Kodak: Structural Erosion and the Fall

Download PDF

¶ I. The Company That Owned Memory

In 1976, Kodak sold 90% of all film and 85% of all cameras in the United States¹. It employed 120,000 people, its research laboratories were among the six most significant industrial facilities in the world², and the word Kodak functioned as a synonym for photography — much as Xerox for copying. The Kodak moment was not a marketing slogan, it was a linguistic fact. The company owned the category of memory.

In 2012 Eastman Kodak Company filed for protection from creditors under Chapter 11³. The shares, which in the mid-1990s had reached 94 dollars, were trading at 36 cents. Between these dates lay 36 years during which the company invented a digital camera (1975, before its commercial peak⁴), obtained hundreds of patents on digital imaging, developed key OLED display technologies, and lost the war it had itself invented.

Kodak saw the future — it had it in its own research before any of its competitors. What Kodak did not know how to do was to act on what it saw.

The fall of Kodak must be reconstructed as the anatomy of a structural erosion that ran for decades beneath the layer of visible prosperity. The actors were able, educated, devoted to their company; the organization had all the data, all the technology, all the capacity — and was structurally unable to use what it had.

¶ II. Steven Sasson and the Camera That Was Not Allowed to Exist

In December 1975 a Kodak engineer, Steven Sasson, completed the prototype of a device later shown to be the first digital camera in the world⁵. It weighed 3.6 kilograms, stored a black-and-white image at 0.01 megapixel on a magnetic cassette, and required 23 seconds to take a picture, which then displayed on a television screen. From today's perspective, a museum artefact; in 1975, a proof of principle that an image could be captured, stored and displayed without film, developing emulsion, or a chemical laboratory.

Sasson's presentation of the prototype to senior Kodak management has, in retrospect, become the most cited episode of the entire story line. Management saw the project as technically interesting, strategically irrelevant. The principal concern: should digital technology ever find commercial use, it would cannibalize film, and film was at the time the source of 70% of Kodak's profit⁶.

This reaction is not irrational. It is a classical instance of the innovator's dilemma, which Clayton Christensen⁷ formulated in 1997. Christensen showed that successful companies often cannot react to disruptive innovations — technologies inferior to existing products on the main metrics in early stages, but cheaper, simpler, or more accessible to new market segments. Disruptive innovations initially serve marginal markets that are economically uninteresting to the established player. By the time the technology improves enough to threaten the main business, it is already too late: the new player has built production capacity, knowledge, and a customer base that the established player can no longer catch up with.

Christensen's analysis of Kodak⁸ is among the most cited applications of the theory. What is essential: Kodak's decision to ignore digitalization in the 1970s was locally rational. For the company's economy at the time, an investment in technology that would cannibalize 70% of profits would have been strategic suicide. The fall of Kodak did not begin in 2000 or 1995. It began in every annual budget from 1976 onward, in which the choice between strengthening film and strengthening digital was made, and in which strengthening film was, in the short term, always the more advantageous choice.

This mechanism has a structural parallel in population genetics in the concept of an evolutionary trap: a locally optimal strategy that leads to long-term catastrophe. Short-term fitness maxima do not necessarily lie on the path to long-term sustainability; they may be inverse to it. Optimization for the current gradient ignores the topology of the fitness landscape. Whoever cannot read the topology is condemned to be led by the local maximum into a fall.

Christensen's model does not explain why Kodak could not overcome the innovator's dilemma, although it had decades to identify it, and why other firms in similar situations sometimes overcome it.

¶ III. Three Decisions No One Made

Over the 1980s there were several decision moments at Kodak that could have been turning points. None was used to alter the structural trajectory. Three examples illustrate the pattern.

1981 — the warning from Sony. Sony introduced the camera Mavica (Magnetic Video Camera), a device with an electronic sensor that stored an image on a magnetic floppy disk. The Mavica was technically imperfect and commercially insignificant, but Kodak took its release seriously enough to commission an internal study on the future of photography. Vincent Barabba, Kodak's vice-president for strategic marketing, led an analysis that concluded: digital photography would replace film within 10 years in the professional segment and within 25 years in the consumer segment⁹. The report was presented to management in 1981. Conclusions were accepted at the rhetorical level, not translated into capital allocation, research priorities, or restructuring of production capacities. The study disappeared into the archive.

