1. Innovation and Schumpeter’s Theories
2. Incremental vs. Radical Innovation
3. Henderson – Clark Model
4. S-Curves
5. The Teece Model
6. Abernathy – Utterback Model
7. Disruptive Innovation

As I wrote in the first part of the series “despite the popularity we can say that innovation management is still an immature science. There is no dominant theory on the field and little agreement among managers and academics alike regarding what affects a company’s ability to innovate”.

The absence of a dominant view implies that we need to consider different theories for different situations, possibly using the insights gained in one model to complete the flaws of another.

This article is the seventh part of the Innovation Management Theory series, you can check the previous six clicking here.

Disruptive Innovation

The term disruptive innovation as we know it today first appeared in the 1997 best-seller The Innovator’s Dilemma. In the book Harvard Business School professor Clayton Christensen investigated why some innovations that were radical in nature reinforced the incumbent’s position in a certain industry, contrary to what previous models (for instance the Henderson – Clark model) would predict. More specifically he analyzed extensively the disk drive industry because it represented the most dynamic, technologically discontinuous and complex industry one could find in our economy. Just consider that the memory capacity packed into a square inch of disk increased by 35% per year, from 50 kilobytes in 1967 to 1,7 megabytes in 1973, 12 megabytes in 1981 and 1100 megabytes in 1995.

Christensen describes how one of his friends was responsible for that choice when he commented that “those who study genetics avoid studying humans because new generations come along only every thirty years or so, it takes a long time to understand the cause and effect of any changes. Instead, they study fruit flies, because they are conceived, born, mature and die all within a single day. If you want to understand why something happens in business, study the disk drive industry. Those companies are the closest things to fruit flies that the business world will ever see”.

Sustaining vs. Disruptive Innovation

The central theory of Christensen’s work is the dichotomy of sustaining and disruptive innovation. A sustaining innovation hardly results in the downfall of established companies because it improves the performance of existing products along the dimensions that mainstream customers value.

Disruptive innovation, on the other hand, will often have characteristics that traditional customer segments may not want, at least initially. Such innovations will appear as cheaper, simpler and even with inferior quality if compared to existing products, but some marginal or new segment will value it.

The disk drive industry

The first disk drive was developed in IBM’s San Jose research laboratories, around 1954. It was as large as a refrigerator and it could store 5 megabytes of data. By 1976 $1 billion worth of disk drives was being produced annually, divided between integrated producers (IBM, Control Data, Univac, and others) and OEM producers (Nixdorf, Wang, Prime, and others).

By 1996 the disk drive market was worth $18 billions, but out of the many companies that were operating in 1976 only IBM was still in the market. About 129 firms entered the market during that period, and 109 of them ceased to exist. Most of the technological discontinuities that emerged in the industry were sustaining innovations. For example in the 1970’s the oxide disks started to reach a physical limit (in terms of bytes of information contained), forcing the leading companies to develop an alternative. IBM, Control Data and other incumbents invested more than $50 million developing thin-film coatings, and virtually all of the established firms managed to keep their position in the face of such sustaining innovation.

In contrast, there have been very few disruptive innovations over the same period, but those were responsible for the downfall of established firms. As Christensen highlights the most important disruptive technologies were the architectural innovations that shrunk the size of the drives, from 14-inch diameter disks to 8’, 5.25’ and 3.5’, and then from 2.5’ to 1.8’.

The passage from 14-inch to 8-inch disks

The 14-inch disk drives were produced to supply mainframes, and the two parameters mainframe producers would consider as a performance measure were the capacity and the cost per megabyte. Around 1980 some new firms ( including Micropolis, Priam and Quantum) developed smaller 8-inch drives, but those packed 10 to 40 megabytes of capacity while mainframes were demanding 400 megabyte drives. The leading companies producing 14-inch drives could have developed 8-inch drives internally without much of a trouble, after all the technological innovation involved was simple and architectural in nature. Why they did not, you might ask? Because their main customers, the mainframe manufacturers, were not interested at all in the smaller hard drives.

The new entrants would not be able to sell the 8’inch drive for mainframe producers; consequently they were forced to look for new applications that would eventually value the characteristics of their product, mainly the reduced size. They found such application in the minicomputer. Manufacturers likes DEC, Prime and HP were willing to pay a higher cost per megabyte in order to get smaller disk drives.

