Optimizing product architecture is the highest leverage activity in prod- uct development and has the greatest potential for ensuring success. But, as with product definition, the importance of this stage is often ignored by merely assuming that the product architecture will be the same as previ- ous or competitive products.
One of the biggest causes of suboptimal product architecture starts with the seemingly innocuous step of building a breadboard “just to see if it works.” Breadboards are designed to prove functional feasibility and are usually built with the materials on hand (not widely available production- grade materials) in the most expedient way (not the most manufacturable way). Further, breadboards are built by prototype technicians, who can usually make a single unmanufacturable unit work as a matter of pride.
Product architecture optimization and manufacturability are rarely even considered at the breadboard phase, based on the assumption that those tasks will be done “later.”
Unfortunately, once the breadboard “works” and is demonstrated to management or customers—you guessed it—there is a strong temptation to “draw it up and get it into production.” The unfortunate result is the company ends up mass- producing breadboards forever! Basing produc- tion designs on breadboard architecture misses the biggest opportunities to make significant reductions in cost and development time.
As shown in Chapter 1, 60% of a product’s lifetime cumulative cost is determined by the concept/ architecture phase of a project (Figure 1.1). By the time design is completed, 80% of the lifetime cumulative cost is deter- mined. By the time the product reaches production, only 5% of the total cost can be influenced. This is why cost reduction efforts can be so futile, because cost is really determined by the design itself and is very difficult to remove later.
Similarly, other important design goals, such as quality, reliability, ser- viceability, flexibility, customizability, and regulatory compliance are most easily achieved by optimizing the product’s architecture.
Thorough up- front work greatly shortens the real time- to- market and avoids wasting time and resources on firefighting, change orders, and ramp problems, as shown in Figure 3.1.
The time to stable production can be cut in half because of the thorough up- front work, which minimizes the need for firefighting and change orders and makes the manufacturing ramp- up several times faster. Note that the concept/ architecture phase went from a trivial amount in the traditional model to an order of magnitude more in the concurrent engi- neering model. More thorough up- front work decreases the post- design activities from almost three- fourths to less than one- half of the product development cycle. It is more efficient to incorporate a balance of design considerations early than to implement them later as changes.
Figure 3.1 emphasizes one of the most important principles to reduce the real time- to- market: thorough up- front work. This graphic, and its profound implications, generate much discussion in the author’s in- house DFM trainings.2 In fact, at one Fortune 50 company, we spent an hour discussing this graphic at all four seminars.
While developing the Prius, the Toyota team avoided the temptation to jump right into detailed design:
“With the extreme time pressure, the temptation would be to make a very fast decision on the hybrid technology and get to work on it immediately.
Instead, the team reexamined all its options with painstaking thorough- ness, … considering 80 hybrid types and systematically … narrowing it to 10 types. The team carefully considered the merits of each of these and then selected the best four. Each of these four types was then evaluated carefully through computer simulation. Based on these results, they were confident enough to propose one alternative, … six months later.”3
Concurrent Engineering Time Line Concept
Architecture Design
Firefighting & Change Orders Ramp
Stable Production in Half the Time Typical Time Line
FIGURE 3.1
Traditional vs. front- loaded time lines.
As engineers and managers realize the importance of thorough archi- tecture optimization, they ask what more should be done in the order- of- magnitude increase in the concept/ architecture phase shown in Figure 3.1 and how this can actually reduce the final timeline so much. The key ele- ments of an optimal architecture phase are described here.
Product definition defines what the customer really wants and mini- mizes the chance that the product will be subject to change orders to reflect new customer needs that were not anticipated in the beginning.
Lessons learned should be thoroughly investigated and understood to learn what worked well and what caused problems in previous projects (see Section 3.3).
Issues should be raised and resolved before proceeding further, thus minimizing the chances that these issues will have to be resolved later when each change is harder to implement and when each change may, in turn, induce yet more changes.
Concepts should be simplified with clever, elegant designs, fewer parts, part combinations, higher levels of silicon integration, modular oppor- tunities, and so forth. Design efforts with ambitious cost goals may need help from design studies to generate breakthrough ideas.4
The architecture should be optimized for the minimum total cost, for designed- in quality and reliability, for manufacturability and serviceabil- ity, and for flexibility and customizability.
