Modularity in New Market Formation

Date May 30, 2016
Speaker Ronald Allen SANCHEZ (Professor of Management, Department of Innovation and Organizational Economics, Copenhagen Business School)
Commentator SHIBATA Tomoatsu (Professor, Graduate School of Economics and Management / Faculty of Economics, Tohoku University)
Moderator HASEGAWA Hiroshi (Deputy Director for cross-sectoral policy coordination, Manufacturing Industries Policy Division, Manufacturing Industries Bureau, METI)


The modularity revolution, adoption, and challenges

Ronald Allen SANCHEZ's Photo

Ronald Allen SANCHEZ

Integral architecture is a term I have often heard in Japan. The view in Japan seems to be that integral architectures are best for achieving commercial success, but I have yet to find an adequate definition of what "integral architecture" actually means. If "integral architecture" means a design that is well-integrated esthetically and functionally, then an integral architecture could contribute to commercial success. However, if "integral architecture" refers to an optimized design intended to be good at doing only one thing, then I believe it would be virtually impossible to succeed in today's diverse and rapidly changing markets with "integral designs." The wide configurability and rapid upgradability of modular designs mean that product designs derived from modular platform architectures are the key to commercial success in almost every industry. Modular designs—not one-off optimized designs—are taking over every major industry worldwide. Therefore, strategies based on single-purpose designs following some unclear notion of "integrated architecture" could seriously limit the competitiveness of Japanese products in global markets, as well as in Japan.

Another idea that I have encountered in Japan is that modular architectures lead to product sameness and "commoditization" of product markets, with resulting low or no profitability. In fact, the opposite is true. Modular architectures make it possible to configure a range of product variations and technologically upgraded products quickly and at low incremental cost. The evidence around the world is clear: Modular architectures enable greater product variety, faster upgrading, and greater speed to market, and greatly reduced product costs.

This is why I refer to the "Modularity Revolution." Modular architectures represent a fundamental revolution in product design technology and are now clearly a dominant strategy in virtually every kind of product market in the world today. It is unclear to me why this is not recognized in Japan today, especially when some Japanese companies (like SONY Corporation) have been such competent and successful users of modular architecture strategies.

My experience in researching and consulting with companies around the world in implementing modular architectures in support of modular strategies has been that the real challenges in implementing modularity are not technical, but managerial and organizational. Managers need to understand modularity in order to understand the significant changes that must be made to an organization and its processes in order to implement modular design processes and market strategies—but most senior managers are too busy (or too proud or afraid) to take the time to learn about modularity and about the new rules and new roles an organization must put in place in order to implement successful modular product creation processes.

Modularity is a tool for supporting various kinds of market strategies, and the ways modularity is used will therefore vary with a company's specific strategies and business situation. Thus, one company may use modularity successfully in ways that are quite different from successful uses of modularity in another company. For example, the research I am doing with Professor Tomoatsu Shibata has described Volkswagen AG's modularity strategy as based on what we call "inside-out" modularization: a platform of common components is adopted that then constrains exterior body designs. We have described Nissan Motor Company Ltd.'s modular strategy, however, as "outside-in" modularization: major body components define an architecture that constrains the mechanical components used in its vehicles. The two different approaches appear to be well-aligned with each firm's rather different business situation. Thus, with modularity, "one size does not fit all," and modular architectures must be created to support each firm's own business situation and strategy.

Research has clearly established that, in many industries, the use of modularity concepts in both product and process designs enables new organizational capabilities and product strategies. When one or more company in an industry begins to adopt modular product architectures and processes, it profoundly transforms the structures and dynamics of the industries. Vertically integrated production systems are replaced by networks of component developers and producers and final product assemblers and marketers who cooperate "upstream" in defining and developing modular architectures, and then compete "downstream" to see which companies can use modular architectures in serving various market preferences. These dynamics of modular industries create "co-opetition" that leads to much more efficient, adaptable, and evolvable industries than vertically integrated industry structures have been able to achieve. Thus, it is no coincidence that the most technologically dynamic and diverse industries in world—computing devices, telecom products, consumer electronics—are all based on modular architectures, as well as on widespread use of industry-standard common components.

In this seminar, we will consider how government policies may support the development of modular industries. A policy principle that I will emphasize is that effective industrial policies will seek to encourage—but never command—the use of modular product architectures in industries. As a case example for discussion, we will briefly consider the role of open-system modular architectures in the rapid development of the electric two-wheeled vehicle (E2WV) industry in China.

