History of IBM

1888–1924: The founding of IBM

Tabulating Machine Corporation plant in 1893.

IBM's history dates back decades before the development of electronic computers when it developed punched card data processing equipment. It originated as the Computing Tabulating Recording Corporation (CTR), which was incorporated on June 15 1911 in Endicott, New York a few miles west of Binghamton.

CTR was formed through a merger of three separate corporations: Tabulating Machine Company (founded 1896 in Washington D.C.), the Computing Scale Corporation (founded 1901 in Dayton, Ohio) and the International Time Recording Company (founded 1900 in Endicott, NY). The president of the Tabulating Machine Company at that time was Herman Hollerith, who had founded the company. The key person behind the merger was financier Charles Flint, who brought together the founders of the three companies to propose a merger and remained a member of the board of CTR until his retirement in 1930.

Thomas J. Watson Sr., the founder of IBM, became General Manager of CTR in 1914 and President in 1915. In 1917, the CTR entered the Canadian market under the name of International Business Machines Co., Limited and in February 14 1924, CTR changed its name to International Business Machines Corporation.

The companies that merged to form CTR manufactured a wide range of products, including employee time-keeping systems, weighing scales, automatic meat slicers, and most importantly for the development of the computer, punched card equipment. Over time CTR came to focus purely on the punched card business, and ceased its involvement in the other activities.

Early Education and R&D

It was in the middle of the Depression, in 1932, that Watson formally established an Education Department at Endicott, then IBM's main plant; with the first three engineers graduating in 1934. By 1937 almost half the 32,000 factory employees at Endicott were enrolled; in 52 classes, studying 24 different subjects. Tool-making was an especially important skill being taught, but even the factory supervisors and executives had to attend its "Supervisor's School".

The education facilities were conveniently placed near the 'Homestead', the clubhouse which was the hub of IBM's leisure complex; as well as being the hotel for customers attending education. It was a near idyllic location for doing business and, according to his son Tom Jr., Thomas Watson Sr.—who had a permanent suite of rooms there—said it was his favorite IBM location.

The earliest research, at Endicott, focused on product needs, indeed largely on quality control, rather than the more fundamental investigations which were later, in the 1980s, to earn IBM two consecutive Nobel prizes for Physics. For example, the paper-testing laboratory was then especially important; where, by 1937, Endicott was producing up to 10 million cards a day. But a 'Methods Research Department' which, in the context of punched cards, was effectively the start of IBM's 'software development' (albeit that it then revolved around the wiring of plug-boards rather than computer programs), was started at much the same time.[1]

World War II

During World War II 274 mobile IBM punched card units followed the front-line controlling the logistics of supply; and, in the process, exposing a new generation of users to the benefits of data processing. In any case IBM's business, in common with that of many other manufacturers, grew substantially on the back of the increased economic activity generated by the war effort. From a revenue of $34.8 million in 1939 sales climbed to a high of $143.3 million in 1944, before falling off slightly at the end of the war.

Towards the end of the war the research arms of the services started to develop requirements for the first true computers; the nuclear weapons industry perhaps being the most significant potential customer. In terms of commercial applications, however, the Bletchley Park code-breaking work in England was very much a side event; though it did produce Alan Turing as one of the great theoreticians of computing. IBM's contribution, in 1943, was the Mark I, an electro-mechanical monster 51 feet long by 8 feet high and containing 530 miles of wiring! It cost $200,000 (plus a donation of another $100,000 to cover operating costs) and was installed at Harvard.

Even with this example IBM did not really get into computing in its modern form. It built a further electro-mechanical marvel, the Mark II, the IBM Selective Sequence Electronic Calculator (SSEC) with 21,000 relays and 13,000 vacuum tubes, which it grandly installed in its headquarters in 1948. The real action was, however, elsewhere at the University of Pennsylvania where the ENIAC (Electronic Numerical Integrator And Computer) team were developing the first electronic computer; unveiled at the University of Pennsylvania in 1946. Even in 1949, despite Watson's alarm at the progress shown by the UNIVAC (which had emerged from the ENIAC project), IBM was only just starting to design its first family of true, electronic computers; the 701. At that stage it was probably at least two years behind UNIVAC; which was subsequently, in 1950, sold to IBM's competitor, Remington Rand, which a year later also purchased the other group working in the field, Engineering Research Associates (ERA), to become the undisputed leader in computing. [2]

1930s–1940s: Holocaust era

Browning Automatic Rifle

M1 Carbine

During World War II, IBM manufactured the Browning Automatic Rifle and the M1 Carbine. Allied military forces widely utilized IBM's tabulating equipment for military accounting, logistics, and other War-related purposes. There was extensive use of IBM punch-card machines for calculations made at Los Alamos during the Manhattan Project for developing the first atomic bombs; this has been notably discussed by Richard Feynman in his book, Surely You're Joking, Mr. Feynman!. During the War IBM also built the Harvard Mark I for the U.S. Navy, the first large-scale automatic digital computer in the U.S.

In 2001, author Edwin Black published IBM and the Holocaust (ISBN 0-609-80899-0), a book documenting how IBM's New York headquarters and CEO Thomas J. Watson acted through its overseas subsidiaries to provide the Third Reich with punch card machines knowing that the machines could help the Nazis conquer Europe. The book extensively quoted from numerous IBM and governmental memos and letters chronicling how IBM in New York, IBM's Geneva office and Dehomag, its German subsidiary, were intimately involved in supporting Nazi oppression. Black also published IBM's own internal reports admitting that these machines made the Nazis much more efficient in their efforts. For example, IBM's European manager J.D. Schotte in 1940, after WWII broke out, reported back to IBM in New York “in Germany a campaign started for, what has been termed … ‘organization of the second front.’” He elaborated, “In military literature and in newspapers, the importance and necessity of having in all phases of life, behind the front, an organization which would remain intact and would function with ‘Blitzkrieg’ efficiency … was brought out. What we had been preaching in vain for years all at once began to be realized.” Schotte’s memo made clear that only at IBM’s initiative did the militarists comprehend what magic they could achieve with Hollerith automation. “Lectures on the punched card system were held by our representatives before officials of the general staff of various countries and, with our men, the study of possible applications was begun. ...progress was rather slow, and it was not until about eight or nine months ago [summer 1939] when conditions in Europe clearly indicated that a war was more or less unavoidable, that the matter became acute.”

