3   THE EMERGENCE OF IBM AND THE CULTURE OF THINK

This business has a future for your sons and your grandsons and your great-grandsons, because it is going on forever. Nothing in the world will ever stop it.

—THOMAS J. WATSON, 19261

ON FEBRUARY 15, 1924, the Wall Street Journal published a small article, buried on page 3, announcing that the “International Business Machines Corp. has been incorporated under the laws of New York to take over business and assets of Computing-Tabulating-Recording Co. Capitalization is the same as for Computing-Tabulating-Recording Co. and shares will be exchangeable.” The article went on to explain that the name change “has been made because of increasing growth of the company business and the development of additional lines of business devices. No change is involved in the names of the subsidiary companies, the Tabulating Machine Co., International Time Recording Co. of New York and Dayton Scale Co.”2 That is how the world learned about the existence of IBM, unless they worked in Canada, where employees had known their employer as IBM since 1917.

When Watson was remaking C-T-R, he already thought the company’s name was awkward and uninspiring, but changing it was the last thing he had on his mind; too many other issues required his urgent attention. Why he decided to rename the firm adds to our understanding of IBM’s early evolution. On February 13, 1924, Watson wrote a letter addressed to all employees to announce the change, explaining: “Our new name is particularly adaptable and suitable to our business, in view of the fact of our increasing growth, the consistent development of additions to our line, and our products covering such a wide range in the field of business machinery.” He concluded, “We are confident that this change in name will be beneficial to the business, and that the members of our field organization will find it of direct value in introducing our company and the products which we manufacture and distribute.”3

The event was also important to C-T-R’s customers and vendors. They may not have read the press announcement but received a letter announcing the change, and their salesmen rushed forward to explain how wonderful it would be. Then there was this little story in the Wall Street Journal three months after IBM’s incorporation that may not have been true but hints that everything comes down to execution. The article began by stating that some people were confused by the change in the old name, which it called “unwieldy,” and maybe this really did happen:

The other day an uptown merchant called up a friend in Wall Street and inquired if he had ever heard of a new concern called the International Business Machines Corp.

“We have a big order from them,” he said, “and I am trying to check up on their credit rating. My partner has just gone downtown to demand a balance sheet from them.” “Well,” said the broker, “in their last balance sheet they showed $800,000 cash and about $8,000,000 current assets. Their position seems pretty good. How big was the order?”

“About $100,” said the merchant. “Wait a minute, I want to head off my partner.”4

Implicit in that story was the real business problem of retaining the positive reputation of C-T-R among customers, investors, employees, and the media while leveraging its new name to build business momentum. Watson and his colleagues spent the 1920s and 1930s creating a brand image that reflected their positive view and plans for the future but that also translated into transactions, profits, growth, and prestige. They battled ignorance of what their products could do, invented new ones, hired people, expanded operations, overcame the worldwide economic tragedy of the Great Depression, and endured an antitrust challenge, the first of several that would threaten to scar IBM. As historian Robert Fitzgerald has pointed out, American firms were early adopters of various marketing and market segmentation practices, many of which were in evidence at IBM. The company did not go after mass markets—its products were too specialized—but the IBM case demonstrates that practices used by large manufacturing and distribution firms were common at IBM, such as market segmentation, use of advertising, and education of customers.5

The name change ultimately signaled a larger transformation under way. While it occurred halfway through the years discussed in this chapter, it had been a long time coming, as the Canadian example suggested. Whether the firm should have had a different new name than IBM was less important than that Thomas Watson felt it was time to declare a grander purpose for the company. The tone of his comments, the nature of the company’s communications, and the way its staff interacted with the media and with customers evolved almost as a step change after the adoption of the new name. He was declaring that IBM wanted to become a major player in both the office appliance industry and as a major corporation.

