2.2 On modularity
2.2.1 Theories related to modularity
國
立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
“Building up architectural knowledge about the component was recognized as a critical success factor for suppliers to win design competition”
2.2 On Modularity
This section deals with the general concept of modularity so to provide the reader with solid background information on the research. Even though theory of modularity is fairly simple, I spend quite a lot of space to cover this topic thoroughly, so that the reader can get a deep understanding of all the challenges in modular product development. The basic modular theory is introduced followed with the major challenges to modularity – the architectural innovation and the modularity trap.
2.2.1 Theories related to Modularity
Modularity can be seen as a general concept, which help us understand systems and their organization. According to Schling (2000) modularity is an abstract term (a continuum) which describes the degree, to which system`s components can be separated and recombined.
Therefore all systems are characterized by some degree of coupling between components, and only very few systems have components which are completely inseparable and cannot be
recombined. This is the reason we can say that almost all systems are, to some degree, modular.
Simon(1995) in his pioneering research shows examples of modularity from very wide specter of situations and postulate, that modularity might be found in almost all entities around us – social, biological or technological. Simon use the familiar example of biological organism to introduce the very basic concept of modular system and its ability of decomposition ‐ “which is composed of organs, which are composed of cells, which contain organelles, which are
composed of molecules and so on.
Simon use the term “hierarchically nested systems” – meaning that at any unit of analysis, the entity is a system of components and each of the components is, in turn is a system of finer components, until we reach a point, where the components are “elementary particles” or until
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
question is, whether the systems can be put back together and still be functional in the same manner, as before – and is it necessary for them to be re‐configured in the same, original way to keep working? This is where we can differentiate between the high and low levels of modularity (Sanchez, 1995).
High level of modularity are those systems, whose components might be disaggregated and recombined into new configurations – and possible substituting many new components into the configuration of the system with minimal or none loss of functionality. These components are relatively independent on each other, and the only dependence is to the overall system – the architecture of the system.
Still, there always will be some configurations, which would be more powerful – the
components in that particular combination would overall provide better system output than other configurations. This optimization is a crucial concern during the design of modular artifact.
The designer must take into consideration the possible advantages (trade offs) of fully modular and decomposable product, or product with lower modularity, however having possibly higher efficiency. Schling (2000) describes such phenomena as “synergistic specificity”. These are situations, when through the combination of components we can achieve functionality unobtainable through combinations of more independent components (components with higher modularity). Such an architecture functions will be unchallenged by more modular systems, however later changes into these architectures are very difficult to do, as the components are more tightly organized and more deep interconnected.(Schling, 2000).
Baldwin and Clarke (2000) define modular systems as systems, which are composed of units (or modules) that are designed independently but still function as an integrated whole.
Modularity means building complex products or process from smaller subsystems, that can be designed independently yet function together as whole. They do n comprehensive research of the different communication patters among the components, but also companies responsible for them.
The modular design allows the creators to use components designed by other entities than the creators (people, but also companies or organizations) so the end product will still work – this is due the compatibility of the components with each other and the interfaces.
Since not only one company is usually in charge of more complex products, this brings us to knowledge management and information sharing across these organizations. Baldwin and Clarke (2000) specify different types of information, accessible to only selected classes of users or designers – the visible information is the architecture designed by the architect. This
information is accessible to everyone who wants to participate on the modular design, and the
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
participants must follow it in order so their modules would be compatible with the overall architecture. The hidden information however is created by the designer of the module, and they do not need to enclose it to others – since this would leverage their competitive advantage.
In plain words – as long as the component is functional within the architecture and follows its purpose, its designers does not need to share any more information than necessary. The visible design rules are therefore decisions, that affect subsequent design decisions, and they should be specified early in the design process and communicated broadly to those involved.
Visible rules might be further divided into 3 categories:
Architecture –specifies what modules will be part of the system and what their functions will be
Interfaces – that describe in detail how the modules will interact, including how they will fit together, connect, and communicate
Standards for testing a module`s conformity to the design rules (e.g. can module X function in the system?) and for measuring one module`s performance relative to another (how good is module X versus module Y?)
The hidden design parameters are decisions, which do not affect the design beyond the local module – they have no influence over any other modules. They can be chosen late and and changed often and they do not have to be communicated to anyone beyond the module design team – so long they fit the original architecture and do their functionality does not interfere with any other modules in the system, or they change the functionality of the whole system. In other words, the hidden rules might be also described as component know‐how.
The standards are the main advantage of the modular design. Once established and articulated to the suppliers, they allow a face speed competition as well as price reduction and
innovation – since they allow numerous firms to experiment with a variety of implementations, and this resulting complexity far exceeds what could be produced inside a single firm. This experimenting and miss and match process allows a great deal of innovation, and the development of an product platform is much speeded. (Chessbrough and Kusunoki, 2001)
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2.2.2 Modular product design and Modular Organization Design
Literature differentiates between Modular product design, modular organization design and also Modular manufacture design (Ernst, 2005).