Customer demand for capacity was growing at 25% every year, while producers of 8’inch disk drives found that with sustaining innovations they were able to increase their disk capacity by 40% every year, almost twice as fast. Notice that most disruptive innovations will improve faster than what is demanded by mainstream customers, meaning that after some time disruptors should be able to attack established firms as the figure below illustrates.

After some years the 8-inch drives were offering an inferior cost per megabyte than 14-inch ones and the capacity was already enough to supply lower-end mainframes. The incumbents of the 14-inch generation witnessed their markets being invaded; and obviously it was too late to react. Only one third of 14-inch producers managed to make the transition into the new technology, but eventually all of them went out of the market.

The story repeats itself

The same pattern was observed when Seagate introduced the 5.25-inch disk drive, its capacity of 10 megabytes was of no interest to microcomputer producers. Seagate and the other 5.25-inch drive producers (Computer Memories, International Memories, Miniscribe and others) needed again to find a new application for their product. As Christensen writes “they went by trial and error, selling drives to whomever would buy them”.

The application was found in the desktop computers, and just like with under the 8-inch generation 5.25-inch disk drives improved their performance via sustaining innovations faster than what minicomputer customers were demanding. After a couple of years 5.25 drives invaded the 8-inch market, and virtually all the leaders under the 8-inch technology struggled out of the market.

By now I think you are already guessing what happened when the 3.5-inch disk drives emerged right? Almost all of the established players under the 5.25 generation were again forced out of the market.

What are the reasons?

How could we possibly explain such pattern? Clearly it was not a matter of technological complexity, incumbents were perfectly able to cope with the architectural innovations that shrunk the disk drive sizes. Some of the leading 5.25-inch producers even developed the 3.5-inch drives internally before new entrants, but they shelved the innovation as soon as their mainstream customers demonstrated no interest in them (5.25 drives were used for desktop computers while the 3.5-inch ones would be employed in notebooks). According to Christensen the crucial factor to understand is the concept of value network, described as “the context within which a firm identifies and responds to customers’ needs, solve problems, procure inputs, react to competitors and strive for profits”.

First of all, operating under such value network might lead a company to “listen too much” to its main customers. As a result it will not recognize potentially disruptive innovations that serve only marginal customers. Secondly large companies will not be interested in small markets; they hardly offer significant growth opportunities. Again this will lead companies to completely ignore the disruptive innovation or to wait until the market is “large enough to be attractive”. That is exactly when new entrants attack incumbent’s turf, and by that time it is usually too late.

What is a possible solution?

In order to solve both of these problems organizations should create an independent business unit whose size matches the emerging market. Quantum Corporation, a leading producer of 8-inch drives, recognized that 3.5-inch drives could have some applications in the computer industry, but they were not sure what those applications were exactly. Instead of shelving the project they created a spin-off unit to develop such 3.5.inch drives. After ten years the 8-inch market had completely disappeared while their small venture had grown to become the world largest disk drive producer.

Other examples of disruptive innovations:

• telephone (disrupted the telegraph)
• semiconductors (disrupted vacuum tubes)
• steamships (disrupted sailing ships)

Professor Kim Clark recently published a working paper where he argues that companies should invest in architectural knowledge to create sustainable competitive advantage. In his own words: “by studying the underlying cause-and-effect relationship in a complex architecture, a firm can identify “bottlenecks” and redesign the interfaces of key components to make them more modular (…) and it can outsource more activities without sacrificing either performance or cost”.

According to the paper this was the strategy used by Sun Microsystems against Apollo Computer in the 1980s and by Dell against Compaq and other PC manufacturers in the 1990s. I think a similar approach is also being by used by Chinese manufacturers in many sectors including motorcycles and telecommunications equipment.

The motorcycle producers, for instance, developed products with clear standards and specifications so that any supplier would be able to provide individual parts without the problem of incompatibility. The result is a combination of modular innovations from the side of the suppliers and architectural innovations from the side of the manufacturer.

In my opinion this pattern will be observed in many other industries. It fosters innovation and collaboration creating win-win situations. To read the full paper click here.

The Abernathy – Utterback Model

Fluid Phase – the first phase they called fluid phase, where technological and market uncertainties prevail, a great deal of changes take place conteporaneously and outcomes may vary significantly. It is almost a large experimentation game in the market place. The manufacturing process relies on high-skilled labour and general purpose equipment, there is almost no process innovation and the many, small firms competing will base their advantage on differentiated product features. Competition will not be as fierce as in later phases because companies have no clear idea on potential applications for the innovation, nor on what direction the market might grow. There is low bargaining power from suppliers since no specialised materials are used in the production. The major threats come from the old technology itself and from the entrance of new entrants if the innovation was radical and competence-destroying.