There was a great deal going on at IBM besides a change in name, which makes it possible to view the announcement as a signal to the company’s employees and its market. In chapter 2, we focused largely on Watson the leader and what he did during his first decade at C-T-R. In this chapter, we continue to discuss his role but with additional emphasis on the actions of the firm at large. This chapter begins to tell the wider story of IBM as an institution, although Watson continued to dominate the further formulation of its ways of operating. We focus on developments in the 1920s, but because many developments and trends continued through the 1930s, they are explained here, such as the arc of technological and product development, which is best understood as a whole over the entire period. Chapter 4 takes up events of the Great Depression of the 1930s, a time that threatened the existence of so many corporations. That IBM came out of it financially stronger, which was a unique experience for an American company, was part of a broader pattern that historian Mira Wilkins neatly termed “the conquest of markets.”6

This chapter begins with a discussion of the transition from C-T-R to IBM, a case of marrying desired image to corporate strategy. Changing the name of a company is never straightforward, being rare and fraught with complications, so IBM’s experience needs consideration. We discuss the unfolding of IBM’s strategy and business operations in the 1920s, continuing a discussion begun in chapter 2. We identify the effects of Watson’s sales culture at work. Since he tied strong sales operations to technological innovations, we examine the evolution of IBM’s products. Finally, we add a fuller account of the role of customers, because part of Watson’s way of doing things was to nurture a tripartite interaction of customers, sales, and product development. As chapter 5, on the World War II years, demonstrates, the combination of strong sales, a modern product line, and a data processing ecosystem facilitated IBM’s substantial expansion during the war years.

HOW C-T-R BECAME IBM

As Watson expanded the business outside the United States, he set up independent companies owned by C-T-R, as required by local laws. He established one in Great Britain, another in Germany, still others in France and Italy, and so forth. He experimented with different names, even using his own family name, but usually with the concept of the International Business Machines Corporation. To use the word of Watson biographer Kevin Maney, that name had “heft,” on par with names such as General Motors and United States Steel.7 In 1917, Watson used the IBM name when he set up Canadian operations and when the company moved into Latin America. It fit nicely with his worldview.8

Despite some resistance from board members, and of course from Charles Flint, the board gave Watson permission to move forward and so, on February 5, 1924, he filed the necessary paperwork to change the company’s name. He then expended considerable energy selling the idea to employees with speeches, a new company newsletter, and other correspondence. Managers began taking down C-T-R signs in front of their branch offices, replacing them with IBM ones. As part of his branding initiatives, Watson introduced a new logo. Figure 3.1 shows one of the old C-T-R logos alongside the new IBM one. The new world logo remained in use until 1947, when a plain “IBM” replaced the global image. The 1924 logo captured the essence of Watson’s business being big and global. Using black for the logo connoted that IBM was serious, yet fashionably modern. It would be decades before IBM would become known as “Big Blue.”

Figure 3.1

C-T-R/IBM had only two logos until the end of the 1940s. The IBM logo did not become a highly recognized image until the late 1940s, although it was respected earlier by its customers. Photo courtesy of IBM Corporate Archives.

IBM’S STRATEGY FOR GROWTH IN THE ROARING 1920S

IBM grew in terms of bricks and mortar, number of people, and products built, transported to customers, and leased. IBM’s growth was physical, yet the company grew as fast as many digital companies of the early twenty-first century. Executives of many companies, and business professors, have long been fascinated with IBM because of its ability to grow profitably, normally thinking only of the mainframe computer years, but the 1920s and 1930s provide an even more remarkable story of growth and prosperity. The numbers tell an outstanding story, suggesting the complexity of the issues IBM faced, which demonstrated the results of sound execution.

In 1923, the year before Watson changed the company’s name, C-T-R generated $11 million in revenue, and by 1939 that had tripled to $38 million; profits had quadrupled. The company began 1924 with 3,161 employees, nearly a third outside the United States. In 1939, IBM employed over 11,000 people, still with a third outside the United States (see table 3.1). It therefore remained very much an American company, but one that paid attention to expanding its business worldwide. The salient fact is that its workforce had grown in size by three and a half times in just 16 years! Almost as impressive as any of these numbers is that IBM dominated its chosen markets. Eighty percent of all tabulating machines around the world came from IBM, as did nearly all the cards these machines consumed, both providing a continuous, safe revenue flow to the company.9

Table 3.1

IBM’s revenue, earnings, and employee population, 1914–1940, select years (revenue and earnings in millions of dollars)

Year

Revenue

Net earnings

Worldwide employee population

1914

4

1

1,346

1918

9

1

3,127

1923

11

2

3,161

1928

15

5

5,102

1933

17

6

8,202

1938

34

9

11,046

1940

45

9

12,656

Source: Emerson W. Pugh, Building IBM: Shaping an Industry and Its Technology (Cambridge, MA: MIT Press, 1995), 323.