Open and closed systems

Confusion also exists with regard to the meaning and significance of open vs. closed system designs. Open-system product and process architectures are designs that interested companies are legally and technically able to use. In contrast, in closed system architectures, some firms either are unable to understand an architecture technically or are not allowed to use the architecture because of patent protection or other legal restrictions. A computer with a USB interface for memory devices would be an open-system architecture in that respect, while SONY's proprietary "Memory Stick" interface would be an example of a closed-system architecture.

With global talent and technology capabilities on the rise around the world, the idea of one company being able to create a stand-alone architecture that can take on the world is becoming untenable. Even Apple Inc. is under pressure now to make its products more open-system, largely owing to the commercial success of Samsung's Android-based open-system devices.

I think the prior era of commercial success achieved with "stand-alone" proprietary architectures is historically interesting, and some Japanese companies did very well during this era. But commercial success today and in the future requires use of modular architectures, of co-opetition in networks of companies rather than vertically integrated industrial structures, and of open-system modular architectures to maximize configurability and upgradability of the products offered by an industry (while also lowering product costs, often substantially).

Nature of modularity

Our discussion today is structured around four key questions:

  1. What is modularity?
  2. How does modularity enable new firm capabilities and market strategies?
  3. How does modularity enable new market structures and industry dynamics?
  4. How can modularity become an important component of industrial policies and support new industries and technological development?

Let me just also clarify that although the title of this speech is "Modularity in New Market Formation," this discussion is also very relevant to the revitalization of established industries. The transformation of industries from proprietary, vertically-controlled "integral architectures" to open system modular architectures can lead to a substantial renewal of design creativity and new market development and growth.

What is modularity?
Modularity is a special kind of product and/or process architecture that is designed to enable rapid configuration of new product and process variations and rapid technological development of improved and higher performing product variations.

Therefore, in order to define modularity, we first have to explain the meaning of architectures. Architecture is a two-part concept that is a way of describing the technical structure of any design, whether it is hardware, software, or a service process. Architecture is, in effect, a way of describing what makes a system-design work technically. Taking product architecture as an example, we would start by asking what we want the product to do for customers. How are we trying to provide utility and value to the user? What will this product do that will make it valuable? We then define the specific functionalities that would make the product useful and attractive to customers. Then, we technically "translate" these product functionalities into sets of functional components. The first step in defining an architecture is therefore to define the set of functional components that will make up the product architecture.

The second part of the concept of an architecture is also essential. We next have to fully specify the component interfaces; namely, the inputs and outputs of each component that will determine how the components interact in the product design as a system.

We can apply the same two-part definition to a process design. In that case, we start with the question: What is the process supposed to do for internal and/or external customers? What functionalities does the process need for it to be useful to the intended customers? We then define the activity building blocks, or process components, needed to deliver those functionalities. We then have to fully define the interfaces between the activity building blocks in order to define how the processes interact when functioning as a system.

Architectures then become modular when the way that we decompose the architectures into sets of functional components is standardized. That is, we adopt a standard set of component types for delivering the product's functionalities to the customer (rather than making different combinations of component types or adding other functions and components as an afterthought). The first step in modularization is to define a set of functional components which is then standardized for a given architecture.

We next define the interfaces between the components in a very special way. We first define the interfaces for each component to allow the substitution of a range of component variations in the architecture without having to change the designs of any other components or any other aspect of the architecture. We then standardize those interfaces for the lifetime of the architecture. Not allowing changes to the current interfaces greatly expedites development processes. The substantial increase in development speed achievable when interfaces are defined and standardized as a first step in the development process enables a "fast-cycle" architecture development process. Any change in interfaces is a change architecture, and it has to be managed as a strategic change. In other words, we don't just make changes in interfaces in the engineering design process because a change in one part of a system design may have "ripple effects" that lead to technical problems in other parts of the system design. We therefore reserve any changes in component types and interfaces for a next generation architecture in which we may make a number of strategically-motivated changes to interfaces and/or components.

We can use modular architectures to create what I call a strategic platform. I define a strategic platform as a carefully and strategically coordinated product architecture, and supporting process architecture intended to accomplish specific strategic objectives. We then concurrently define and co-develop the product architecture and process architecture to accomplish the defined strategic objectives. The process architecture must have the flexibility to take advantage of the ability of the product architecture to generate product variety and upgrades.

How do we create strategic platforms to promote product variety and technological change? First, we follow a principle that I call "one-to-one mapping" in decomposing the overall functionalities that we want to have in a product into specific functional components. We strive to understand the functions, features, and performance levels needed by our intended customers and/or the market in general and work to create functional components that individually deliver just one of these important functions, features, or performance levels. In effect, we try to capture each aspect of the product that is important to the customer in a single component (or subsystem) in the technical structure of the architecture, because when we begin to introduce new component variations into the architecture, they will be seen as significant variations or changes in the product by the market.

