Few manufacturers have the luxury of substantially redesigning a product after it is in production. The chance to engage in redesign, for whatever reason it occurs, must be fully exploited as the great opportunity it really is. In the case of Alderberry's products, the need to produce approximately the same quality, at a higher volume and lower price provided the impetus. There was little choice: redesign the product or abandon it. There are always more goals that can be achieved than merely making something cheaper, or easier to build faster. These goals are mere factors in the larger problem of making a product more salable. Designing for Assembly (DFA) may help you sell your products at a low price, but it does not insure that you will sell enough of them to pay for the trouble you spent developing the methods that produce them.

To ensure that a product design fulfills the demands of the marketplace, goals must be established which reflect "absolutes" established by production and materials, and "needs" established by marketing and financial capabilities. These themes recur throughout the text. This focus is the result of years of personal experience "discovering" apparent voids in the marketplace, assuming these voids represented "needs", designing things to address these "needs", spending substantial amounts of money figuring out how to produce them, and then, finally, trying to figure out how to market them.

This manual began its life as a crash-course in "Concurrent Design" or "Simultaneous Engineering", the method of assessing all factors involved in the design of a product simultaneously . It grew other features and attributes as a result of an ever-changing design specification.

In the early to mid-1980's I explored patenting several ideas and applications of these ideas, in preparation for moving into large-scale production to exploit what appeared to be obvious market niches. I pursued developing musical notation software to allow written music to be generated by playing, and easily edited or taught by computer. I also developed, or put together a team to develop: 1) a flashlight with variable brightness, and 2) software which would auto-correct between the colors you see displayed on the computer screen and colors you see printed on paper.

Central to this activity was the ever-present belief in my own "lack of originality", my belief that my discovery of any void would soon be followed by similar discoveries of the same void by others, and that the first to fill the void would have the opportunity to create a "niche" and sell products to those who discovered the need next..

Based on nearly 20 years of this experience, I am not alone in my conviction that this pattern of product development leads to failure far more often than it leads to success. Standing in Costco or Target, it can be quite bewildering to be confronted by the enormity of the consumer market, and the incredibly narrow range of options actually available to the consumer. Few (if any) of the available products appear to be well designed or even well made. Why do people buy plastic yard furniture? Why do they buy all these things that use disposable batteries? Why don't they want to buy what I want to make?

A formal design process

Among the founding fathers of Industrial Design in the United States, (Henry Dreyfuss, Walter Dorwin Teague, Raymond Loewy and Norman Bell Geddes), Bell Geddes stands out as the man whose design process included assessing the needs of the marketplace. He was the first to set out an explicit design method and establish a design studio where this method was applied. The method he applied was this:

  1. Determine the precise performance requirements for the product.
  2. Study the methods and equipment used in the client's factory.
  3. Keep the design program within the budget.
  4. Consult experts on the use of materials.
  5. Study the competition.
  6. Conduct consumer-use surveys on existing products in the field.

Assessing customer needs

The most important opportunity available in a redesign of an existing product is the chance to create a better fit with the needs of the consumer. The craft of determining and of manipulating the perceived need of the customer is highly developed. However, the cost of addressing an existing need is far lower than the cost of creating the perception of need.

In the US alone, the advertising industry consumes close to $36 billion a year, and much of this money is spent solely to create the perception of need for products that fail to serve any real purpose or need in the customer's life or work, except insofar as they fulfill these created needs.

As a result of this incredible expenditure of money, the "science" of identifying potential customers, and assessing their perceptions and needs is well developed, and there are many consultants who can provide this service. There are also comprehensive manuals available on running surveys and assessing the results. The process presented below reflects, in summary form, a "state of the art" process that can be implemented without enormous investment, and can be run without consultants.

Whether you use this system or another is of little importance. The critical necessity is to ensure that you are really querying the opinion of the person you interview, rather than "creating" your answers by the way you design the questions or record the responses.