Regarding the Romanian census for example, another IBM manager in Europe, W.C. Lier, wrote back to Thomas Watson's personal legal advisor in October 1941, “One of the first matters discussed with them, was that of the Romanian census and the machines destined for this business, which are actually blocked in Poland.” A day earlier, Lier had sent a more formal letter to Watson personally to assure: “On the occasion of my visit to Berlin, I also settled a few pending matters, such as the machines blocked in Poland [and] the Romanian census … I am addressing separate reports to the executives concerned in New York.” Lier also reported to IBM NY, “I obtained an interview with the Romanian Commercial Attaché who immediately endeavored to obtain the freeing of approximately seventeen machines at present blocked in Poland from the Devisenstelle [Foreign Currency Office] and the German Authorities … I have been given every assurance as to the satisfactory outcome of this demand.” The transfer of those machines to IBM’s Romanian subsidiary was then effected, and the Romanian census which located all Jews was undertaken and automated.

The 2003 documentary film The Corporation showed close-ups of several documents including IBM code sheets for concentration camps taken from the files of the National Archives.

Shortly after the war, IBM recovered the profit made from their Hollerith departments in the concentration camps, the printing of millions of punchcards used to keep track of the prisoners, the custom-built punchcard systems, and other business activities.

IBM asserts that, since their German subsidiary came under control by the Nazi authorities, they were not responsible for their role in the holocaust.. As in many cases, when the US entered the war, the Reich left in place the original IBM managers who continued their contacts via Geneva. The company has steadfastly avoided commenting in detail or explaining its conduct despite continuous inquiries from media and communal leaders.

1950s: Air Force and airline projects

In the 1950s, IBM became a chief contractor for developing computers for the United States Air Force's automated defense systems. Working on the SAGE anti-aircraft system, IBM gained access to crucial research being done at MIT, working on the first real-time, digital computer (which included many other advancements such as an integrated video display, magnetic core memory, light guns, the first effective algebraic computer language, analog-to-digital and digital-to-analog conversion techniques, digital data transmission over telephone lines, duplexing, multiprocessing, and networks). IBM built fifty-six SAGE computers at the price of $US30 million each, and at the peak of the project devoted more than 7,000 employees (20% of its then workforce) to the project. More valuable to the company in the long run than the profits, however, was the access to cutting-edge research into digital computers being done under military auspices. IBM neglected, however, to gain an even more dominant role in the nascent industry by allowing the RAND Corporation to take over the job of programming the new computers, because, according to one project participant (Robert P. Crago), "we couldn't imagine where we could absorb two thousand programmers at IBM when this job would be over some day, which shows how well we were understanding the future at that time" IBM would use its experience designing massive, integrated real-time networks with SAGE to design its SABRE airline reservation system, which met with much success.

1960s–1980s: Success

T-REX Corporate Center was originally one of IBM's research labs where the IBM PC was created.

IBM was the largest of the eight major computer companies (with UNIVAC, Burroughs, Scientific Data Systems, Control Data Corporation, General Electric, RCA and Honeywell) through most of the 1960s. People in this business would talk of "IBM and the seven dwarfs", given the much smaller size of the other companies or of their computer divisions (IBM produced approximately 70 % of all computers in 1964).

In 1970, GE sold most of its computer business to Honeywell and in 1971, RCA sold its computing division to Sperry Rand. With only Burroughs, UNIVAC, NCR, Control Data, and Honeywell producing mainframes, people then talked of "IBM and the BUNCH."

The original IBM PC (ca. 1981)

Most of those companies are now long gone as IBM competitors, except for Unisys, which is the result of multiple mergers that included UNIVAC and Burroughs. NCR and Honeywell dropped out of the general mainframe and mini sector and concentrated on lucrative niche markets, NCR's being cash registers (hence the name, National Cash Register), and Honeywell becoming the market leader in thermostats. General Electric remains one of the world's largest companies, but no longer operates in the computer market. The IBM computer, the IBM mainframe, that earned it its position in the market at that time is still growing today. It was originally known as the IBM System/360 and, in far more modern 64-bit form, is now known as the IBM System z9.

IBM 360

Only 18 months after work had started on the 1401 Tom Watson Jr set the labs to work on what was to become the 360 range of computers. This was perhaps IBM's biggest gamble, and one on a scale that has rarely been equalled by governments let alone by companies. It was a decision on the part of Tom Watson Jr that can only be described as truly courageous.

The initial R & D reportedly cost $500 million, but this was merely the tip of the iceberg. The total investment, including six new plants built around the world (with IBM switching from being the world's largest purchaser of electronic components to being the world's largest producer of these!), was estimated to have cost $5 billion over 4 years. It was indeed, at the time, the most costly privately financed programme in history; according to Tom Wise it cost more than the Manhattan (atomic bomb) project up to the time of Hiroshima. It eventually absorbed 2,000 programmers; when problems with the 'leading edge' software caused the delivery schedule to slip near disastrously. It also added 50,000 to IBM's total workforce world-wide.

Most adventurously of all it was incompatible with the preceding ranges of machines; at a time when the Honeywell 200 had been announced as the first 'plug compatible' (that is a machine that simulated, or 'copied', the 1400) to which the customers' existing programs could be easily transferred. It was the last time IBM could afford the luxury, for by the time of the launch of the subsequent 370 computer the customer base was too valuable, and too tightly locked in to the technology, for any such risk to be commercially viable.

It is understandable that Bob O Evans, then head of IBM Federal Systems Division, described it at the time to reporter Tom Wise as "You bet your company...". He did though add the much less publicised rider "....but it was a damn good risk and a lot less risk that it would have been to do anything else, or do nothing at all".

The gamble was of course a success, and made IBM virtually unassailable until the 1990s. The 360 range was announced on April 7th 1964 and the first 360/40 installed (albeit late; but with the production problems resolved) a year later in April 1965.