We then strategically partition the architecture into sets of common components that do not need to vary across product variations (and which can therefore be massed produced at low cost), and sets of differentiating components that confer product variety in the eyes of the market.

We then succeed in product markets by adding differentiating components to common components to create modular product variations that are perceived by various kinds of customers as better because they are more suited to the preferences of different kinds of customers. To create variety that is seen by customers as significant, we need to map the functions very carefully and clearly into specific functional components. Volkswagen seems to be very good at this, and in its various platform architectures, functions (and associated components) that are not seen as important by customers are standardized and used in common among a platform's vehicles, while the functions and components that are perceived by the market as important are made changeable and varied across the product models configured from the platform architecture.

Components should also be technically decoupled so that within any given component type, other components can be freely substituted into the architecture without affecting the functioning of the component. This requires technical isolation of each functional component within a system design. During development, coordination of component development is necessary and is achieved through standardized interface specifications, but after development is completed we want to be able to freely plug and play an intended range of component variations in the product architecture.

I like to use a simple example to explain the "Power of Modularity" to enable product variety and change. Imagine a product that has 10 different types of functional components, each of which can have 10 different component variations, so that we have a set of 100 components that can "plug and play" in the architecture. We can then configure 10 billion (1010) product variations from this set of only 100 component variations. Thus, enormous amounts of variety can be leveraged from a well-chosen set of functional components, and if our supporting process architecture is flexible enough to handle the product variations, there's little or no additional cost to offering product variety other than the cost of developing new component variations.

As an example, consider the truck architectures made by two companies in Europe. Truck companies in Europe producing tractor trailers usually have to provide at least 40 different variations of engine size and transmissions to meet various use conditions, and, in addition, truck bodies are essentially customized to meet the preferences of different drivers for sleeper beds and various comfort features. One company's design solution was to create about 40 different designs requiring over 15,000 different parts. A smarter company produced over 60 different product variations based on 600 modular components. Think of the differences in product complexity and costs for Company A versus Company B, and also the differences in process complexity that have to be managed in the two companies. As you might expect, Company B has survived and prospered, while Company A has struggled.

Strategic partitioning can also be illustrated by examples of small products like clothes irons and powered toothbrushes. I want to emphasize, however, that the principles I will illustrate using these examples are directly applicable to automobiles, airplanes, moon rockets, software, banking services, and any other kinds of products. The principles of one-to-one mapping and strategic partitioning are generally-applicable design principles that can—and should—be used in creating any kind of architecture.

Philips has a personal garment care unit in Singapore that makes many millions of clothes irons a year for markets around the world. At any point in time, it provides between 300-400 product variations to markets around the world, and there is a constant need to refresh these designs every six months to create ever-changing variety in the markets. Philips has strategically partitioned the product architecture of its irons into common components used across all models (which are mass-produced in an automated factory), and differentiating components that are used to create hundreds of product variations, are changed a couple of times a year to "refresh" product lines, and are made in batch processes for assembly at the last point in the production process.

The result of this strategic partitioning is a broad range of product variations and low cost per product. Strategic partitioning of product and process architectures to enable late-point differentiation in production processes can significantly reduce product costs, increase speed to market, and reduce production complexity. In fact, studies show that the time and cost of product development can be reduced by as much as 80%. Product cost can be further reduced through disciplined re-use of existing components and by designing components for commonality and reuse. We can also improve the predictability of new product introductions because the discipline of the modular development process in eliminating technological uncertainty from development results in greater development productivity and speed.

We can also reduce customers' operating costs and complexity. When customers are involved in their products and can understand the continuity of product architectures, they may understand better how a firm intends to provide them with upgraded products in the future. This understanding may provide the basis for establishing long-term customer relationships.

Advantages of modular development over conventional development

In conventional product development, when a product is being developed, different components are likely to be at different stages of technological development. Thus, conventional product development processes typically try to develop new technology (in the form of new component designs) while also trying to create new product designs based on new components. The mixing of technology development and product development in the same process is not effective. When component designs and their interface specifications are evolving during a development process, frequent redesign of components is required. My research has shown that in a broad cross-section of industries, 60%-80% of conventional product development time is spent on redesign because the interfaces were not defined and standardized at the beginning. The other problem with conventional product development processes is that redesign takes a lot of time, often causing the development process to fall behind schedule. In an effort to stay on schedule, ad hoc "compromise" component designs and interface specifications are often adopted. As a result, the product architecture that emerges from conventional development processes is usually not flexible, configurable, or evolvable, but rather an inflexible "compromise" solution that only does one thing, often not very well.