The system of collecting and processing information provides a means to acquire, organize and analyze information that is "broadly" useful. This means that the utility of this process extends beyond assessing customer needs. You could gather a great deal of important information about your employees' perceptions of the workings of your own organization using this process. Following the rationale provided previously I assume that you will assemble a "Design Team" using the talent already available in your organization, and that you will assign research and review tasks on the basis of interest, rather than on the basis of the existing organizational structures. The goal is to assess issues related to marketability.

In order to define the key aspects of the new product that insure marketability, the team must identify the new product's target customers and their needs, and it must develop a method to assess the degree to which the new product addresses these needs.

How to do it

You must actually gather customer data.

  1. Develop criteria to identify and qualify customers.
  2. Select qualified customers to interview.
  3. On the basis of the qualified customer list, identify nexus between common characteristics and interests of this set of individuals (refinement of the original customer qualifying criteria developed in steps 1 and 2 above) and characteristics of your product, and on the basis of these intersections, develop a script.
  4. Observe customers using existing products.
  5. On the basis of what you see, revise the questions & order in which they are presented in the script.
  6. Hold one-on-one interviews and ask appropriate questions*. If possible, have prospective customers use and review a similar product during their interview.
  7. Record dialog, and if possible, physical information as well. Record with multiple channels, to encode the maximum available bandwidth: If possible use a tape recorder, a cam-corder, and a note-pad.

Develop / extract data

Relative Importance Translation of Customer's Need
Category of concern
1Stated Requirement
2Stated Requirement
3Stated Requirement
4Stated Requirement
5Stated Requirement
6Stated Requirement
7Stated Requirement
8Stated Requirement

Organize data into a hierarchical customer needs matrix

WHICH PRODUCT PARAMETER Perceived ImportanceTranslated concern

Establish the basic metric for assessing customers' needs

Develop a product specification

The next stage: calibration & validation

It makes little sense to rush into the development of concepts until the specification is complete. Once the specification is working, it should be calibrated against products already on the market. It should also be used internally, by Production and Marketing staff, to document any internal concerns raised about the priorities and the metrics chosen to represent customers' needs, and to validate the numerical values selected. This is a very important step basic to "concurrent design" and it may identify substantial problems and allow them to be eliminated before they are "designed in". Marketing people may not be sufficiently familiar with realities of production to distinguish enormous cost differences between apparently similar materials or manufacturing methods. Bell Geddes referred questions of materials to experts in the field. Consider your production people experts-- they probably are. Once this has been done, it is time to move forward toward a product design.

Internal review and revised specification

As discussed in the previous chapter, metrics provide quantifiable means of translating customers' expressed requirements into specifications. Specifications are essential tools for choosing among alternative concepts or proposals. They allow evaluation of the success of each proposed concept in meeting the requirements of the market. Specifications must be expressed in terms of measurable values.

The same metrics can serve a variety of purposes. In these examples, different members of the design team use the same metric table to evaluate the concept from different directions, assigning different values and "importance rankings" for each metric. The customer will likely assign one level of importance to a particular detail, your production people another. A good example is the "green factor". Virtually everyone in the United States believes themselves to be an environmentalist. This means that they enjoy clean and beautiful surroundings, and believe that if given the opportunity, they do their part to keep it that way. On the other hand, the amount of "extra" money that consumers are actually willing to pay to ensure that the materials and processes used by your company are "green" appears to be quite limited. Therefore, customers' "importance" rankings must be listened to, but your actions must be tempered by your own experience and observation of actual market behavior as well as the results of interviews and surveys.

Production looks at the needs list, and the metrics developed to assess compliance with those needs, but assigns a value based on cost to achieve that specification using available technology, and ranks them according to their vision of "most bang for the buck".

Finance looks at the same metrics and assigns a different value, based on concern over the cost of the money required to achieve them, or perhaps on the potential to leverage the purchase of new manufacturing technology.