Although the project was Tom Watson Jr's greatest business achievement as reported outside IBM it was actually masterminded by his heir apparent, Vincent Learson, with Gene Amdahl (covering technical matters and later to found his own competitor to IBM), Albert Williams (who had the critical, and unenviable, task of keeping Wall Street happy with the massive investment programme), and Dick Watson (handling the production plan). [3]

Antitrust

IBM's success in the mid-1960s led to inquiries as to IBM antitrust violations by the U.S. Department of Justice, which filed a complaint for the case U.S. v. IBM in the United States District Court for the Southern District of New York, on January 17 1969. The suit alleged that IBM violated the Section 2 of the Sherman Act by monopolizing or attempting to monopolize the general purpose electronic digital computer system market, specifically computers designed primarily for business. Litigation continued until 1983, and had a significant impact on the company's practices. In 1973, IBM was ruled to have created a monopoly via its 1956 patent-sharing agreement with Sperry-Rand in the decision of Honeywell v. Sperry Rand, a decision that invalidated the patent on the ENIAC.

At the time of the launch of the 360 range IBM was under considerable competitive pressure from the large 6600 computers offered by Control Data. To counter this IBM announced, in August 1964, the 360/91, followed over succeeding months by a number of replacement machines, Only a few of these were ever produced (reportedly only 25 machines; and at a loss of $110 million) and in 1967 IBM stopped selling them. [4]

Control Data, understandably, was less than happy with this spoiling strategy (as they saw it) and claimed these were only 'fighting machines' (such as had been used by Watson Sr. at NCR) purely designed to keep Control Data out. It is possible that IBM simply outstripped its own capabilities and just couldn't meet its promises; even IBM does not usually contemplate a $110 million loss to fund a fighting machine. Although CDC instituted an anti-trust suit in December 1968 a settlement was reached in 1973. From IBM's point of view, perhaps the most important aspect of this settlement was that CDC was required to destroy its index to the millions of documents were about to be also used by the Department of Justice; and without which their case was to be much less powerful! [5]

The history of the CDC anti-trust suit, and the host of copycat suits by other companies that it spawned is worthy of a book in itself (and indeed there is at least one such book). Ultimately, though, it adds little to one's knowledge of IBM as a whole. Of the literally millions of pages (to be precise 66 million pages and more than 2,500 depositions) that have been archived in the case, those that have been abstracted—apparently almost exclusively by the prosecution in the case—were designed to show IBM in a bad light. More important, they were not representative of how IBM then normally did business; no matter how reprehensible, or even illegal, each individual item so reported might be considered to be. [6]

The most important impact, though, was that it possibly prompted the US Justice Department in January 1969 (at the very end of Johnson's presidency) to mount its own anti-trust suit. This one was not so easily settled. It consumed the efforts of hundreds of lawyers and probably hundreds of millions of dollars over a decade and a half; until it was eventually withdrawn in January 1982. It is arguable that some of the most dubious of its practices were stimulated by the various cases brought against it. [7]

The key effect of this, though, was that, for a critical decade and a half, IBM was under the constant scrutiny of the US government; which was ready to pounce on the slightest transgression, and almost continuously threatened break-up on the lines of Standard Oil earlier in the century. The ethical standards promoted by Tom Jr might possibly have faded into obscurity under the pressure of normal business priorities. Under the scrutiny of the US Justice Department, on the other hand, they became the very backbone of IBM. [8]

Thus, the anti-trust suit, which has usually been portrayed as a threat to IBM, was in perverse reality a critical element in its success; and indeed was essential to the development of the later company, founded more on philosophies than profits. The paradoxes did not cease with the departure of Thomas J Watson Sr and the succession of TV Learson and Frank Cary. [9]

It is perhaps not coincidental—though somewhat anti-intuitive—that soon after the pressure of 'anti-trust' had been removed in 1982 (compounded by the fact that there was no longer a member of the Watson dynasty in power to oversee its moral values) IBM's results (for 1985/1986) deteriorated. [10]

Competition in the 1970s/1980s

(Source: David Mercer's history of IBM

Some of the changes after the end of the Watson dynasty were forced upon IBM by external developments:

The Emergence of Leasing Companies

The early 1970s saw the arrival of computer leasing companies. These firms would purchase IBM equipment and then lease it to customers over a longer period—typically seven years—than that used by IBM to calculate its own rental charges. (IBM used a four-year payback period, though later, when IBM wanted to encourage customers to move from lease to purchase, this dropped to three years.) The difference allowed the leasing companies to undercut IBM and yet still make a substantial profit.

But the main problem for IBM was the loss of 'control' of its customer base. The leasing companies owed no loyalty to IBM and would happily sell anyone's equipment. Some of them even started to market their own lines of hardware.

The embarrassment for IBM was short-lived because of the pricing policies pursued by the most aggressive leasing companies. When IBM announced a new range of mainframes earlier than expected, the financial argument that was the foundation of the leasing companies' business evaporated, because customers wanted to switch immediately to the new IBM machines offering substnatially better price-performance. The moral for the leasing companies—and also for Lloyds, which had foolishly underwritten much of their business—was that borrowing long and lending short is a gamble.

The BUNCH of the 1960s, together with minicomputer vendors such as DEC and Hewlett Packard, had by then found niches which avoided direct competition with IBM's DP division. They offered products that IBM could not offer profitably and as such, they were tolerated, if not welcomed, by IBM.

Plug-Compatible Manufacturers

A more important problem was created in the 1970s by the 'PCMs'—plug-compatible manufacturers—working on the principle that if you can't beat IBM, then look like it but charge a lower price.

Initially, they competed in the area of peripherals—disk, tape and memory. IBM's response, to shake off the likes of Memorex and Telex, was to release its most advanced technology. This strategy held good for the best part of a decade, but as IBM started to lose its way in the second half of the 1980s, new competitiors were able to beat IBM in these areas, by offering better products.