Don’t compromise the up- front work by rushing into design for the illusion of “early” progress because up- front work is unfamiliar, or because of temptations or pressures to launch the first working prototype into production.
3.2.1 Thorough Up- Front Work at Toyota
The 2006 book, The Toyota Product Development System, emphasized the importance of thorough up- front work at a company that some say is four times more efficient than the typical product development process.
“The ability to influence the success of a product development program is never greater than at the start of a project. The further into the pro- cess, the greater the constraints on decision making. As the program pro- gresses, the design space fills, investments are made, and changing course becomes increasingly more expensive, time consuming, and detrimental to product integrity.”5
The book provides many details about “bringing together your bright- est, most experienced engineers from all functional disciplines to work col- laboratively, thoroughly thinking through all of the critical project details, anticipating problems, applying lessons learned, creating precise plans, and designing countermeasures from a total systems perspective… .”
3.2.2 Thorough Up- Front Work at Motorola
Motorola’s most effective product development projects “invested rela- tively large amounts of effort early in the initial design phase so that most, if not all, of the problems that appeared later in the implementation phases had already been considered.”6
3.2.3 Thorough Up- Front Work at IDEO
Contrary to the apparent paradox, IDEO firmly believes that if they build many models early, it will “slow us down to speed us up. By taking the time to [model] our ideas, we avoid costly mistakes such as becoming too complex too early and sticking with a weak idea for too long.”7
3.2.4 Avoid Compromising Up- Front Work
Make sure the following problems do not compromise thorough up- front work.
3.2.4.1 Slow Processes for Sales and Contracts
Make sure that a slow sales and contract process does not delay the start of product development, especially for fixed delivery deadlines. Encourage customers to decide quickly and order early. Streamline the sales and con- tract process. Avoid contractual delays because of cumbersome or onerous terms and conditions. Plan ahead to start NPD as early as possible, even before the formal signing of a likely contract or award.
3.2.4.2 Rushing NPD for Long- Lead- Time Parts
When designs are based on parts with long lead times, the product’s up- front work may be shortened to make time for their delivery before
launch. The solution would be to eliminate long- lead- time parts with the techniques described in Section 5.19.2.
3.2.4.3 Rushing NPD for Early Evaluation Units
Similarly, pressures for early evaluation units could shorten the early up- front work. See solutions in the next section.
3.2.5 Early Evaluation Units
Early evaluation units are sometimes needed for qualifications, alpha tests, beta evaluations, trade shows, contractual obligations, or test markets, but hopefully not to get late customer feedback.
First, do all the thorough up- front work for the good of the overall proj- ect, but add the goal that some form of evaluation units can be delivered early. Make sure there are enough resources available.
Avoid the temptation to compromise the up- front work just to get units out early. Be careful that the structure of the early units does not specify or imply suboptimal product architecture of the production units.
Prove the feasibility of the functionality with a physical evaluation unit without necessarily achieving all the space, shape, interface, or weight constraints while, for contracts, presenting plans to achieve these goals in the production units
Maximize the proportion of time spent on the concept/ architecture phase by accelerating early the unit design and building process with para- metric CAD, modular design, rapid prototyping, additive manufacturing,8 short- run vendors/ partners, expedient tooling, or extensive “hogging” on CNC machines that would be replaced with more cost- effective processes in high- volume production. Two especially effective techniques are
• The SWAT team. Use skilled technicians to build, test, and validate evaluation units quickly, making sure they feed back lessons learned promptly to the design team.
• The parts store. If part delivery time for the evaluation units is so long that it threatens to shorten thorough up- front work, order all the possible part candidates before the BOM finalizes. These may appear to add cost or may not have an obvious funding source. However, getting these units out earlier will generate much more money in
increased profits. Actually, many of the parts in the “store” might be obtained at no cost as samples, which suppliers would gladly provide for free just to have their parts considered.
Keep track of the total cost and resource demands and make sure these provide the basis for planning and estimating, bidding, and negotiations on subsequent projects.