In modular development, the essential first step is to fully define and freeze the component interfaces. Any technologically new component (whose interfaces cannot be reliably specified yet) should be developed offline in a decoupled development process. When you are finally certain that you understand the system behavior of a new component in the intended use context and are confident that reliable interface specifications can be defined for that component, then you put the new component design into the design library of available components for use the next generation of architecture.

In modular architecture development processes, we do not try to develop new types of components while also trying to configure a new architecture. Attempting concurrent development of components without first standardizing (freezing) the interfaces results not in "concurrent engineering," but rather in what I call "concurrent chaos." Some companies have tried to use what they call "overlapping development" or "overlapping problem-solving" methods to accelerate product development. This method seeks to compress development time by beginning development of some components midway in the development process of other components. The problem in this method is that as further component development takes place, component designs and interface specifications typically have to be changed, and much redesign work is still needed. The reworking of designs is not only time-consuming; it can also be a very contentious process. I estimate that overlapping problem-solving may achieve a 20%-25% reduction in development time—but only at great human cost of stress caused by disagreements and frequent redesigns of components. By comparison, my research shows that disciplined modular architecture development processes can reduce development time by as much as 80%.

In fast-cycle architecture development, higher performing upgraded products are brought to market not by trying to develop new technology while trying to develop a new product, but by developing new technology "offline" and then introducing new, proven component designs into next generation architectures. Philips has stated that when it converted to this fast-cycle architecture development process, it was able to bring new technologies and new products to market four times faster than when it was using conventional technology and product development processes.

Another example is Philips' electric toothbrush business. Philips was close to exiting that market; it was losing market share and was becoming overwhelmed by the complexity of their development and production processes, trying to design one-off designs for about 80 or 90 different product variations. By adopting a modular development process, Philips was able to increase product variations to more than 300, reduce delivered product cost by 48%, reduce lead times from six weeks to five days, and improve order fulfillment rates from 80% to 99%.

When producers can cooperate to define a common architecture, it encourages component suppliers to enter the market and create a supplier base because they now know what components are going to be used in the industry, and their potential customers multiply. It's a major incentive to establish a supplier base. This in turn has a similar effect on the subcomponent suppliers and investment and cooperation among them. Adopting a common platform throughout an industry, or at least platforms that share common components, lowers the risk of investing in starting a new industry. In addition, businesses and consumers increasingly do not like to be locked into proprietary systems, and appreciate the flexibility and cost savings of open system modular architectures.

Industrial policy to promote modularity

The first industrial policy focus in promoting cooperation among firms and architectural standards should be to find ways for firms to agree on a common architecture as the basis for a new industry or for revitalizing an industry. For example, the emergence of electric cars is an opportunity to do that. The adoption of industry standard architectures has been happening spontaneously in many industries in Europe and the United States. Increasingly, people in high-tech industries realize that they are all better off cooperating at the early stages of a new market formation. In this regard, I think Japan may face a special problem. Traditional Japanese business culture is very competitive—our uchi against all other uchi—but this go-it-alone thinking must be changed. Today successful strategies are based on co-opetition: cooperate upstream, compete downstream.

To create industry standard architectures, the kinds of components to be used in products and the kinds of interfaces that will be used between the components must be defined and standardized. Suppliers cannot develop new components for new kinds of products whose component types and interfaces are uncertain.

An example from China is the rapid development of an industry providing electric two-wheel vehicles (E2WVs). I want to emphasize that China did not make policy decisions specifically mandating use of a common modular architecture for E2WVs. Rather, China happened to make certain policy decisions that tried to address concerns about environmental degradation and urban crowding that had the effect of encouraging the emergence of a common modular architecture for E2WVs in the Chinese market. The use of a common modular architecture for E2WVs substantially contributed to rapidly rising product performance, falling costs, and rapid growth of the market in China.

Initially, Chinese policy makers passively encouraged the growth of an E2WV market because it would help to reduce urban pollution from other kinds of vehicles. China's policy initiatives also sought to control the size of E2WVs for safety reasons, as E2WVs would be the first powered vehicle many Chinese consumers would drive. China wanted E2WVs to be drivable safely by nearly anyone, so it limited the power of E2WV motors to 0.4 kW. This limitation enabled the product to use standard "off-the-shelf" electric motors and bicycle components, opening up access to an existing supplier base of pre-existing components. As the market grew and the limitations of existing batteries and speed controllers became apparent, companies were able to focus on developing better battery and speed controller technologies. A few companies became very good at these and became component suppliers to their competitors. A lot of money can actually be made by selling superior components to competitors. But then how does a company compete in its product market if it sells its superior components to its competitors? By competing on the "softer" parts of a product offer, like styling, customer service and support, giving better warranties, etc. Competing on building better kinds of customer interactions is increasingly what determines commercial success in global markets today. Subsequently, these pioneering E2WV companies grew domestically and succeeded in export markets because of their new appreciation of and capabilities in customer service and support.