Representation: focus on the functions your product provides for the customer

Think back to Henry Ford and try to reverse engineer the design process he used from the lore that surrounds him: The line about any color you want, as long as its black is classic; his Model-T was a remarkably integrated design; nearly all the decisions that were reflected in it, from sticker-price to color, addressed well defined aspects of both the manufacturing capabilities of the period and marketplace into which the product was being launched. Cost was defined as a percentage of the income of the consumer. Options were evaluated in terms of their impact on both the cost to produce and the cost to maintain. His accountant was outraged that the cost to manufacture and sell the first cars was far above the sale price. Ford reasoned that the service / repair parts market (now called "consumables") that was created by these cars would quickly eclipse the gap in initial pricing.

Developing a design specification from customer needs

Developing a design concept from a design specification

Product development models: inductive and deductive

The problem of fitting solutions to problems is fundamental to our lives. It is so intuitive that we take it for granted. The processes we use for solving design problems can be roughly broken into two classes: fitting solutions and crafting solutions

The deductive process

Fitting solutions = Deductive reasoning = We develop solutions by identifying needs, and then applying solutions acquired through study of previous designs. This sort of work is the basis of most research and most design work. Solutions to most common problems are pulled from handbooks and catalogues. Sometimes however, this process does not produce the answers we need. Sometimes the solutions are not available in the handbook, regardless of how clearly the consumers' need are defined.

Options available
Relative Importance 12 34 5
Customer's Need < 6# / cubic foot > 5 years continuous duty > 2000 AH / per cell Can withstand 3000 g impact Can survive typical 70 mph crash.
Light-weight Lithium Battery
Long service life Lead Acid
High "burst" charge rate Electrolytic Capacitor
Mechanically Durable Polyethylene - Air
Non-explosive N/A

The inductive process

Crafting solutions

Inductive reasoning = We "craft" solutions by identifying needs, and then by identifying and filling gaps in our current knowledge. Such solutions are rare, and tend to be expensive. Proponents of personal electric-powered commuter vehicles have been waiting for a technological break-through in battery design (lightweight, high charge rate, and non-explosive) for nearly 20 years. No amount of reprioritization can overcome fundamental obstacles expressed in these tables.

Objectives tree

Functional analysis = requires an organized list of all expected inputs and all desired outputs.

Functional decomposition

Functional Decomposition (Divide and Conquer) is a standard engineering practice based on building a relatively short list (6-8 is ideal, less than 10 max.) of the most critical functions that define the concept. Functions are based on requirements defined through market research. Functions are isolated from the concept as a whole (decomposition), and then the full range of solution options available for each function is analyzed. The assumption of this process is that in isolation, each function can be decomposed to a scale where it becomes solvable, and that once the solutions have been identified, the overall concept can be reassembled.

In the real world, this linear approach often fails due to a phenomenon known as a combinatorial explosion. In practice, the range of solution possibilities for each isolated function will usually be large and the effect of combining these large sets of possible solutions is the creation of an enormous set of possibilities: a solution fog. Since these solutions cannot actually be assumed to fit together equally, selection of any particular combination of function-level solutions demands the evaluation of each against all the others, looking for positive and negative interactions. The optimal combination of solutions is the one with the best ratio of positive and negative interactions.

Prior to the development of CAE and CAD applications, such analysis was dismissed as impossible, or at least as not cost-effective. Now, desktop computer applications allow multiple solutions to many engineering and design problems to be evaluated and ranked rapidly.

Tools for representation

As stated previously, representation is the fundamental process of design. The end product of engineering is the artifact. The characteristics of the artifact are its attributes.

Representation is the process of characterizing the need for which the artifact is being designed and the process that ultimately yields the artifact as well as the attributes of the artifact.