However for the purposes of this paper, I will work with the first two only, since they are closely connected with the topic of product development and knowledge management, where the third deals with manufacture strategy and tactics.
All modular systems might be described as “loosely coupled” (Weick, 1976). This means that modularity involves independence between the particular modules of the system, and changes of one module will not affect design of any other modules. This independence is intententional.
In computing, these might be described as systems, where the components use little or no knowledge of the definitions of other separate components. All they need, is the access to the visible rules and as long it is compatible with the architecture, they can work and improve their own product (component) – until they reach the physical technological constraints of the given architecture.
It is very important to note, that the term “loosely coupled” is strictly abstract. So for example a computer, even though the modules and all the components are tightly integrated and all are physically interconnected (and sometimes even touching each other) on the motherboard, we can say that all the modules are “loosely coupled” – as they are interchangeable (by substitutes) and thus create a platforms for “economies of substitution” (Garud and Kumaraswamy, 1993) ‐ a range of component variations in order to configure a range of product variations.
For the reasons explained earlier, the modular product architecture is flexible – by substituting modules and using standardized interfaces between components enables variation. Particular modules, which over time become bottlenecks might be easily exchanged with more powerful and suitable ones. This “mixing and matching” provides the whole organization with strategic flexibility and creates potentially large number of product variations, distinctive functionalities, new features and/or performance levels (Sanchez, 1994).
Once the outputs of particular processes and components are clearly specified and explicitly announced, the design of these might be than partitioned into tasks, that can be performed autonomously and concurrently by loosely coupled organizations (von Hippel, 1990). The interfaces between the modules would be a building block for inter‐firm or inter organizational communication channels for loosely coupled development process – so as long as the designers follow the design rules, the final product will be compatible with the rest.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
This is in contrast with the traditional tightly coupled organization structure, coordinated by a managerial authority hierarchy ‐ an organization design typically achieved within single firm.
Within such a company, the engineering effort would follow methodology of constant optimization – and this tries to obtain the highest level of product performance within some cost constraint or the lowest cost for a product meeting a minimum performance constraint (Sanchez and Mahoney, 1996). Such a design typically leads to product designs, which are integrated, with “tightly coupled component designs” – so any change in one component would inevitably bring an necessary change to other components. For this reason, systems with such a structure would need much more extensive managerial coordination and care during any design processes. Moreover, the officer (or team of managers) in charge would have to have very extensive knowledge of most of the processes and components involved. In contrast, the modular organization can effectively eliminate a big deal of managerial and coordinatinal work by simply articulating what is the architecture about, and from this point the need for
managerial coordination is lowered.
Therefore it save to say, that modular product design would inevitably lead to modular organization design – the organization of the network of companies, involved in the
development of particular product or process would ultimately adapt the architecture of that given artifact – and they would be organized in very same way as are the components in the artifact.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2.2.3. Dominant Design
Both the theoretical studies as well as the empirical evidence support the premise, that the new technologies are not fully developed when entering the markets. The life of an technology is often characterized by a changing cycles of turbulent development, experimenting, learning – by‐ doing and subsequent reevaluation of the learned and its application.
When a new technology arrives to the market, there is a big deal of confusion. Even though there might be a good understanding and consensus on what the new technology ought to do, there is just a little agreement on what the major subsystems of the product should be, or how should they be connected – put together.
As the product architecture is non existent during this phase or in better case underdeveloped, the companies will experiment with different technologies and processes as they will try to put all the necessary subsystems they wish to incorporate into the new product together – and many concepts will occur during this period.
However, it is usual that only a small number of these concepts will survive on the market competition and those that survive will become the “dominant design”.
According the Utterback’s model (1994) there are 3 phases in every industry to set an Dominant Design. The first phase is the so called “Fluid Phase” – when a great number of technological solutions compete with each other. At this time, neither the problem nor the way the problem should be solved are very well defined and there is high variety of different solutions and approaches to the issue.
During the Transition phase, there is a quick elimination and shakeout of majority of ideas – and companies connected with these ideas as well – until the third and final phase – the dominant design emerges and only very few companies remain. As from now – since there s already and dominant design on the markets – the innovation effort would move from product innovation to process innovation. (Utterback, 1994)
“It is (dominant design) equivalent to the general acceptance of particular product architecture and is characteristic of technical evolution in a very wide range of industries.” (Clark, 1985) Also, the companies will starting from now on to try to figure out how can they differentiate their products from their competitor`s based on the same dominant design.