The problem of plug-compatible mainframes was less susceptible to easy solution, and by the end of the 1970s, it became a major threat to IBM's dominance. Paradoxically the catalyst was one of those responsible for the success of the 360. After managing the technical side of the 360 launch, Gene Amdahl had left IBM. Unhappy with the hybrid technology IBM was using, he wanted to move on immediately to LSI (Large Scale Integration). An alternative view put forward by Taiyu Kobayashi is that Amdahl's conflict with IBM had been over the number of models in the 370 range. This conflict deepened when IBM closed his research facility. Whatever the reason, he set out in 1970 to beat IBM at its own game of launching leading-edge CPU technology.

He chose, however, not to produce a new architecture, but instead to duplicate the 360 architecture he knew so well, having designed large parts of it. The resulting plug-compatible CPUs operated just like IBM's ... except that the hardware inside was more advanced and cheaper. Worse from IBM's point of view was that, having been in contact with Fujitsu since 1969, Amdahl allowed that company to take a 24% share of his new company in 1972, when he was short of funding. Amdahl gave Fujitsu access to the technology it needed to catch up with the IBM System/370, which the Japanese had seen as a great competitive leap forward. In 1974 it was further agreed that Amdahl computers would be produced by Fujitsu in Japan, for sale in the US. In effect, Amdahl had lost control to Fujitsu.

The Japanese

Japan having been so successful in eliminating competition in shipbuilding, cars and consumer electronics, the government agency MITI announced in the 1970s that it was setting its sights on the computer market. Japan was particularly adept at developing current technology to a higher standard of quality and at a lower cost.

This announcement caused panic among much of the computer industry, but was received with relative equanimity by IBM. Japan's previous inability to develop successful software, coupled with the lack of widespread success by any of the Japanese mainframe manufacturers, had left IBM's senior management largely unworried. Indeed, IBM was still the clear market leader in Japan itself, and its S/370 launch had destabilised most of the Japanese computer industry.

After its success with the VLSI memory chip programme, which eventually gave Japan dominance in this field, MITI's programme overreached itself in the early 1980s, with its announcement that it was funding the development of the 'fifth generation' of computers, focussed on artificial intelligence.

Amdahl and Fujitsu

By 1979, Fujitsu had, with the assistance of technology acquired from Amdahl, overtaken IBM in the Japanese market for mainframes. It was even making a profit. The shock finally sank in at IBM, causing panic in much the same way that it had for the rest of the industry five years earlier. Japan now appeared potentially unbeatable to IBM's management. Fujitsu had gradually built on technology which was demonstrably superior to IBM's. The word, therefore, went out from IBM's headquarters in Armonk that Japan was its most feared competitor.

The message was clear and unambiguous. Other competitors could be tolerated but no mercy should be shown to any Japanese entries. IBM's international subsidiaries set up organisations to provide the muscle for competitive situations where PCM vendors were involved. As always, in the IBM of that time, such activities were to be kept spotlessly ethical, despite unfortunate internal references to the yellow peril.

Overreaction

Although Japan continues to make advances in the related field of robotics, its fifth-generation project, which got underway in 1982, turned out to be a failure, and was finally buried in 1992. The computing market proved to be resistant to the heavyweight tactics that Japan had applied so successfully to other markets. In particular, success in mainframes required software that was largely beyond the capabilities of a Japanese industry isolated by its own character set, Katakana.

Even though the Japanese challenge wasn't successful, its impact on IBM was massive, by way of its effect on the IBM culture. IBM announced the 303x range at prices that stunned competitors, but they were still matched by Amdahl. A religious war broke out between IBM's water-cooling and Amdahl's air-cooled techology.

IBM set out a new objective of becoming the lowest-cost producer, so that it could not be undercut by Japan. Production costs were to become the source of IBM's competitive advantage. Over the next five years, investments in new plants put spending on 360 in the shade. The argument ran that by building highly sophisticated and monumentally expensive plants, the barriers to entry for potential rivals would be prohibitive.

Speaking to investment analysts in November 1985, John Akers said "...our investments in research, development and engineering, plant property and equipment ... $32 billion over the past five years, an amount we expect to exceed in the next five years". However IBM's factories, rather than the less sophisticated but more flexible ones of its competitors, became obsolete far quicker than expected.

Lowest Cost Producer

The impact of the new commitment to be the 'lowest-cost producer' on the subsequent decline of IBM cannot be overestimated. Even IBM management admitted this was a strategic blunder, eventually rescinding the policy in 1987. But plants at the time took a number of years to commission, so the effects remained with IBM for more than ten years after the original decision. In particular, the relative inflexibility of these plants put IBM in a very weak position given that product life cycles had dropped to three years or less. Even under the capable management of John Opel, IBM was unable to match the speed with which new computing trends emerged. In John Akers' hands, the problem became a crisis.

The biggest casualty caused by the new commitment was one of IBM's most cherished beliefs: lowest-cost producer was in direct conflict with customer service. This conflict weakened IBM's responses, especially in later dealings with third parties. Indeed, it could be argued that the strength of the IBM's beliefs was best demonstrated by the problems which arose when IBM abandoned them.

Buying In

CEO John Opel introduced many of the ideas which, in the academic world, came under the title of the Make or Buy decision. Transaction Cost Economics, as described by the likes of Oliver Williamson, were also brought to bear on IBM's strategy. IBM had always purchased products and services which other comapnies produced more efficiently, although the IBM procurement process set very high standards for suppliers to meet.

In the 1980s this theme was boosted by the use of contract staff wherever possible. For example, all new secretaries were hired as contractors rather than as IBM employees.

The most critical make or buy decision which IBM has ever made came with its creation of the Personal Computer. In order to get the product designed as quickly as possible, the skunkworks set up by John Opel decided to buy in the processor from Intel and operating system from a tiny fledgling company called Microsoft.