1986 — the APS dilemma. In cooperation with Fuji and several other players, Kodak invested a billion dollars in the development of Advanced Photo System (APS), a new film format intended to extend the life of analogue photography by improving the user experience. APS was launched in 1996. By that time, the digital camera was commercially available and dropping in price at the rate of Moore's Law. The APS investment was a strategic bet that film had another 20 years of life ahead of it. In retrospect it is generally considered one of the worst capital allocations of the 1990s in American industry, not for the size of the loss (absorbable within Kodak), but for the signal it sent internally: management is betting the company on film¹⁰.

1989 — the marginalization of the digital team. Kodak's research team for digital imaging, led by Sasson and later his successors, was achieving breakthrough results in the 1980s: the first megapixel sensor (1986), the first professional digital camera (Kodak DCS-100, 1991), key patents for JPEG compression and for digital archiving. The team was organizationally isolated within Kodak — it sat in a separate building, reported under a different budgetary structure, and its head had limited participation in the company's strategic decisions. Innovation took place inside the firm on its periphery. Its outputs did not flow into the strategy of the whole¹¹.

The third episode goes beyond Christensen's innovator's dilemma. Kodak had the digital technology. It had the research team. It had the patents. What it did not have was an organizational bridge between research and strategy, a mechanism that would carry innovations arising in a peripheral team into the firm's core. Such a bridge exists in some organizations as a deliberate layer (skunk works at Lockheed, Bell Labs at AT&T from 1925 to 1984, PARC at Xerox in the 1970s), but in 1980s Kodak it was systematically missing.

Michael Tushman and Charles O'Reilly described this phenomenon in the 1990s as the problem of the ambidextrous organization¹². An established firm that wants to survive a disruptive change must simultaneously maintain the efficiency of its existing operation (exploitation) and build capacity for a new operation (exploration). These two environments have different cultures, motivational structures, metrics of success. Whoever tries to keep both in one organizational frame without explicit structural separation followed by integration loses either efficiency, the capacity to innovate, or both. An ambidextrous organization requires that management actively maintain a bridge between these two environments. A bridge that Kodak never built.

Tushman and O'Reilly, in a later analysis of 90 firms across various industries, documented that firms with an explicitly separated exploration unit plus an active leadership bridge between exploitation and exploration had more than triple the success rate in overcoming disruptive shifts than firms with an integrated structure¹³. Kodak's organizational architecture was typical for large established players of its time.

¶ IV. Culture as Structural Fact

The second layer of failure was culture — in the sense Edgar Schein described as the system of shared assumptions a group has developed over time to address problems of external adaptation and internal integration, and which have been effective enough that they are taken to be valid and are taught to new members as the right way to perceive, think and feel¹⁴.

Kodak's culture had been shaped for decades by one fact: the firm sat at the centre of an industry that was technologically not changing. Film emulsion in 1980 was structurally identical to film emulsion in 1930, improved in parameters (sensitivity, grain, colour fidelity), unchanged in its fundamental architecture. Kodak's engineers, marketers and managers worked within a paradigm stable for more than three generations of workers.

Chris Argyris and Donald Schön formulated in 1978 the distinction between single-loop learning and double-loop learning¹⁵. Single-loop is learning within existing assumptions: the organization improves what it does without questioning whether it should be doing it at all. Double-loop is learning that questions those assumptions themselves. Argyris documented that, in stable environments, single-loop learning becomes dominant and double-loop learning gradually atrophies, because the environment provides no occasion for it. When an environment then arises that demands double-loop, the organization is not, in capacity terms, ready for it.

Kodak in 1990 was a masterpiece of single-loop learning. To make better film, distribute it more efficiently, market it more persuasively, reduce production costs by thousandths of a cent per metre of film stock — that the firm could do at world level. What it could not do was the question of whether film makes sense at all. The culture did not, organizationally, recognize that question as legitimate. An engineer who in 1985 had raised it loudly would not have been punished explicitly; he would have been gently redirected to „more important" projects.