Representation of the need and the processes are both critical factors in developing a design. In recent history, few events have called this necessity into such clarity as the collapse of the 4th floor mezzanine in the Hyatt Regency Hotel. This catastrophe was the direct result of inadequate representation of the underlying needs to which the design was directed. The use of the term "inadequate" is based on what appears to have happened, not on the laws that govern the assignment of responsibility or the distribution of responsibility assessed in the insurance settlements. This lack of representation allowed the contractor to improvise based on a misunderstanding of the engineering of the structure, and this led to the collapse of an entire floor of the building (as a result of loads which overwhelmed the supporting structures which were never intended to handle such loads). The engineer's plans were actually clear, showing a solution that would have worked (the engineering was sound). Had the building been constructed according to plan, it would not have collapsed. However, the design called for supporting fasteners that did not readily exist, and lacking clear instructions to the contrary, the contractor "improvised" and substituted available fasteners, hanging the structure from a floor above, pulling both down together.

Black box representation

During the Initial stages of design development and clarification, BLACK BOX or "schematic" representations are used to abstract the functional elements of the design solution. In the course of this abstraction, detail is kept deliberately spare, to allow the overall "architecture" of the concept to remain paramount. As the design process progresses, the BLACK BOXES are replaced with more detailed representations, reflecting a more completely defined understanding of the needs the box must fulfill and the solution options that it offers.

Detail within the BLACK BOX need never be developed if it represents an "off the shelf" component (i.e. a battery) or an element well understood in common practice (i.e. by most members of the design team).

One of the keys to developing efficiency among collaborators on a design team is the development of a well defined vocabulary of common practice, which allows discussion at an appropriate level of abstraction without the burden of unnecessarily detailing the obvious. The importance of establishing a solid basis for communication is not to be trivialized. Everyone comes by knowledge through different routes, and words acquire many meanings. Checking and rechecking for consensus on the meaning of terms among team members is essential.

Accurate specification of the required inputs and outputs of a BLACK BOX allows connection to other design elements, according to the architecture of the concept. This is another perfect job for Post-it Notes®. Use Post-it Notes® to represent the black boxes during the early stages of project definition. Add pages to the black box specification as its functions and requirements become more clear.

Transparent boxes

TRANSPARENT BOXES can be used alongside BLACK BOXES to abstract innovative or original practice.

Transparent boxes often contain schematics. Basically, schematics are abstract specification drawings showing the inter-relationships of one or more BLACK BOXES.

Schematic representation

Schematic representation is still paramount throughout the Design Process. The tendency to leap to a detailed representation of the artifact and its attributes, rather than exploring and documenting the relation between functions and objectives, shortchanges and potentially short-circuits the Design Process, and is to be avoided. The discipline of electronics is based on schematic representation. Since the performance characteristics of the components represented in the schematic are well known, simulations of overall circuit behavior are relatively easy to accomplish.

The evolution of analog electronic circuitry has progressed very rapidly from discrete components, which required complex schematic representations and analysis to integrated circuits which are now treated as complex multi-functional black boxes. In the electronics industry, these black boxes are characterized by their input and output specs and their performance.


As described in detail in the previous section, an overall design specification is developed from the abstractions called metrics. This is a process which enlists participants from a variety of disciplines who develop inclusive lists of necessary functions (performance criteria) which must be provided by the product, and matches these functions to listed components or structures required to provide these functions, selected to fulfill these criteria, and connected to provide all necessary functions.

Turning a design specification into a product

The design specification provides both metrics and values which represent the objectives of the design: these objectives reflect the customer's requirements. Key aspects of the specification are the order qualifying and order winning requirements determined through analysis of customer needs. These requirements include the price of competitive products. If a similar item is available in the Hold Everything Catalog for $69.00 plus $5.00 handling plus $12.00 shipping, a value for the metric "retail price" might be "under $87.00". From this number, wholesale price can be deduced. If you are selling into a distribution channel that demands 60 points of markup, the target cost to produce the finished piece, including necessary profit, is then around $35.00.