Third Parties

Most important of all was the infatuation with 'Third Parties' as the solution to one of IBM's remaining cost problems; that of the 'channels of distribution'. Although I had by then signally failed in my attempts to explain the principles of mass marketing to IBM's UK board, they had at least recognised that the number of customers it needed to contact was about to expand dramatically; and were even more aware of the cost of the IBM sales-force in this context. Their thinking then went along the lines of 'if this can be passed to small businesses, which are not crippled by IBM's pay scales and, in particular, its very high overheads, the cost problem can be resolved'; and IBM would become the lowest cost producer in this area too. This distracted IBM's attention from the potential problems in its existing sales operation; and especially from the problem of extending coverage of the IBM customer service umbrella to much wider groups of end-users. The future of IBM, with its confident prediction of $100 billion by 1990, was to be in the hands of these cheap and efficient third parties—who were expected to account for more than half of IBM's sales in 1990.

The Switch to Outright Sale

Hiding many of the emerging problems were the other supposedly sound management practices which were introduced at the same time. Rental of machines, which was the previous basis of IBM's customer relationship, meant that the customer could easily return them and switch to a competitive offering.

Traditionally the answer had been that the sales-force held them in place by superb relationship marketing. John Opel, though, found a much simpler expedient; that of selling the machines outright. This was easily achieved—from 1982 on—by weighting the pricing structure in favour of this and having the ever faithful IBM sales-force reverse their previous practice; to remove the 'danger' of competitors replacing rental machines.

Unfortunately, in turn, this introduced a number of less obvious dangers. The first of these was that the 'selling campaign', which had previously been almost continuous since the sales personnel had to defend their exposed rental machines almost daily, was now reduced to a short burst of activity once every two or three years. The wise sales professionals still continued to maintain the relationships through these periods of 'non-activity'. The unwise—of whom there proved to be not a few—used the time for the more traditional sales activity of unproductively chasing new prospects. Indeed, in research conducted by Andersen Consulting in 1993, the most frequent and significant complaint made by IBM customers (37%) was that IBM had lost touch with the market-place; though this was closely followed by the 35% that said it had failed to keep up with its more nimble competitors. IBM's arrogance as also seen (by 11%) as a major cause of IBM's decline.

Although it might have seemed almost a change in 'accounting practices', this break with the hidden factor which had previously driven IBM's selling practices was to ultimately prove a central factor in IBM's later downfall. As Mills & Friesen pointed out, "Put bluntly, customers in the past had bought solutions and reliability from IBM; then IBM moved them down to buying boxes. Soon many customers concluded that they could get cheaper boxes elsewhere." 'Solutions' was not to feature again in IBM's strategy until the end of the 1990s!

Tom Watson Jr., in his 1990 autobiography, was also critical of this change "In spite of Frank's [Cary] success, he made a few decisions that caused me to wish I was still running the show. In particular I was alarmed when he and his eventual successor John Opel rapidly phased out the rental system, shifting billions of dollars worth of business to outright sales. They did this partly to free up capital that would otherwise have been tied up in rental machines. But it bothered me because rentals had been crucial to IBM's success. Rental contracts wedded us to our customers, gave us a powerful incentive to keep service top-notch, and made IBM stable and essentially depression proof."

The Cost to Customer Service

More insidiously, but as might have been predicted from its undermining by the new 'objectives', IBM's philosophy of 'customer service' now began to be subject to cost constraints, and was even charged for. The systems engineers, who had in practice been IBM's most powerful sales personnel, were withdrawn and 'sold' as a service; so that, in the 1990s, branches no longer had any staff undertaking this key role. The net result was that customer service deteriorated, even when IBM was providing it face to face. Of course, it became little more than a joke when 'IBM' dealer personnel were involved. Senior management began to talk of selling commodities (especially in terms of the PC), where the only deciding factor was to be the price. This was something that IBM had never considered in its three quarters of a century of history. It had always sold customer service, so that price alone was never a deciding factor. Even when price could not be avoided it had talked about 'price performance' (where the IBM service was a key element of the 'performance'). By the late 1980s, however, it was mostly talking just about price!

Selling the Family Silver

A less obvious effect of the outright sales policy, and one which—amazingly—IBM management did not see, was that IBM's revenue (and especially its profits) were artificially inflated in the short term. It was inevitably a one-time source of income which would be exhausted in two or three years, but IBM's management seemingly overlooked this key fact! At around the same time, IBM had changed its very conservative accounting practices, to adopt more conventional ones. In particular, investment in R & D was no longer paid as part of current costs, but was depreciated over future years. More controversially, the long-term 'capital value' of the leases (which IBM was now selling to replace the previous rental it had offered), were shown as income in the year in which the sale was made. Again, this inflated current income at the expense of the future.

These inflated figures set up unrealistic expectations of the future for the outside world; especially for IBM's stockholders, but also for all its other stakeholders. The figures which IBM senior management worked to showed annual growth rates of 20–30%; and IBM's sales-force was accordingly targeted with these growth rates year upon year. Unfortunately, the underlying figure (excluding the one-off, outright-sales) was actually below 10%. Though this was still a good performance, not dramatically different to IBM's historical record over previous years, it unfortunately did not justify the 100,000 extra staff who were then being hired.

This myopia had two dramatic consequences. The first was that it distorted the strategy which IBM management was setting. IBM's board really did believe that the growth was totally sustainable; and actually would lead to their dream of IBM becoming the first $100 billion per annum corporation. At the time this was central to IBM's strategy, and coloured everything it did. Not least it justified recruiting the extra 100,000 staff which it would, a decade later, have to lay off again (in the process abandoning IBM's core belief in lifetime employment). The second important consequence was that the IBM incorporated the figures as a basic premise for its success. Anything less than 20%, let alone an unthinkably low 10%, growth per annum was seen to be an unacceptable failure. The reality was that, after the one-off effect ran out, IBM's growth rate fell to normal. Indeed, as one might expect of such processes involving mortgaging the future, it actually fell to zero and even showed a slight decline. This was not seen as 'business as usual', as it might have been in any other organization, but was deemed to be a dismal failure; by the markets and especially by the frantic IBM senior management. The fact that this 'failure' (even when the figures returned to 10% growth in 1987) still represented a dramatic success in terms of other organisation's results was ignored, and the culture suffered its first (massive) loss of confidence—and began its long slide into oblivion.