There was plenty of credit to go around for this success. Dominating all major decisions was Watson himself. His engineers executed well, too, as we will see later, while his sales force grew and matured over time. Luck helped in the 1920s, when the economies of the industrialized world expanded; bad luck followed in the 1930s, when they shrank; but more good fortune came when Watson refused to let the Great Depression slow him down and the U.S. government launched Social Security and other economic and welfare programs.

For the kinds of products IBM sold, it needed a healthy economy, especially in the United States. The start of the 1920s was terrible, but the economy shifted to prosperity, increasing GDP and per capita income by 30 percent during that decade. New industries emerged and rapidly expanded. These included radio, movies, automobiles, appliances, construction, electricity, and chemicals, a long list. The 1920 U.S. census reported for the first time that more people lived in cities than in the countryside. New and old companies needed office appliances, including everything IBM was selling except more cheese slicers. IBM rode the wave, with revenue rising from $10.2 million in 1922 to $20.3 million in 1931. Because profits also grew, making Watson a wealthy man, while his senior executives profited handsomely from their IBM stock. Another generation of management would enjoy the same experience in the 1960s, but in the 1920s, as in the 1960s, it was not a free ride; results had to be earned.

IBM had its rivals to contend with, such as Burroughs, NCR, and Powers (later as Remington Rand), which also enjoyed expanding demand for their products, although not to the same extent as IBM. Often ignored by historians was Underwood Elliott Fisher, created in 1927 to sell typewriters and business machines.10 It had an excellent sales force, comparable to IBM’s, and a popular product line. The Powers Accounting Machine Corporation was folded into Remington Rand that year as part of another merger effort like Flint’s of 1911. NCR did marginally well, but by the second half of the 1920s it was not diversifying sufficiently beyond the cash register business.

IBM made a few acquisitions to strengthen its technology portfolio and to keep potential rivals at bay. IBM retained the business model of leasing hardware, selling the cards needed to feed these machines, and offering tailored terms. The sales force designed accounting processes for their customers that used equipment to lower operating costs and improve efficiencies. Throughout the 1920s, Watson and his sales executives kept telling everyone at IBM that they had barely tapped the opportunity. Leasing and card revenues were sufficient to fund the manufacture of more equipment, support R&D, and pay dividends.

Card sales became increasingly important as a cash cow for IBM. By the mid-1930s, it sold over 4 billion cards annually at $1.05 for a box of 1,000. This was such a lucrative business that IBM’s first two antitrust lawsuits with the U.S. government focused largely on these pieces of stiff paper. ITR’s customers also used cards, as every worker needed to punch in and out daily, each using one card per week. Card sales also strongly supported ITR’s financial performance. Between ITR and Tabulating Machines, demand for cards kept growing, only slowing a tad during the Depression years. Cards were also more profitable than leasing equipment. By the late 1930s, card sales made up 15 percent of the company’s revenue.

To turn over every rock to look for revenue, IBM expanded its services business, offering smaller enterprises the chance to do their accounting work at an IBM data center. That offering introduced smaller enterprises to data processing in an almost ad hoc manner and in the process introduced them to IBM’s technology, so that when they eventually leased their own equipment they understood and were already dependent on IBM’s equipment and procedures. That expanding IBM ecosystem rested on its technology. For new and smaller users, the spread of IBM’s equipment became more formularized through its Business Services Department, established in 1925 and put under the control of Frederick W. Nichol (1892–1955). Pushing hardware and services were twin features of IBM’s strategy that reinforced the ecosystem. In 1932, IBM reorganized its services business into the Tabulating Machine Services Bureau, better known as the IBM Service Bureau Division, the organization used to expand services in the United States and later worldwide.11 This business can be viewed as quintessential IT outsourcing, today known as cloud computing. IBM’s decision to expand service bureau work during the Great Depression proved prescient, because some customers could not afford to set up or expand their existing data processing operations but wanted to save on their own operating costs by using tabulating equipment. IBM’s offering became a convenient option. Government and commercial users took advantage of IBM’s data centers, located in major cities. Uses were as diverse as in one’s own data centers: accounting, inventory control, human resource management and record keeping, marketing surveys, and sales management.12