Problem solving

Problems can be classified along a range from easy to difficult, according to the constraints on their solution.

Solutions can be classified as routine, innovative or creative.

Solving design problems

Applying concurrent engineering

Concurrent engineering implies that the design of a product, and the systems to manufacture, service, and ultimately dispose of the product are considered from the initial design concept.

-------- J. A. Rehg

Experience in companies large and small has established that the three major aspects of manufacturing (marketing, product design, and production engineering) are best approached simultaneously. The cruel joke is that everyone knew this a long time ago, but somehow, we all forgot.

Design - a process of refinement

This information is developed "internally". It is based on the economics of the kind of business you are in: the scale at which you currently operate, and the "appetite" of the organization you will need to field the kind of product you are designing.

The market opportunities are evaluated based on the market's presumed appetite for the product (which drives production volume, and thus production capacity requirements), and the perceived value of the product (which drives ultimate cost, quality and margin considerations).

Requirements are the production and design evaluation criteria. Requirements allow design and production options (including technologies) to be evaluated (ranked) against criteria chosen to reflect their ability to satisfy market requirements (especially price) which result from design and production decisions. Assessment of market opportunities yields requirements for sales, expressed below as prerequisites. Each of these prerequisites offers an opportunity to balance the basic variables of price, quality, service, and function in different ways. Each combination which appears to satisfy the criteria characterizes a different concept.

Output 2 = constraints

These prerequisites provide constraints on options. These constraints become essential Design Criteria. Inputs come from examination of existing products, existing markets, emerging markets, and analysis of existing production capacity. Capacity analysis is based on:

Output 3 = metrics for evaluation of concept / choice of design technology

Production design

Production design is always a process of refinement. Between the finalization of the Concept and the implementation of the design are many iterative steps, which generally attempt to reduce cost without adversely affecting other prerequisites. These steps are generally discussed as Design for Assembly and Design for Manufacture in section [5].

Production design is the final translation of a thoroughly defined concept into an artifact.

A general design process

Clarification of objectives / development of constraints stage

Initial market definition (niche) established.

Cost and production target definition established.

Abstraction - description of the general problem

Abstraction - visualization of possible solutions

Synthesis stage

Using all the various ways we have developed to answer the question "How can we do that?" we must assemble the best solution. This is sometimes called design enrichment. This will involve combining elements from a variety of solutions in a design synthesis.

The synthesis stage normally involves approaching the overall problem as isolated problems which must be solved. This process, called functional decomposition, breaks the overall design into small, individually solvable problems (divide and conquer), based on the functions the product must provide to the customer and the functions each part must provide to the overall design.

This process leads to a set of solvable problems at the level of individual design elements, but also leads to a combinatorial explosion of possibilities, which tends to overwhelm the designers with options while providing little guidance for choices among these options. This can create as much difficulty as it eliminates.

Analysis stage

Development of metrics for the evaluation of alternatives and selections.

Analysis of differences between alternatives requires detailed itemization of cost and customer needs. Tools are based on financial analysis and metrics developed previously.

CAE and CAD software allow the computer to assist the designers by evaluating vast numbers of possible solutions, then testing to see which fit together in synergy and which interact in contradictory ways. In traditional engineering analysis, metrics are developed to allow mathematical analysis of the performance of individual design options. Evaluations of the interactions between design elements can also be performed, if metrics are developed to allow an evaluation. This process has led to breakthroughs in the fields of mechanical engineering of structures, simplification of optical systems through the development of aspheric lenses, consumer electronics, computers, and automated manufacture systems.

Evaluation stage

Analysis of these metrics allows development of a detailed design specification. At this stage, design is still largely abstract, and shape or material may not even be chosen, but the functions it must provide have been thoroughly defined.