Linked with culture is cultural inertia¹⁶. A culture shaped over decades has its own inertia and actively resists attempts to be rewritten, even when the rational arguments are unanswerable. Resistance does not manifest as open conflict; it manifests as slowness, as „we need to think about that", as „we need more data", as the informal corridor exclusion of people proposing radical change. Whoever has no place in the culture has no voice. Whoever has no voice has no influence. Whoever has no influence has no capacity to rewrite assumptions.

In Kodak in the 1990s this had a concrete expression. People with digital experience were often hired from outside (from Microsoft, Hewlett-Packard, the technology industry of California). These newcomers either left the company within a few years (frustrated with the slowness of decision-making) or gradually assimilated into the existing culture and lost their original instincts. Both outcomes meant that the firm did not retain a population whose input was indispensable to its long-term survival. Population biology has a precise name for this: failure of immigrant integration. The immigrant carrying new variability either flees or assimilates so completely that they cease to be an immigrant; in either case, the variability is lost¹⁷.

A third cultural feature was the internal homogeneity of the decision core. Most senior management in the 1980s and early 1990s came from long-tenured employees with a background in marketing or in the financial side of film. Engineering, technology and software perspectives were under-represented. When the firm faced decisions that required technological intuition (a judgement of the pace of Moore's Law applied to the price of digital cameras), the decision core lacked the structural capacity to produce that intuition; all the models of thinking within it had been trained on a different kind of problem.

Homogeneity had a further consequence: the price of criticism. In a culture where the decision core shares the same operating frame, the critical voice always comes from outside. An external critic has no legitimacy in the internal hierarchy, so their voice carries less weight, even when their arguments are empirically stronger. The more homogeneous the core, the more readily criticism is eliminated by structure rather than by argument. Argyris introduced for this mechanism the concept of defensive reasoning: the way organizations rationalize the rejection of input that would require rewriting their identity assumptions.

The convergence of three factors — single-loop dominance, cultural inertia, decision-core homogeneity — produced in 1990s Kodak a situation in which information about the future was available but unprocessable. Barabba's 1981 study was clear, Sasson's prototypes were clear, the trajectory of digital sensor prices was clear. What was missing was the capacity to convert signal into decision.

¶ V. The Transition Moment 1996–2003

If the fall of Kodak is to be described as a process, its most dramatic phase was the period between 1996 and 2003 — seven years during which digital photography moved from the marginal segment into the mainstream, and Kodak had simultaneously available all the resources to win in this transformation, and all the structural conditions to lose it.

In 1996 Kodak reached its historic peak in revenues (16 billion dollars) and market capitalization (nearly 30 billion). In the same year it launched APS, a bet on the continuation of film. In the same year a consumer digital camera, the Casio QV-10, was available for 750 dollars and was selling at growing pace. The adoption curve was exactly where Barabba in 1981 had predicted it would be¹⁸.

Between 1996 and 2003 Kodak did several things in parallel. First, it invested intensively in maintaining the film business: marketing campaigns, distribution expansion, the APS rollout. Second, it tried to enter the digital market: it launched a line of consumer digital cameras (EasyShare from 2001), bought the online photographic service Ofoto (2001), developed inkjet printers for home printing of photographs. Third, it sought new areas of expansion: medical imaging, document imaging, business printers.

The problem was not lack of activity. The problem was fragmentation of strategic attention. None of these activities received a sufficient concentration of resources to win in its competition. Kodak's consumer digital cameras competed directly with Sony, Canon and Nikon, firms with stronger engineering bases, deeper electronics competence and faster decision-making. Kodak's cameras were good enough, not exceptional. In electronics, being good enough is a recipe for commoditization and price pressure.

Inkjet printers ran into Hewlett-Packard and Epson, who had decades of competence advantage in producing print heads. Ofoto ran into the fact that, by 2005, people were sharing photographs through Flickr and later through Facebook, and that the photograph had ceased to be primarily a printed artefact; it had become a digitally shared experience. Kodak's strategic assumption — that people would still want photographs printed — was being empirically refuted in real time, and Kodak could not react fast enough.

Kodak reacted to the digital transformation. The reaction was surface-level — addressing visible symptoms (we lack digital cameras, so let us produce them; we lack an online platform, so let us buy one) and leaving aside the structural condition from which the symptoms flowed. The firm was organized around printing the image on a physical substrate, and the future of the image did not require a physical substrate. Without rewriting that condition, every reaction was a tactical patch on a strategic crack.