Thus it was that John Opel fatally undermined the three core competences of IBM; enshrined in its three beliefs:

Respect for the Individual—although he would not have admitted it, probably not even to himself, John Opel's reign started the downward slide in its relations with its employees. Although they still remained in theory, many of the cherished personnel policies were discretely circumnavigated in practice. Since the time of Thomas J Watson the 'relationship with its employees' had always been its greatest strength. In the years running up to the 1980s this had, in particular, been reflected in the great flexibility, and indeed desire for change, which had enabled IBM to survive in such a turbulent industry. That flexibility was to be needed even more over the next decade, and yet IBM was having to rely upon an increasingly demoralised and alienated workforce.

Customer Service—again without any clear intent—but following the policy of 'lowest cost producer'—IBM's management had been seduced away from providing the highest levels of service to its customers. Worst of all, its most visible customer service, that provided to its millions of PC customers, had been sub-contracted to dealers most of whom were—correctly—seen by their (and hence IBM's) customers to be little more than charlatans.

Pursuit of Excellence—this was a more insidious process. The PC Independent Business Unit—then minuscule in IBM terms—had no choice, if it was to meet its very tight deadlines, but to source its hardware and software outside of IBM. This in effect gave Intel and Microsoft, in particular, their 'patents' on IBM's most visible developments. That the original managers in PC group negotiated such damaging agreements (especially with Microsoft which actually bought the operating system, 'on behalf of IBM', for only $75,000 and then made literally billions of dollars from it) is perhaps understandable. What is incredible is that over the following years (and indeed nearly a decade) no one in IBM developed a successful policy to counter the threat.

Many commentators would argue that IBM's problems were a result of the changes in the market—which the successful IBM of old was not able to handle (though it had by then been handling such changes, often against far greater odds, for the best part of three quarters of a century). The irony is that in reality it was IBM itself, very clearly led by John Opel, which had initiated (and then strongly promoted) these changes.

The Ratings Slip

The scale of the problem was made obvious when the business magazine Fortune released its 1987 poll (of 6,000 business executives) as to which was the top rated US corporation. Since the poll began IBM had been rated top overall every year. In 1987, as IBM's image began to be equated (by its own wish!) with that of its dealers, its position slumped to seventh. Worse still in 1988 it plummeted to thirty-second. By the beginning of the 1990s it was placed below 200 out of 300 runners. How were the mighty fallen! It took a decade, until 1998, for IBM to return to a number one spot in Fortune's Most Admired Companies—and even then only in the 'computer, office equipment' category.

Even so, the IBM's results in 1988 began to recover (after three bad years), and Akers seems to have relaxed. He returned to the tactic of urging IBM'ers to cut bureaucracy. Such tactics might have worked for Tom Watson Jr. But Akers did not have the same leverage, and certainly did not have the time, though he never quite seemed to realise this.

Risc and Risk

Sometime around the mid 1980s one of IBM's major developments in the laboratories was a range of RISC based minis, which eventually emerged, in one form, as the very successful RS/6000 workstations. There was, however, another project which was just as important—and was linked to this. This was the extension of an IBM version of UNIX to all its low end mainframes (the ones in the mid-range 4300 series of mainframes just as much as the AS400 and the, later, RS/6000; as well as the PC). This had reached an advanced stage of planning and indeed was only a few months away from launch. For example, detailed plans were in place to retrain the whole sales-force in UNIX—a major investment, of many weeks of training, for each of them. It was intended that this would answer the problem of incompatibility, since the reason for this was to be largely found in the software not the hardware. Using the new RISC-based processors, UNIX would have allowed most of the IBM range to talk to each other as did the DEC range. The only exception would have been largest processors; and they would have had interfaces added to their software to overcome even that problem. Converging five IBM systems into one, it would have added a great degree of 'open systems' capability, allowing in other (hardware and especially, IBM hoped, software companies to 'add value'), in much the same way that the PC had achieved the same results. In retrospect, it is possible to now see that it could have put IBM in a commanding position in terms of the actual market developments which, by then beyond its control, came about in the 1990s!

In practice, this major development never saw the light of day. It may be that there was an insuperable technical problem (and the abandonment of the related 'Fort Knox' hardware may, indeed, have had this effect), but that would have been unlikely to have held it back indefinitely. Some believe that this was the result of the clashes between the various laboratories involved. Perhaps more likely was that that IBM management had begun to realise the risks which 'open systems' brought with them! On the other hand, the open systems philosophy was by then unstoppable; and the abandonment only meant that IBM lost out twice. It finally had to accept the inevitable in the 1990s, but by then it was one of the laggards! It was another of the early changes of direction which later seemed almost to paralyse John Akers and his management team, for there was no obvious replacement—other than mundane enhancements to the power of the computers. This was something at which the competition were much better than IBM!

Fort Knox Tumbles

'Fort Knox' had (according to Paul Carroll) absorbed the effort of 4,000 developers over four years. This was intended to use RISC based architecture across the whole of the lower end of IBM's mainframe and mini range—to allow these, in combination with UNIX based software, to talk to each other (and to combat DEC's increasing strength with its VAX range). At the time this was one of IBM's most important, and best kept, secrets. Paul Carroll claims that it was, once more, one of IBM's failures in development. However, despite the assertions of Frank Soltis, there was no good 'technical' reason to abandon it. Perhaps it was again the 'Open Systems' approach (of the UNIX software more than the hardware) which caused IBM to rethink the project; and to start down its path of fatal indecision.

Price Cutting

The inertia of the consumer service factors (perhaps most cynically summed up in the phrase 'Nobody ever got fired for buying IBM') carried IBM forward through its 'successes' of the first half of the 1980s and the real problems didn't emerge until mid-decade. These problems came about because IBM management was starting to think in commodity pricing terms. On the other hand, it had long been almost a religious dogma at IBM that it did not cut prices, no matter how big the customer was.

