quality, permitting it to compare the new design with the scores of the competing products.

The great strength of the house of quality is that it provides a common language and framework within which the members of a project team may interact. The house of quality makes the relationship between product attributes and customer requirements very clear, it focuses on design trade-offs, it highlights the competitive shortcomings of the company's existing products, and it helps identify what steps need to be taken to improve them. The house of quality is used in settings as diverse as manufacturing, construction, police service, and educational curriculum design.30 Advocates of QFD maintain that one of its most valuable characteristics is its positive effect upon cross-functional communication and, through that, upon cycle time and the product/customer fit.31

page 255

Design for Manufacturing

Another method of facilitating integration between engineering and manufacturing, and of bringing issues of manufacturability into the design process as early as possible, is the use of design for manufacturing methods (DFM). Like QFD, DFM is simply a way of structuring the new product development process. Often this involves articulating a series of design rules. Figure 11.6 summarizes a set of commonly used design rules, along with their expected impact on performance.

FIGURE 11.6 Design Rules for Fabricated Assembly Products

Source: Adapted from M. A. Schilling and C. W. L. Hill, 1998, “Managing the New Product Development Process,” Academy of Management Executive, vol. 12, no. 3, pp. 67–81.

As shown in Figure 11.6, the purpose of such design rules is typically to reduce costs and boost product quality by ensuring that product designs are easy to manufacture. The easier products are to manufacture, the fewer the assembly steps required, the higher labor productivity will be, resulting in lower unit costs. DEKA Research makes a point of bringing manufacturing into the design process early, because as founder Dean Kamen points out, “It doesn't make sense to invent things that ultimately are made of unobtanium or expensium.”32 In addition, designing products to be easy to manufacture decreases the likelihood of making mistakes in the assembly process, resulting in higher product quality.

The benefits of adopting DFM rules can be dramatic. Considering manufacturing at an early stage of the design process can shorten development cycle time. In addition, by lowering costs and increasing product quality, DFM can increase the product's fit with customer requirements. For example, when NCR used DFM techniques to redesign one of its electronic cash registers, it reduced assembly time by 75 percent, reduced the parts required by 85 percent, utilized 65 percent fewer suppliers, and reduced direct labor time by 75 percent.33

Failure Modes and Effects Analysis

Failure modes and effects analysis (FMEA) is a method by which firms identify potential failures in a system, classify them according to their severity, and put a plan into place to prevent the failures from happening.34 First, potential failure modes are identified. For example, a firm developing a commercial aircraft might consider failure modes such as “landing gear does not descend,” or “communication system experiences interference”; a firm developing a new line of luxury hotels might consider failure modes such as “a reservation cannot be found” or “guest experiences poor service by room service staff.” Potential failure modes are then evaluated on three criteria of the risk they pose: severity, likelihood of occurrence, and inability of controls to detect it. Each criterion is given a score (e.g., one for lowest risk, five for highest risk), and then a composite risk priority number is created for each failure mode by multiplying its scores together (i.e., risk priority number = severity × likelihood of occurrence × inability of controls to detect). The firm can then prioritize its development efforts to target potential failure modes that pose the most composite risk. This means that rather than focus first on the failure modes that have the highest scores for severity of risk, the firm might find that it should focus first on failure modes that have less severe impacts, but occur more often and are less detectable.

FMEA was originally introduced in the 1940s by the U.S. Armed Forces and was initially adopted primarily for development projects in which the risks posed by failure were potentially very severe. For example, FMEA was widely used in the Apollo Space Program in its mission to put a man on the moon, and was adopted by Ford after its extremely costly experience with its Pinto model (the location of the gas tank in the Pinto made it exceptionally vulnerable to collisions, leading to fire-related deaths; Ford was forced to recall the Pintos to modify the fuel tanks, and was forced to pay out record-breaking sums in lawsuits that resulted from accidents).35 Soon, however, FMEA was adopted by firms in a wide range of industries, including many types of manufacturing industries, service industries, and health care. A recent PDMA study found that firms report using FMEA in 40 percent of the projects they undertake.36

Computer-Aided Design Computer-Aided Engineering/Computer-Aided Manufacturing

Computer-aided design (CAD) and computer-aided engineering (CAE) is the use of computers to build and test product designs. Rapid advances in computer technology have enabled the development of low-priced and high-powered graphics-based workstations. With these workstations, it is now possible to achieve what could previously be done only on a supercomputer: construct a three-dimensional “working” image of a product or subassembly. CAD enables the creation of a three-dimensional model; CAE makes it possible to virtually test the characteristics (e.g., strength, fatigue, and reliability) of this model. The combination enables product prototypes to be developed and tested in virtual reality. Engineers can quickly adjust prototype attributes by manipulating the three-dimensional model, allowing them to compare the characteristics of different product designs. Eliminating the need to build physical prototypes can reduce cycle time and lower costs as illustrated in the accompanying Theory in Action. Visualization tools and 3-D software are even being used to allow nonengineering customers to see and make minor alterations to the design and materials.

Computer-aided manufacturing (CAM) is the implementation of machine-controlled processes in manufacturing. CAM is faster and more flexible than traditional manufacturing. 37 Computers can automate the change between different product variations and allow for more variety and customization in the manufacturing process.

A recent incarnation of computer-aided manufacturing is three-dimensional printing (also known as additive manufacturing), whereby a design developed in a computer aided design program is literally printed by laying down thin horizontal cross sections of material until the model is complete. Unlike traditional methods of constructing a model, which typically involve machining a mold that can take several days to complete, three-dimensional printing can generate a model in a few hours. By 2015, three-dimensional printing was being used to create products as diverse as food, clothing, jewelry, solid-state batteries, and even titanium landing gear brackets for supersonic jets.38 Biotechnology firms were even using three-dimensional printing for use in creating organs by depositing layers of living cells onto a gel medium.39 This method has recently begun rapidly replacing injection molding for products that are produced in relatively small quantities.

three-dimensional printing A method whereby a design developed in a computer aided design program is printed in three dimensions by laying down thin strips of material until the model is complete.

page 257

Theory in Action  Computer-Aided Design of an America's Cup Yacht

Team New Zealand discovered the advantages of using sophisticated computer-aided-design techniques in designing the team's 1995 America's Cup yacht. The team had traditionally relied on developing smaller-scale prototypes of the yacht and testing the models in a water tank. However, such prototypes took months to fabricate and test and cost about $50,000 per prototype. This greatly limited the number of design options the team could consider. However, by using computer-aided-design technologies, the team could consider many more design specifications more quickly and inexpensively. Once the basic design is programmed, variations on that design can be run in a matter of hours, at little cost, enabling more insight into design trade-offs. Computer-aided design also avoided some of the problems inherent in scaling up prototypes (some features of the scaled-down prototype boats would affect the flow of water differently from full-scale boats, resulting in inaccurate results in prototype testing). The team would still build prototypes, but only after considering a much wider range of design alternatives using computer-aided-design methods. As noted by design team member Dave Egan, “Instead of relying on a few big leaps, we had the ability to continually design, test, and refine our ideas. The team would often hold informal discussions on design issues, sketch some schematics on the back of a beer mat, and ask me to run the numbers. Using traditional design methods would have meant waiting months for results, and by that time, our thinking would have evolved so much that the reason for the experiment would long since have been forgotten.”

 Source: M. Iansiti and A. MacCormack, “Team New Zealand,” Harvard Business School case no. 9-697-040, 1997.