In this phase a company can follow two strategies. Firstly it can try outmaneuvre the competitors and establish its product as the “dominant design”(explained in the next phase). This strategy will involve agreements with distributors and marketing investments to affect customers’ perceptions. Alternatively the company can try to take control of complementary assets and wait for the appearence of the dominate design; then once the standard becomes clear it will try to secure most of the profits basing its competitive advantage on the distribution channels, supplier contracts, complementary technologies, value-added services, and others.

.

Transitional Phase – as producers start to learn more about the technology application and about customer’s needs some standardization will emerge. Usually by this time the acceptance of the innovation starts to increase and the market starts growing, signals that we are entering into what the authors called the transitional phase. The convergence pattern in this phase will lead to the appearance of a “dominant design”, which is a product degisn whose main components and underlying core characteristcs do not vary from one model to another, it often comes out as a new product syntethised from individual innovations introduced independtly in previous product variations. In Utterback words “the dominant design product has features that competitors and innovators must adhere if they hope to command significant market share following” (Utterback, J. M. – Mastering the Dynamics of Innovation, Harvard Business School 1994) .

Winning the battle for the dominant design is desirable because it will enable the firm to collect monopoly rents (given imitability is not so high or Intelectual Property Rights can be applied). Even is the standard is “open” the developer can build complementary products or enhanced versions faster, possibly establishing a new standard in the future. Microsoft managed to establish Windows as the dominant design for graphical operating systems also thanks to its previous dominant position with the MS-DOS operating system. The threat of new entrants on the transitional phase is linked to the technology involved in the innovation, if it is proprietary incumbents are favoured. Firms in this phase will use strategies to consolidate their product positioning and start increasing production capacity and process innovation in order to face the next phase, the specif phase.

.

Specific Phase – after the appearance of the dominant design competition will shift from differentiation to product performance and costs. Companies now have a clear picture of market segments and will therefore concentrate on serving specific customers. Manufacturing will use highly specialised equipment and employing high-skilled labour become less important since there is a commoditisation taking place, which in turn means that bargaining power of both suppliers and customers will increase.

Competition becomes more intense and the market moves towards an oligopoly. As a consequence incumbets are able to secure their position through supplier relations, distributtion channels and other complementary assets that will create entry barriers to new entrants.

.

Summing up, the Abernathy – Utterback is one of the most complete and solid models we have covered so far. The analysis they carried was very broad, ranging from technological impact upon products and processes to market dynamics and competition to organizational structure and strategic decisions within companies.

The Teece Model

David Teece clarified that two factors – imitability and complementary assets – will have a strong influence in determining who will ultimately profit from an innovation. Imitability refers to how easily competitors can copy or duplicate the technology or process underpining the innovation. There are many examples of barriers a company could use to protect itself from imitation, including intelectual property rights, complex internal routines or tacit knowledge.

Consider the case of RC Cola, it was the first firm to introduce a diet cola on the market, but since it could not protect itself from imitation soon Pepsi and Coca-Cola jumped in, and using their complementary assets (distribution channels, brand name, etc.) they appropriated all the profits of the segment. Complementary assets, therefore, are equally important. They include any activity that gravitates around the core innovation such as distribution channels, reputation, marketing capabilities, strategic alliances, customer relationships, licensing agreements, among others.

Analysing the two dimensions we conclude that if imitability is high and complementary assets are freely available or unimportant it will be difficult to make money out of the innovation (exeptions can be made at the very short run). If instead complementary assets are tightly held and important and imitability is once again high, the holder of such assets will be the one profting on the innovation, independtly of who developed it like in the diet cola case. If imitability is low the innovator will find himself in a much better position. When complementary assets are not controlled by other economic actors he will be able to collect most of the profits being generated. When, on the other hand, complementary assets are important and tighly held negotiation will take place, profits will be shared in proportion to bargaining power of the parts involved.

The Teece model can be used not only to predict who will profit from an innovation but also to understand what company will have higher incentives to invest in certain innovations. The threat of imitation and the importance of complementary assets had already being used under other frameworks, but usually they were employed individually or were used to dissect the overall market structure (i.e. Porter’s Five Forces) and not the innovation dynamics. The major flaw one can find the this theory is the lack of empirical evidence, which results from the diffculty to isolate the imitability and the complementaity effects from other factors.