The problem first surfaced when this dogma could no longer be sustained, as PC dealers took over as its point of contact with IBM's largest customers, rather than the small businesses the IBM strategy had called for them to support! Even then, there was no problem for the first two years as customers queued to buy the rationed PCs. The problem emerged when supply eventually exceeded demand. Most of the dealers had no answer other than to cut price. They had almost no technical expertise, and as little marketing talent, so their only answer was to be cheaper than everyone else. The result was predictable. Their discounts were ratcheted up—dealer vying with dealer to buy business at the lowest price. Eventually it soaked up almost all of the 35% discount or so that the larger dealers received; and sometimes went beyond, as dealers—whose mass marketing knowledge was even worse than IBM's—foolishly 'invested' in the future, with the lowest prices! In the process, IBM found that its prices (for the PC was still seen as being IBM's no matter who sold it) were widely, and very publicly, discounted. Indeed, they were massively discounted—far more deeply than even its most aggressive mainframe competitors had offered in the past! To counter the prices offered in this way, even its own mainframe sales-forces had to start discounting the PC.

Dealer Prices

The problem for IBM was not just that it could not fix its dealers' prices, that would have been illegal, but that it no longer had any justification for a price premium. The product being offered by its competitors was almost exactly the same—it was, after all, 'IBM compatible'. Worse still, the key justification which IBM had offered in terms of 'price performance', its fabled customer service, no longer applied. By common agreement PC dealer salesmen ranked somewhere below used-car dealers. In any case, what little service there was could be obtained, from exactly the same dealers, by IBM's compatible competitors!. So, even beyond what was offered in the way of discounts, IBM eventually had to lower its published prices; in an escalating race which saw published prices for comparable packages fall, just two years later in 1985, to around a quarter of the launch levels.

Mainframe Prices

IBM found it easy to dissociate itself from what it saw to be very special price-cutting on the PC, but its customers did not! The typical scenario which started to emerge in the mid 1980s saw the IBM account manager about to close a multi-million dollar sale on a mainframe being asked the usual question about discounts, and giving the usual answer; 'IBM never gives discounts'. The difference this time was that the customer no longer was willing to accept this, but pointed out that only a few weeks previously the same account manager had negotiated a 35% discount on a much lower value of PC business! Under these circumstances, the account manager was glad to get away with a 10% discount on the mainframe sale, and the branch were happy to give it—they desperately needed the business. Later, in one of the decentralisation moves which did work, the sales-force were able to give discounts without reference to senior management; and needless to say they did! It is difficult to believe that some of the larger discounts which then appeared, reportedly reaching as high as 50% on key business—coming very close to covering only IBM's variable costs—did not require the approval of senior management . That alone should have set off very loud alarm bells.

Voluntary Redundancies

The almost inevitable corollary of price cutting is cost cutting. IBM had always been efficient, apart perhaps from the intractable problem of administration overheads, and John Opel's 'lowest cost producer' drive had increasingly focused attention on this aspect of IBM's business. John Akers, making the mistake of so many other executives when faced with problems, went much further. Abandoning the salesman's charm which had been his main forte previously, he became a ruthless management activist. In 1986 he was forced to announce the first of the packages to slim back IBM from the level of 400,000 employees which had resulted from John Opel's earlier, misguided attempts to reach $100 billion turnover by 1990. By that time it had at last become obvious even to Armonk that this target was unreachable. The package (euphemistically called FAP, Financial Assistance Package), which included attractive early-retirement and severance offers, took 10,000 jobs out of the US workforce as IBM went into reverse. In the following year, 1987, according to Bob Heller 15,000 employees left IBM semi-voluntarily, while another 21,000 were re-deployed; 11,000 of them added to the sales force, although the evidence showed that despite their very expensive retraining, most of these also left the company over the next few years.

By 1991 John Akers was looking to cut staff to around 300,000, from the peak of 407,000—which had resulted from the misguided expansion of the early 1980s. This reduction was supposedly to be achieved without layoffs; thus maintaining the fiction of 'lifetime employment'. But, instead, pressure—sometimes amounting to unbearable stress—was being put on staff throughout the organization. They were, in effect, being forced to take one of the number of 'attractive' packages which were available to those who chose to leave. In theory, though, IBM's lifetime employment was still intact, even John Akers could not yet face that break. In a 'question and answer' section of the 1991 Annual Report, he said "We believe full employment reinforces pride and morale…", though he did add later in the report "…If further significant reductions are required, we will reassess full employment…". In rea ias all the staff knew - the philosophy was by now only a theoretical concept; the practice, of which they became increasingly afraid, was widespread redundancies!

The Cost of Cost Cutting

Cost-cutting became the major theme for IBM and, when the 1989 results crashed, the financial effects of this - together with a 'cushion' of extra funds for 1990 - were hidden amongst the restructuring costs; as another 10,000 jobs went. The process was discretely labeled - in some parts of IBM at least - as its 'Career Transition Programme' (the transition in question being that out of IBM!), but known irreverently - but accurately - to employees as 'cash to piss-off'. From 1986 to 1991, IBM cut its headcount by 21,000, closing 19 US plants; though, once more, nobody was actually fired - it resulted from normal attrition and the very attractive early-retirement plans.

It has been claimed that "…the policy of not firing people accounted for 80% of the debt IBM accumulated in the early 1990s." It is argued that, rather than maintaining morale, this resulted in nothing less than raw fear; and staff kept their heads down awaiting the redundancies to come!

Losses

When the major losses started, in 1991 (with a $47 million loss - the first in three quarters of a century - but admittedly after major restructuring charges), the cost-saving effort was redoubled. IBM until then had traditionally embraced a five point scale of individual performance, of which only the bottom '5' position (to which very few sank) was deemed 'unsatisfactory'. Akers then changed it to a four point scale. More important, he insisted on a bell shaped distribution - 10% of employees would be '1's, 40% would be '2's and another 40% '3's, and - crucially - 10% would be '4's (and hence deemed unsatisfactory). The fours would be further divided by making half of them 'Four-Check'; with the statement that if they did not improve rapidly they would be offered a severance package or simply fired. There might have been a great deal of sense in improving a system which clearly had its flaws, but - in the context of IBM's philosophies (especially that which incorporated 'lifetime employment' ) - it was almost the final blow.