The economic model. Kodak's razor-and-blades model rested on selling cameras at a low margin and earning long-term profit from selling film. Customer lifetime value (LTV) was, at Kodak, calculated over decades of continuous film consumption from a single customer. In the digital world this model did not hold: the camera was sold once, no recurring consumption existed. The transition to digital would structurally rewrite the firm's economic model, and no firm voluntarily rewrites its economic model unless catastrophe forces it to. By the time catastrophe did, it was already too late.

Empirical comparison: Fujifilm, the second-largest film producer in the world, went through a structurally similar crisis at the same time and survived it. Two reasons recur in the literature¹⁹. First, Fujifilm had a more diversified product portfolio (in addition to film, it produced chemicals, optics, paper, magnetic media), so the transition was not all-or-nothing. Second, Fujifilm invested in deep strategic analysis of adjacent markets into which its competence in chemistry and thin films could be transferred. The result was expansion into cosmetics (Astalift, drawing on Fujifilm competence in collagen and antioxidants), pharmaceuticals, and flat-panel displays. By 2020 Fujifilm was larger and more profitable than ever before. The same starting position, a different structural reaction, a different fate.

¶ VI. Bankruptcy, Patent Sales, and What Remained

Between 2003 and 2012 Kodak's fall accelerated. Revenues fell at double-digit rates. The workforce shrank from 70,000 in 2003 to 18,000 in 2012. The main factory in Rochester, where film rolls had been produced for the entire world, was gradually emptied. Rochester, for a hundred years a synonym for Kodak, was losing its economic identity²⁰.

In January 2012 Kodak filed for protection from creditors. During the bankruptcy proceedings the company sold 1,100 digital patents to a consortium led by Apple, Google, Microsoft, Samsung, Adobe and Facebook for approximately 525 million dollars²¹. These patents covered fundamental technologies of digital photography, image sharing and photographic streaming. Technologies Kodak had created and could not commercially exploit.

The symbolism of the sale is structurally more significant than its economic value. The companies that bought the patents were exactly those to which Kodak had lost its century-long position in imaging. Apple turned the camera into a standard function of a phone. Google with Android turned every smartphone into a platform for image capture. Microsoft with Windows turned the personal computer into a tool for image processing. Facebook turned the sharing of images into a basic social infrastructure. Kodak invented key technological components for each of these firms. None of them ever needed Kodak.

In 2013 Kodak emerged from bankruptcy as about a tenth of its original size, focused on commercial printing, packaging and specialty products for the film industry. It exists — as a company that sells specialized services in a marginal segment in which it has a competence advantage from the surplus of its earlier infrastructure. The category of memory it no longer owns.

The question of responsibility. Popular discourse points to specific CEOs: Antonio Pérez (2005–2014), George Fisher (1993–1999), Daniel Carp (2000–2005). The individualization of responsibility provides narrative closure. Structural analysis runs against this narrative.

No CEO of Kodak after 1990 had the structural authority to rewrite the firm at the pace that would have allowed it to survive. Any CEO who would seriously attempt radical restructuring (closing the film business in 1995, massive investment in a digital platform, divestment of Rochester operations) would face internal cultural resistance, shareholder revolt, union opposition, a media backlash in the Rochester community, and likely their own dismissal before they could complete the change. The structural trap into which Kodak fell in the 1980s was, by the 1990s and 2000s, structurally inescapable without collapse. The CEO could decide only on the speed of the fall.

Structural erosion runs long before it shows in economic metrics, and at the moment it does show in the metrics, it is too late for structural intervention. Erosion at Kodak began in the 1970s as an invisible shift in technological topology. It was diagnosable in 1981. It was undeniable in 1996. It was catastrophic in 2012. Between diagnosability and catastrophe lay 31 years, and in each it seemed the firm still had plenty of time.

¶ VII. What Can Be Read From Kodak

The fall of Kodak is sometimes reduced to one of two theses. First: Kodak did not know what was happening. Empirically false: it knew, with a head start over most competitors. Second: Kodak's leadership was incompetent. Also false: it was comparably qualified to the leadership of firms that survived (Fujifilm, Hewlett-Packard, IBM in its 1990s reorganization). The actual lesson has several layers worth naming.