As IBM increased its focus on the tabulating market, its interest in scales and cheese slicers waned. Those products were sold with no follow-on sales for supplies, so there was no ongoing sales opportunity. In 1923, Watson sold Dayton Scale to Hobart Manufacturing, a solidly performing company, in exchange for 100,000 shares of common stock. These shares would spin off between $125,000 and $150,000 in revenue each year. This transaction took off the table a troubled business that had absorbed too much of management’s attention over the years, although IBM kept making some scales for industry, a business tucked into a new organization within the firm, International Industrial Scales and Counting Devices. IBM acquired other small rivals in this industrial market in the 1920s and 1930s. By the end of the 1930s, the experiment was over; IBM had become the world’s largest supplier of punch card tabulating equipment.

One feature of life at IBM was its seemingly constant reorganizations to align and optimize products, people, and customers. After its return to prosperity in the early 1920s, IBM reorganized the old Hollerith business with a broader scope, renaming it International Electric Tabulating and Accounting Machines. Fortunately, C-T-R’s tabulating equipment had originally been electric, whereas Powers did not make the transition away from manual machines until the 1920s. Electrification made C-T-R’s, and later IBM’s, equipment easier to use and usually faster, and it appeared more modern, more advanced. Even more product innovations arrived during the Great Depression.

Senior managers paid considerable attention to innovations or, to use economic historian Deirdre Nansen McCloskey’s more precise term for such matters, betterment, which speaks more to the purpose of changes in a technology or product while thinking tactically about the matter.13 In October 1927, for example, Watson called all his senior executives into a conference room in New York, including Otto Braitmayer, James W. Bryce, Frederick W. Nichol, and others, “to talk about the future.” Analyzing IBM’s current circumstances, he wrote off the scale division as nearly useless, and opined that ITR was properly managed and doing fine, though its opportunities seemed limited. But now Tabulating was increasingly becoming the company’s major source of cash and profits. Watson wanted his machines to dominate their market, not simply do well. He told his senior managers that this could be done by controlling all patents related to tabulating, regardless of whether they were held by IBM or others, and that is why IBM was willing to buy out other companies. It was essential that IBM have the most effective, uniquely designed, and patented cards in the business—IBM’s with square holes, Powers’s with round ones—so that rivals could not displace its equipment. IBM would need to expand manufacturing. Watson wanted IBM to have the best R&D and sales operations in the office appliance industry. None of these messages was new to the men assembled in that room.

What seemed new, however, was the intensity of IBM’s growing commitment to this strategy and to the message that the others embrace it as enthusiastically as Watson did. It was vintage Watson: “There isn’t any limit for the tabulating business for many years to come.”14 After a great year in 1927, in which IBM brought in $4 million in profit and was set to increase that by an additional $1 million in 1928, Watson again brought his close associates into the same conference room. This time he lashed out at them for not having done even better. He wanted everyone to push harder, to “never feel satisfied.”15

So IBM in the 1920s was all about growth, opening more offices, and reaching out to new markets in Europe, Latin America, and even Asia. It was also about Watson being completely in charge of “his” company. Everyone of any consequence in the firm reported to him. In the late 1920s, 13 companies made up IBM, and Watson had regional and division executives, the R&D community, and manufacturing reporting to him. Watson biographer Kevin Maney looked at an early organization chart of the 13 companies, which included the IBM holding firm, and saw a list of 14 executives who also worked across the 13 subsidiaries, plus the board of directors, now made up of handpicked men. Maney calculated that there were nearly 200 people reporting to Watson. He called it “a web, with Watson as the spider.”16 Watson was behaving very much like a company founder who, by growing up with the company as it expanded, kept up with what so many other people were doing and, like many start-up founders, had difficulty surrendering authority to his direct subordinates. Nevertheless, he expected them to take initiatives to fulfill his vision. The two staff meetings just discussed were typical.17 In fairness to Watson, while unable to surrender authority, he generously shared praise and credit for successes. Many of his speeches were devoted to praising “the men of The IBM.”