Concurrent engineering follows 5 phases

Architecture of product is established. Initial value engineering (VEG) may be performed at this point. Value engineering is a process in which each aspect of the design (part and subassembly decision) is analyzed individually, in a manner similar to Design for Manufacture / Assembly methodologies (DFM / DFA). The focus is simpler in VEG. Here, the costs are evaluated and make / buy decisions are reassessed. Often, an outside consultant will be retained to manage the VEG process, since by this point in the design process individual "ownership" of some of the previous decisions can create impediments to objective analysis.

Assumes design specification, evaluation criteria, and the overall architecture of the product are complete. Detailed design of individual parts requires:

Integration of parts explored. Sketches begin translation into working drawings.

At this stage, tolerance criteria are developed to govern the fit of individual parts. This is another round of functional decomposition based on the products architecture (in which the relationship between functional elements is explicitly defined) and production tolerances for the physical relationship between adjacent and mating parts are defined.

At this stage, manufacturing cost and assembly time can be fairly accurately estimated. In an effort to reduce this cost, design for manufacture and assembly methodology (DFMA) is applied to each part and subassembly. Modularity and integration decisions are revisited.

DFMA at this stage can result in major changes in the overall "gestalt" of the product. Assure that previously developed criteria and design specifications remain intact throughout DFMA. Carefully consider the impact of all DFM / DFA decisions on serviceability, customer perception of value, metrics, as well as on manufacturing cost.

At the end of this stage, working drawings will be complete or the design specification will be sufficiently detailed such that outside vendors can be approached for outsource solutions.

The paths to new product development

The traditional product development model is proving to be too slow to get products into modern markets. Great progress has been made in "fast prototyping" methods which allow concepts to be evaluated by customers and even field tested long before production tooling is developed. This allows a much closer fit between customer preferences and the finished product. Machines utilizing technologies such as Stereolithography (SL) and Selective Laser Sintering (SLS) will continue to improve and proliferate, and this will make the development of plastic parts faster and cheaper. It is not clear what impact this will have on secondary wood products, except to shrink the cost differential that currently favors the use of machined wood over the use of molded plastic in limited production products.

The traditional product development model

The fast prototyping model

Design for manufacture and assembly (DFMA) process

DFMA is an iterative approach to addressing multiple interwoven design decisions simultaneously. Iterative means that you will perform the same analysis over and over. This process is applicable to prototypes and to products already in production. In the simplest sense, the goal of DFMA is to reduce the number of steps and the number of parts involved in the manufacture of an item.

DFMA has become codified as a set of processes. It is most often applied using a Computer Aided Engineering (CAE) package produced by Boothroyd Dewhurst, Inc. In a sense, DFMA asks: "Is the design ready to be put into production?" CAE provides analytical tools allowing the solution of interacting problems whose interactions render them recalcitrant to manual solutions.

DFMA began as two separate but related disciplines: DFM and DFA. These disciplines have related but sometimes antagonistic goals. DFM stressed the producibility of individual parts and sub assemblies and DFA stressed ease of assembly and overall integration of functions. In the most reductionist sense, DFM asks "How can we make this part cheaply and easily? Which manufacturing steps can we avoid having to do?" DFA asks "What steps can we combine? What parts can we eliminate?"

DFMA has developed into a rigorous discipline which is based on development of a theoretical benchmark for each part's total manufacturing cost and the effect of each possible configuration on assembly time.

The process normally begins by testing the proposed design for ease of assembly because it is assumed that the major areas in which production cost can be reduced are through reducing quality problems and simplifying assembly. The test is performed through the application of DFA rules. Violations are noted, since violations identify areas in need of redesign. Once the design passes the DFA tests, the proposed assembly methods can be evaluated by applying a score-card metaphor to narrow the range of options available.

A table is created showing the cost to implement and the benefits expected for each strategy. Costs are summed and each option is given a score from 0 -100. The total assembly is also given a score based on the sum of the individual scores. This aggregate score allows the comparison of the cost implications of a range of manufacturing options that would otherwise appear to be "apples and oranges" and allows the overall design to be "objectively rated".