IBM PC

The company hired Don Estridge at the IBM Entry Systems Division in Boca Raton, Florida. With a team known as "Chess," they built the IBM PC, released on August 11 1981. Although not cheap, at a base price of $US1,565 it was affordable for businesses — and it was business that purchased the PC. However it was not the corporate computer department that was responsible for this, for the PC was not seen as a proper computer. It was generally well-educated middle managers that saw the potential — once the revolutionary VisiCalc spreadsheet, the killer app, had been ported to the PC as the clone, Lotus 1-2-3. Reassured by the IBM name, they began buying the machines on their own budgets to help do the calculations they had learned at business school.

1990s—the failure of Portfolio Theory

On October 5, 1992, at the COMDEX computer expo, IBM announced the first Thinkpad laptop computer, the 700c. The computer, which then cost $US4,350, included a 25 MHz Intel 80486SL processor, a 10.4-inch active matrix display, removable 120 MB hard drive, 4 MB RAM (expandable to 16 MB) and a TrackPoint II pointing device.

On January 19 1993 IBM announced a $US4.97 billion loss for the 1992 financial year, which was at that time the largest single-year corporate loss in U.S. history. Since that loss, IBM has made major changes in its business activities, shifting its focus significantly away from components and hardware and towards software and services.

When assessing Gerstner's significance to IBM, we can put to one side, for the moment, the massive emigration of IBM talent during his watch. The amount of money taken out of the company and passed to shareholders, of which he had one of the largest individual holdings, can also be ignored in this discussion. If the company strategy had been right, the effect of such actions could have been minimised over the long term.

The irretrievable loss that Gerstner inflicted upon IBM was the loss of many key markets, such as PC operating systems. Much can be blamed on the portfolio management technique that Gerstner brought with him from McKinsey.

Until the 1980s, IBM provided the customer with everything in computing that they wanted: computer, storage, support services, programming, finance etc. IBM probably made a profit on each of these items, but it didn't bother too much with unpicking the individual elements to discover the financial truth. The customer wanted the whole package together, and if IBM meant customer service, it was best if it provided the whole bundle.

In the early 1990s, before Gerstner arrived, IBM was already unpicking these elements, and setting up separate profit centres. Thus for the first time there was a Software Business, even though it depended almost entirely on colleagues in the System Business to generate sales of, for example, the OS/400 operating system.

Gerstner pursued the logic of these separate divisions further. Using portfolio theory, he would let each division run for a couple more years, and if it didn't make a profit or show revenue growth, the division would see its problem offerings discarded. It seemed that Gerstner was prepared to let anything go, if it didn't meet the tests within the timescales. The problem was that IBM was competing against highly focussed competitors for whom the survival of their rival products meant life or death for the company.

Thus, for example, discarding OS/2 was a difficult decision for IBM, but it didn't mean the death of the company or even any executive resignations. Over at Microsoft, the success of Windows 95 and Windows NT, the principal rivals of OS/2, was regarded by Gates and Ballmer as a matter of life and death for the company. No matter how many years it took, Microsoft was determined Windows 95 and NT would win.

Gerstner's willingness to discard unprofitable elements meant that holes developed in IBM's offerings. If, for example, IBM no longer offered maintenance on the hardware it sold, how could it guarantee the overall customer experience?

The problem with Gerstner was that, despite his obvious intelligence, his understanding of IBM's product range was very poor, even at the strategic level. For example, in his autobiography he states that by the mid-1990s, it was obvious that client-server architectures would be supplanted by the Web.

This begs the question, that if Gerstner knew that then, why did he subsequently buy Lotus, whose main product (Notes) was the archetypal client-server product? It is hard to think of another exceptionally bright CEO in any industry who knew less about his or her own company's products. Unaware of the linkages between products and their interdependencies, IBM dropped out of several markets that with hindsight, it shouldn't have.

2000 - 2006

In 2002, IBM strengthened its business advisory capabilities by acquiring the consulting arm of professional services firm PricewaterhouseCoopers. The company is increasingly focused on business solution-driven consulting, services and software, with emphasis also on high-value chips and hardware technologies; as of 2005 it employs about 195,000 technical professionals. That total includes about 350 Distinguished Engineers and 60 IBM Fellows, its most-senior engineers.

A chart showing IBM's revenue and net income, 1980–2005.

A chart showing IBM's patent history, 1993–2005.

In 2002, IBM announced the beginning of a $US10 billion program to research and implement the infrastructure technology necessary to be able to provide supercomputer-level resources "on demand" to all businesses as a metered utility. The program has since then been implemented.

IBM has steadily increased its patent portfolio since the early 1990s, which is valuable for cross-licensing with other companies. In every year from 1993 to 2005, IBM has been granted significantly more U.S. patents than any other company. The thirteen-year period has resulted in over 31,000 patents for which IBM is the primary assignee. In 2003, IBM earned 3415 patents, breaking the US record for patents in a single year.

Protection of the company's intellectual property has grown into a business in its own right, generating over $10 billion dollars to the bottom line for the company during this period. A 2003 Forbes article quotes Paul Horn, head of IBM Research, saying that IBM has generated $1 billion in profit by licensing intellectual property.

In 2004, IBM announced the proposed sale of its PC business to Chinese computer maker Lenovo Group, which is partially owned by the Chinese government, for $US650 million in cash and $US600 million in Lenovo stock. The deal was approved by the Committee on Foreign Investment in the United States in March 2005, and completed in May 2005. IBM will have a 19% stake in Lenovo, which will move its headquarters to New York State and appoint an IBM executive as its chief executive officer. The company will retain the right to use certain IBM brand names for an initial period of five years. As a result of the purchase, Lenovo inherited a product line that featured the ThinkPad, a line of laptops that had been one of IBM's most successful products.

Of late, IBM has shifted much of its focus to the provision of business consulting & re-engineering services from its hardware & technology focus. The new IBM has enhanced global delivery capabilities in consulting, software and technology based process services - and this change is reflected in its top-line.

On June 20 2006, IBM and Georgia Institute of Technology jointly announced a new record in silicon-based chip speed at 500GHz. This was done by freezing the chip to -451 °F (Template:FahrenheitToCelsius °C) using liquid helium and is not comparable to CPU speed. The chip operated at about 350GHz at room temperature.