First layer: surface prosperity masks structural erosion. Kodak's economic metrics were in 1995 the best in the firm's history. The erosion that, in 17 years, would destroy the firm was already advanced in 1995. Any organization whose decision frame rests primarily on economic metrics is structurally blind to most of what could destroy it, because economic metrics are lagging indicators of structural health, not early ones. Early indicators lie outside financial statements: in the share of the decision core with adversarial experience, in the ratio of exploration/exploitation allocation, in the speed at which external immigrants leave the firm, in the slowness of decisions on non-standard opportunities, in the autoreferentiality of internal debate.

Second layer: a competence centre can turn into a competence trap. What makes a firm strong in one paradigm makes it structurally unable to leave that paradigm when conditions change. Kodak's competence in film-emulsion chemistry was at the world summit of industry. In the digital paradigm it lost its value, because the substrate had moved and chemistry had ceased to be relevant — despite its original quality. The depth of competence is, over the long term, also the depth of the trap, unless that competence is continuously supplemented by competences in adjacent areas that could replace the original substrate.

Third layer: structure decides more decisively than strategy. Kodak had, after 1981, strategic documents that correctly diagnosed the digital future. What Kodak did not have was an organizational structure that would convert those documents into decisions. Strategy without a structural architecture sustaining it is paper. A structural architecture that enables decision under change — ambidextrous separation, bridges between exploration and exploitation, a decision-core population with heterogeneous experiential backgrounds, formalized mechanisms for the regular questioning of existing assumptions — produces the capacity to adapt regardless of the particular strategy the firm chooses at the moment.

Fourth layer: individual responsibility is limited within structural traps. If a firm has been organized for decades to prefer short-term stability to long-term adaptability, no individual actor inside it has the authority to rewrite that system in a reasonable time. The argument points toward earlier decisions about structure rather than toward absolving individual actors. Structural choices that produce long-term traps are usually made in a period of success — short-term maxima do not consider long-term topology. Whoever decides in a period of success shapes the firm's future more deeply than those who decide in a period of crisis.

Fifth layer: between individual reason and collective decision lies a structural layer with its own dynamics. The actors at Kodak were individually rational. Their collective decisions, aggregated over time, led to the firm's fall — the standard output of a system whose aggregation mechanism systematically prefers certain classes of information (short-term, quantifiable, compatible with existing frames) and dampens others (long-term, qualitative, requiring the rewriting of frames). Changing individual actors does not change such a system, until the aggregation mechanism is changed.

Kodak had all the information, all the technology, all the resources. What it did not have was the structural position from which it could read that information in a way that would yield a decision capable of saving the firm. This position is an architectural decision that has to be made before it is needed — individual talent or collective effort will not produce it in a crisis. Had it been made in time, it would have seemed unnecessary, because the firm would still have been doing well.

¶ VIII. The Erosion No One Stopped

In 2024 the building of Kodak's main research centre in Rochester — Building 82, in which Steven Sasson completed the first digital camera in 1975 — was partially demolished²². Some parts were preserved as a historic landmark. Most were torn down to make room for a logistics centre for online retail.

Kodak exists. Its current market capitalization is in the order of hundreds of millions of dollars, less than one percent of its 1996 peak. It produces specialty film products for the cinema industry, where part of demand is stabilized by directors such as Christopher Nolan who prefer physical film. It produces commercial printers. It runs a small laboratory for film archiving. It lives off the long tail of its formerly dominant position.

The whole arc of this history — from complete dominance through decades of structural erosion to bankruptcy and to a small existence in marginal segments — takes about 36 years. Thirty-six years during which the firm was never without information; it had data, warnings, diagnoses, prototypes, research reports and Vincent Barabba's 1981 study. The structural capacity to convert it all into decision was what it lacked.

The fall of Kodak was an organizational failure to decide on the basis of information that was available. Between knowing and deciding lies a layer of structural architecture, invisible in good times and revealed in its absence at the moment when one needs to stand on it.