The net effect of DFMA is the development of "integrated" products that contain fewer parts and are less expensive to assemble and manufacture. However, this often raises the development cost and the cost of tooling required. This process is what got us into our spiritual crisis in Costco and Target at the beginning of this section.


  • Prototypes.
  • Re-engineering Production.
  • Prerequisites

    Completed design specification.

    Defined Quality Assurance metrics.


    1. Calculate cost of manufacture.

    2. Develop cost reduction strategies.

    3. Test cost reduction strategies.

    4. Re-calculate cost of manufacture.

    5. Iterate process until cost reduction is acceptable.

    Sequence (the humble art of covering your tracks)

    One of the more important side-effects of the application of the Design Process (and Industrial Design in general) to commercial products has been the development of a rigid set of customer-expectations around fit and finish, and the promulgation of generally accepted procedures for achieving "commercial" finish quality on manufactured items.

    As a result of a number of factors entirely unrelated to quality, consumers have come to expect a very high level of "finish" on the products they buy. Most often these standards are related to the requirements of the processes and materials used in manufacture, rather than any particular decision. Fiberglass boats have very shiny surfaces because it is easier to get shiny boats out of the mold, not because they look better that way. However, this attribute of molded products has had a major impact on manufacturers working in other materials by raising consumer's expectations.

    Close examination of your competitor's products, looking for changes (particularly escalations) in the quality of fit and finish should be part of your normal activity. The bottom line usually comes down to how you deal with the "ragged edge". There is always some trivial detail that ends up jumping out like a kid's cow-lick, consuming far more time than it is worth.

    If you were to study the history of product design and list the trend-setting designs of the century, you would find that many of the most important design breakthroughs are related to the elimination of finish problems through planning the assembly sequence and through integrating the functions of a group of parts.

    My favorite example of covering your tracks is found in the attachment of the air intake screen for the cockpit air vent on the Ferrari 250-GT, a luxury 2-seater designed in the early 1960's by Pininfarina. Overall, this may have been the most widely copied automobile in history. Designers in almost every major company "stole" pieces of the car. For instance, the MGB's designers copied the tail-lights, the Chevy Vega copied the nose, and dozens of cars carry the roof-line. But these were merely copies of styling. The best details were the methods Farina's designers used to achieve the illusion of top-quality finish in a limited production machine that by the standard of the period was really very crudely built. The magic of the design is that there are simply no rough edges visible anywhere.

    As far as I can tell, little thought was given to maintenance. It was probably assumed that the car would end up splattered against a tree long before anything wore out.

    The trick used on the air vent screen was to use the nuts on the shafts that power the windshield wiper arms to hold it down. These nuts serve three functions: they hold the wiper arms onto the shafts, they hold the wiper mechanism below the shafts in place, and they hold down the stainless steel escutcheons which cover the whole messy operation and hold down the air-screen which keeps leaves and pine-needles out of the sump. The shafts come up through the cowl in tubes which extend the bottom of the vent sump (which catches leaves and water) above the level of the hood just forward of the windshield. You will be hard pressed to find a better job of integrating diverse but related functions anywhere. This detail perfectly embodies the philosophy of Design For Manufacture and Assembly (DFMA) discussed above, and it carries a deeper lesson: when studying the work of a master, you should emulate his genius, not ape his styling.

    The "right price" (the perception of value)

    The preceding discussion undoubtedly seemed like a total digression, talking about astronomically expensive sports cars from the 1960s. But it was not.

    One of the lessons that we must all face eventually is a very hard one to learn. You must recognize who and what your competitors really are. Unless you are providing a necessity or are developing a platform product (one which assumes the prior purchase of another product), then your competitors are probably not making similar products to yours. Instead, they are making all the other kinds of products that are competing for your particular customer's discretionary money.