Commentator SHIBATA Tomoatsu

I think the main point of Professor Sanchez's speech is the usefulness of modularity in new market formation and the role of policy. What are the policy implications of the example of e-bikes in China?

I completely agree that modularity is a powerful concept and a powerful tool for new market creation. The rapid growth of the PC came from open modular architectures. My research into numerical control (NC) machine tools showed that the modular concept has generated rapid growth in global competitiveness.

I think there are two key factors to making modularity a success: the design rules and the platform. Strategic partitioning is a good example of a design rule. To make a modular design process successful, a platform—a kind of common component shared by different products on which product variations can be leveraged—is required.

In the case of the Chinese E2WVs, intentional design rule-making was not necessary, because industry-standard components already existed and government regulations brought about standardization.

The government can have one of two roles in design rule-making and platform establishment: passive and reactive (i.e., the market is left alone and the government intervenes only when problems occur), and active (i.e., the government plays an active role in coordinating the establishment of platforms).

I would like to emphasize the importance of increasing the use of the modular concept. To maintain their global competitiveness, Japanese companies need to understand modularity, and the government may play a supporting role in this.

Prof. Ronald SANCHEZ

Governments may either enforce and regulate or offer incentives for companies that want to cooperate. The second option is obviously better.

Also, the organizational and managerial changes are the real challenge; not technical challenges. The changes are so significant that top management support for this process is necessary. The lesson for policymakers is that significant change towards modular design practices through a grassroots movement among designers will not be achieved. It has to start at with top management or it will not work.


Q1. I understand what you mean about designers. I also feel that this is a mistake that is often made, and that we must have top management involved

Ronald Allen SANCHEZ
Yes, the biggest mistake that can be made is to treat the conversion of a firm's designs as a technical issue. I have been in companies where top- and middle-management say they are doing modular stuff and complain that it's not working, but I look and it's not modularity. They just put the name "modularity" on a traditional development process. The intellectual understanding of what modularity is really about is critical. The pride of top managers, who don't want to be lectured by professors or policymakers, is an obstacle. Senior managers who still want to learn new ways of thinking are rare, but these are the people that need to carry modularity forward.

Q2. My question is about the potential impact on the labor market. Your idea is revolutionary. This kind of change in a company can change the labor market or the way of working.

Ronald Allen SANCHEZ
Yes, that is a significant issue at the company and national policy levels. When there is a new technology or methodology that is inherently more efficient than a conventional process, for example, enabling more variety creation with an 80% reduction in development time and cost, there will be "extra" people. My suggestion is to let them work with suppliers on developing new modular component variations, or maybe expand product lines using more modular architectures, etc. It is potentially a real problem, though, and a plan for redeploying people is needed when implementation of modular architectures makes a firm so much more efficient. It can also become a political problem, too, if it is understood that thousands of people are going to be jobless. I tell companies to plan ahead. Put new "extra" resources to work by expanding their markets and be more aggressive in developing new products.

Q3. I have a question on your view of Japanese companies. Do you consider them too traditional or outdated to introduce such a new system?

Ronald Allen SANCHEZ
I don't like to overgeneralize, but I think there is still a lot of very traditional management thinking, especially in the more industrial sectors. They may still think that product quality is just about build quality. One structural aspect of Japan that has been a problem in the past: When there is an assembler who is the top of a group where there are tiers of suppliers, and the assembler has to produce variety for the market at low-cost, the ability of the assembler to pass that problem on to the supplier base and say "you have to solve it" is actually a real source of weakness. Large firms should treat the problem of providing increasing variety and change as their own problem, to be solved by adopting modular methods in their own processes and in their suppliers' processes.

SHIBATA Tomoatsu
The important point is how Japanese companies deal with the modular phenomenon. Based on my research on Nissan's common module family concept, I am sure that the organizational capabilities of Japanese companies will be very useful in promoting modular strategies.

Ronald Allen SANCHEZ
The design rules that Professor Shibata is referring to are rules about basically the strategic partitioning of the architecture, which requires deep participation by top-level strategic managers.

*This summary was compiled by RIETI Editorial staff.