As an example, the emergence of the first car brought many different vehicle` concepts. During the experimentation period, there were different engines and propellant systems (electric, steam, gasoline etc., with steering wheels or tillers, with tires or without, wooden or metal wheels etc.). However, once the dominant design arises, the diversity is eliminated and only
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
one concept will eventually prevail. For cars, it was the gasoline engine that provided the traction, but also had the concept of transmission connecting the wheels with the steering, particular chassis style etc. Simply put, the dominant design is a set of rules, that specifies arrange of basic choices about the design, which are not redesigned with every subsequent generation of the product. All the further improvement will take part on the component level, within the framework of the system (architecture). (Henderson and Clark, 1990)
As a consequence, the companies do not need to learn anything new about the competing architectures – since the prevailing architecture is the one of highest priority to them and they can easily turn to deepening their component knowledge.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
2.2.4 Architectural innovation and Modularity Trap
From the incremental innovation standpoint, modular design has huge influence on the innovativeness of an organization. It allows the companies to focus on their own module and innovations inside this particular part. This is true as long, as the architecture of an artifact (and indirectly also the architecture of the manufacturing organization as seen before) remains intact and unchanged. In that way, the “volume of Information” (Ernst, 2004) is reduced and hence the amount of knowledge sharing, that is required for coordination is down as well. The burden of managerial coordination is lowered and these talents may be employed in other activities.
“When standardized interfaces in modular architectures are used to coordinate the product creation…processes, those processes can become self – managing. Both mid‐level and senior managers can redirect much of their time and attention from routine tasks of monitoring, problem solving and intervention in those processes to refocus on essential tasks of strategic…goal setting” (Sanchez and Collins, 2001)
The companies can thus perform that activities, which have the higher value added and that are critical tasks for their undergoing (“core competencies”). Usually, these activities also generate the highest margins. The others activities where they cannot obtain high margins and which are not critical to their business would be simply outsourced.
This is what Langlois (2003)means with his statement, that modularity causes the “visible hand of managerial coordination” to vanish.
In his opinion, “modularity reduces the need for management and integration to buffer uncertainty…and the buffering functions of the management are evolving to the mechanics of modularity and market”(Langlois, 2003).
Both these arguments imply, that when considering modular design, the need of involvement required through corporate management is decreasing.
The important, but unanswered question is – what happens once the architecture changes, or begins to change?
The literature mentioned above does not work with such a possibility and simply ignore this option, where they present modularity as the ultimate modern management tool. However, as the experience goes, that architectural innovations are likely to happen in any industry, either once the old architecture reaches its technological zenith, or when the changes in the market simply knocks out the old architecture as outdated or not to the current customer`s need.
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
Architectural innovation is a solid challenge to all companies, even to those strongly integrated ones. This is due to the fact that it destroys the usefulness of the architectural knowledge for good, and since this sort of knowledge becomes embedded in the structure and information‐
processing of established organizations, this destruction is difficult to recognize and subsequently correct. Although the architectural innovation often does not bring any new technological advance, it involves new set of engineering and scientific principles and often opens completely new markets as well as potential applications.
As explained in the earlier chapters, the difference between product development of the whole system and of only its parts is great. Modular theory is more than suitable for the later, where no need of extensive coordination of development among modules in necessary. However, for the survival in the conditions of architectural innovation, an insight or better understanding of both is crucial for the company.
The architectural innovation is defined as “the reconfiguration of existing product technologies” – “…the reconfiguration of an established system to link together existing components in a new way” (Henderson and Clark, 1990) ‐ in other words the architectural innovation re‐shuffle the existing components, in a new and innovative way. This destroys the usefulness of the architectural knowledge, however – and this is very important ‐ it preserves the usefulness of the knowledge about the components. And that is why it is so important, that any company should try to reach out and have deeper understanding of both sets of the
knowledge, and not just blindly outsource to its suppliers without having better understanding of the components they employ.
Still, Henderson and Clark (1990) point out – that the sole components may not stay the same – untouched – by the innovation. Actually on a contrary, the architectural innovation is often triggered by a change in a single component – that will create new integrations and new
linkages with other components. However, the “core concept” behind the each component – or the knowledge about it‐ it wouldn`t change that much, but rather evolve or adjust to the new system.
The reason why the companies fail to react to architectural innovations is, according to the authors due to the fact, that they concentrate too much on the development of their “core
‧
國立 政 治 大 學
‧
N a tio na
l C h engchi U ni ve rs it y
component, but the whole model of its product needs to be redesigned. This process might be painfully time consuming, as the primary reaction to new threat would in most cases be to just do some changes in number of components, rather rethinking the whole concept.
The second issue is the organization`s adaption to the new architecture. The challenge is quietly
The second issue is the organization`s adaption to the new architecture. The challenge is quietly