    From this reality comes the problem of the market's expected or "realistic" price. This number is based on the consumer's aggregate purchasing experience and is confused by many seemingly unrelated considerations, such as the price of used cars, 14.4 k internal modems, and take-out Chinese dinners.

    The basis of the standard

    When I was building 18 foot racing sailboats, my real competitors were often building sports cars, not sailboats. This was pointed out to me by a customer who ended up buying the sports car instead. Therefore, the standard of finish the customer expected was actually based on the MGB-GT (now the Miata), not on some other manufacturer's sailboat design.

    The customer applied this third standard as the basis of the evaluation of the boats available, rather than an objective comparison of the merits of two or three boat designs. He assumed that the "fun" and the cost of using the boat or the car were approximately equal, and because the automotive standard of fit and finish was virtually unattainable in limited production sailboats, it led to his conclusion that the "value" was not there in any of the boats to warrant purchase. The standard he chose to use was based on a design refined over years of production and an implementation based on tooling amortized over tens of thousands of vehicles.

    This recognition is echoed in The Business of Woodwork, when Norlin reminds us: "even if you are the only bidder, not bidding against a real number, you are bidding against the perception of what it should cost."

    Maintaining quality

    One of the relevant issues this raises is the problem of consistency that often appears when you attempt to export part or all of your production to subcontractors' shops. Your design reflects your own experience, the things you value in your surroundings and what you think works. How do you communicate this to someone else in a contract? How can you assure that subcontractors and even new employees understand what you are trying to do?

    Building in quality

    Initially, the key to building in quality is to develop truly excellent tooling and techniques. The process of developing quality is "scientific" in that it relies on experimentation and on careful documentation of all experiments, and cannot be expected to work without both. The techniques which work must be carefully and explicitly documented. Second, you must document why you do what you do, and if you want to avoid having to pay for the same failed experiments over and over, you must explain what you have already tried that did not work, to your employees, your vendors, and your subcontractors.

    Product specifications

    Fundamental to the development of a product specification is the development of a standardized vocabulary that can be used throughout the project in discussions with customers, employees, vendors and subcontractors. This vocabulary provides the basis for any contract terms. The clearest examples are always physical objects with the areas of concern (critical areas of fit and finish) well marked.

    Transportable tools

    The link between design and CAD and CNC tools has often been stretched pretty thin during the past 2 chapters. Perhaps it would have been better to continually tie the discussion back to CAD tools. However, the link is actually fundamental. Once you have developed a design specification and committed it to production tooling, CAD provides the means to keep track of design history, make relatively quick and easy revisions to drawings, and use CNC tools to cut out patterns.

    This final step, CNC patterns, allows the development of transportable tools. If you design them properly, these tools can be moved from your shop to someone else's and will allow them to make parts to your spec so that you can be assured parts will fit your customers' needs. This is a critical function if you are ever going to explore the possibility of subcontracting or distributed manufacture. You must be able to provide your customers, yourself and your subcontractors with the assurance that the parts built through application of your tools and methods will fit and will be salable.

    In order for this assurance to be lasting and credible, you must be able to establish that any deviation from tooling or procedures you have provided is at the contractor's financial peril. To this end, the notes on experiments your firm has already performed can provide a powerful incentive to avoid replication of past mistakes.

    In a sense, the development of hard tooling is an attempt to use a reduced number of words in more explicit ways to describe the processes we are using, as though we were following a road map. The drawback to following a map, as opposed to following one's intuition, is that it may limit where you can go. On the other hand, following the map substantially increases the odds that we will arrive at the destination and that we will get there on time.

    "It has often been said that a person does not truly understand something until he teaches it to someone else. Actually, a person doesn't really understand something until he can teach it to a computer, i.e. express it in an algorithm.... The attempt to formalize things as algorithms leads to a much deeper understanding than if we simply try to understand things in the traditional way"..

    Donald Knuth, 1973