The practical question every organization reading the Kodak story may pose itself is therefore not „will we be Kodak?"; no one ever knows. A better question: where in our own organization is structural erosion running, currently masked by surface prosperity? In particular systems — firms, disciplines, institutions, relationships, life projects — structural conditions similar to those at Kodak in 1985 are present, and their diagnosis is a task for the kind of attention Kodak then lacked.

¶ References

Munir, K. A. The Demise of Kodak: Five Reasons. Wall Street Journal, 26 February 2012. Lucas, H. C. Jr., Goh, J. M. Disruptive Technology: How Kodak Missed the Digital Photography Revolution. Journal of Strategic Information Systems, 18(1), 46–55, 2009.
Estrin, J. Kodak's First Digital Moment. The New York Times, 12 August 2015 (interview with Steven Sasson). Mui, C. How Kodak Failed. Forbes, 18 January 2012.
Eastman Kodak Company, Form 8-K filing, U.S. Securities and Exchange Commission, 19 January 2012. Kodak Files for Bankruptcy Protection. Reuters, 19 January 2012.
Sasson, S. J. Disruptive Innovation: The Story of the First Digital Camera. IEEE Annals of the History of Computing, 36(2), 84–88, 2014.
Sasson, S. J. (2014), op. cit. Nakamura, J. Image Sensors and Signal Processing for Digital Still Cameras. CRC Press, 2005, ch. 1.
Lucas & Goh (2009), op. cit. Swasy, A. Changing Focus: Kodak and the Battle to Save a Great American Company. Times Business, 1997.
Christensen, C. M. The Innovator's Dilemma: When New Technologies Cause Great Firms to Fail. Harvard Business School Press, 1997.
Christensen, C. M., Raynor, M. E. The Innovator's Solution. Harvard Business School Press, 2003.
Barabba, V. P. The Decision Loom: A Design for Interactive Decision-Making in Organizations. Triarchy Press, 2011.
Gavetti, G., Henderson, R., Giorgi, S. Kodak and the Digital Revolution (A). Harvard Business School Case 9-705-448, 2005.
Lucas & Goh (2009), op. cit. Mendonça, S. The 'Sailing Ship Effect': Reassessing History as a Source of Insight on Technical Change. Research Policy, 42(10), 1724–1738, 2013.
Tushman, M. L., O'Reilly, C. A. III. Ambidextrous Organizations: Managing Evolutionary and Revolutionary Change. California Management Review, 38(4), 8–30, 1996.
O'Reilly, C. A. III, Tushman, M. L. Lead and Disrupt: How to Solve the Innovator's Dilemma (2nd ed.). Stanford Business Books, 2021.
Schein, E. H. Organizational Culture and Leadership. Jossey-Bass, 1985 (4th ed. 2010).
Argyris, C., Schön, D. A. Organizational Learning: A Theory of Action Perspective. Addison-Wesley, 1978. Argyris, C. Teaching Smart People How to Learn. Harvard Business Review, 69(3), 99–109, 1991.
O'Reilly, C. A. III, Caldwell, D. F., Chatman, J. A., Doerr, B. The Promise and Problems of Organizational Culture: CEO Personality, Culture, and Firm Performance. Group & Organization Management, 39(6), 595–625, 2014.
Hartl, D. L., Clark, A. G. Principles of Population Genetics (4th ed.). Sinauer Associates, 2007.
International Data Corporation (IDC) market data 1996–2003 cited in: Hopper, M. The Decline and Fall of Eastman Kodak. Business Strategy Review, 21(3), 14–21, 2010.
Komori, S. Innovating Out of Crisis: How Fujifilm Survived (and Thrived) as Its Core Business Was Vanishing. Stone Bridge Press, 2015.
Brayer, E. George Eastman: A Biography. Johns Hopkins University Press, 1996. The City of Rochester Economic Development Reports 2005–2015.
Kodak Sells Imaging Patents for $525 Million. The New York Times, 19 December 2012. The buyer consortium included Apple, Google, Microsoft, Samsung, Adobe, Facebook, Amazon, HTC, RIM, Huawei and others.
Kodak's Building 82 Partial Demolition Begins. Rochester Business Journal, 2024.

Jan Vytřísal

Continue

Further essays.

All essays →