The last step in concept development is to select a few integrated concepts for further development. One of the primary aims of this discussion is to develop more than one concept for consideration. A decision with only one option is an imperative—there is nothing to decide! The role of the architect in the concept generation phase is often to preserve options, in the face of pressure to down-select early and focus resources on a single concept. The challenge of evaluating concepts is that they may not be detailed enough to understand the potential fit against goals—hence the pressure to down-select early.

Bazerman provides an excellent overview of biases in managerial decision making, [14] which are directly relevant to concept selection. For example, concepts that are easier to recall, concepts that confirm existing beliefs, concepts that contain more information or fidelity, and concepts we have anchored on or are familiar with are all subject to potential overrepresentation in our choices. For these reasons, we will attempt to be explicit in our judgments. We will illustrate a method for screening concepts for the purpose of our Hybrid Car, and then in Part 4 we will discuss a variety of other, more sophisticated methods.

We would like to apply two central screening criteria: a backwards-looking consideration of the prioritized goals and a forward-looking consideration of the potential for good architecture. Recalling that our list of goals was divided among critically important, important, and desirable goals, we could rate each of the concepts using a Pugh matrix (Table 12.1). A Pugh matrix [15] is a concept comparison tool, where each concept is rated according to a number of criteria. Here we use the goals developed out of the stakeholder analysis (Chapter 11) as the criteria. The intent of the Pugh matrix is to provide a simple scale, often showing just three choices: advantages, average, and disadvantages. Here we have chosen a five-level scale running from very advantageous (++) to having many disadvantages (− −).

Table 12.1 shows a qualitative comparison of the vehicle architecture types considered in this chapter, taking a full hybrid vehicle as a reference. Although the baseline Pugh matrix weights all criteria equally, we could also weight them according to their importance (Chapter  11). Overall, the benefits of electrification lie in reduced tank-to-wheel emissions, better fuel consumption, an enhanced ecological image, and support from government incentives that make these new electrified powertrain architectures more attractive. The disadvantages of electrified powertrains lie in increased weight, reduced driving range, higher manufacturing costs, and commercial risks related to the replacement or servicing of the high-voltage battery. These disadvantages are the technological risks that first movers in the new-vehicle market must be ready to improve on to gain and secure an advantage in the market. If the first down-selection was conducted qualitatively with Pugh analysis, we might also apply a quantitative comparison across the good concepts, as discussed in Part 4.

Notice that not all goals differentiate between concepts (a backwards consideration). For example, we would have to move to a level of much greater detail to determine which of the concepts were more likely to engage suppliers in stable, long-term relationships. Therefore, this goal appears to be related more to the development process than to the concept decision. The concept decision will have to be made according to the subset of goals that offer differentiation.

When considering the concept, one should explicitly consider whether it will lead to an elegant architecture (a forward consideration). This is arguably the function of the entire text, but some simple criteria should come easily to mind, such as subjective criteria for whether the solution is simple and pleasing. Further, we could examine whether the functions are mapped to form in a one-to-one mapping, as in Suh’s axiomatic design, or the mapping is more complex. This may be difficult to determine until some architectural definition is done, so iteration is required. We could consider other dimensions of decomposition, such as whether form maps to development teams or manufacturing stations. A subjective down-selection, according to criteria like elegance, is often the final step in choosing two or three concepts for further development.

In summary, the exit from the concept development procedure should identify a small number of concepts for further development, based on a screening against important architecturally discriminating descriptive goals, and on a consideration of how the concepts lead to architectural elegance.