His views became IBM’s policies. Management overwhelmingly did not like unions, so they provided enough benefits to keep workers in Endicott, and later elsewhere, from organizing. Watson fed his vanity through speechmaking and communicating with his employees, but his management team also knew they had to keep their employees content because they were crucial to the expansion of IBM. Especially in Endicott, management did not want them trying to get jobs at the other big firm in town, the Endicott Johnson Shoe Company, run by George F. Johnson (1857–1948), a highly progressive entrepreneur who Watson came to admire and learn from, which usually had better benefits. IBM raised salaries, built employees a golf course, and introduced two weeks of paid vacation, all in the mid-1930s at the height of the Great Depression. Watson believed IBM could afford to construct a large training building in Endicott to educate factory workers, engineers, salesmen, managers, and customers. His optimism was not fully shared by the company’s board of directors, but he built the facility anyway.18 His benevolence mirrored policies implemented by other U.S. corporations.19

IBM’s strategy involved creating a corporate culture that kept everyone informed about the company’s intentions, values, expected behaviors, objectives (targets) individuals and their organizations were expected to achieve, and employees’ actions outside the walls of IBM. That goes far in explaining the many presentations Watson and various levels of managers made to groups of employees, and why Watson spoke to every sales training class held in Endicott and to other classes and groups of customers. In 1934, IBM published an 886-page anthology packed with hundreds of Watson’s missives, adding to IBM’s information ecosystem.20 Each division had its own processes to inform employees, distribute executive memoranda, host departmental or division-wide meetings, and publish newsletters. In 1924, Corporate launched a company-wide internal publication, Business Machines, which became IBM’s primary channel for printed communications to all employees. It reported on Watson’s talks, new products, employees, personnel practices, events in the company, and market conditions.

IBM’s information ecosystem was designed to reinforce its culture of THINK, one based on the use of data and practical thoughtfulness. Recall, too, that by the early 1920s, field engineers and salesmen were relying on their own manuals for product information. The company added weekly reports on sales up the chain of command and did similarly at the factories. Each country had its own board of directors, which was expected to communicate quarterly with New York headquarters in a prescribed manner. Country general managers spoke increasingly with headquarters staff and more frequently wrote reports addressed to New York about daily operations. Executives were expected to be active in this information ecosystem. By the mid-1920s, there had emerged an information ecosystem involving employees and customers and an effective corporate culture supporting that community. The company’s culture facilitated the functioning of the ecosystem, with many activities supporting this community, reflecting company values and desired behaviors. The interlacing of IBM’s “world”—ecosystem—and its culture is a theme that has permeated the history of IBM.

The 1920s became a good time for Watson to build his company, its collection of customers, and its culture. Businessmen were considered “hot,” fashionable. In January 1925, President Calvin Coolidge famously said, “The chief business of the American people is business.” American companies basked in this glow. IBMers were caught up in it, too. Engineers came up with new equipment, and salesmen found new customers and new uses for IBM’s machines. A biographer said Watson still “loved coming up with data processing ideas, which he then dropped like hand grenades onto his beloved engineering teams,” who had to respond by trying to implement them.21 His ideas ranged from coming up with machines for use by bank tellers to developing a heating lamp to keep his machines from breaking down during humid weather.

No history of IBM’s culture would be complete without discussing “Songs of The IBM.” By the mid-1920s and extending deep into the 1940s, even the 1950s in some parts of the company, IBMers sang songs about Watson, other executives, and IBM. Harry Evans, a spunky, short, high-energy executive with a sense of humor, wrote many of them. In 1925, he published the first IBM Songbook, and after that, his creations were sung at sales conventions, branch offices, and factory gatherings, often as tributes to Watson on the occasion of one of his visits. All relied on existing melodies, such as Over There and Auld Lang Syne. If the company had an anthem, it was Ever Onward, with a chorus that shouted out “Ever Onward—Ever Onward! That’s the spirit that has brought us fame!”22 Other companies sang songs as part of how they linked to their employees. Other methods IBM and others deployed included softball teams, picnics, and even a company band at Endicott that lasted for decades.23

IBM’S TECHNOLOGICAL EVOLUTION

No customer of high-tech industrial equipment will long tolerate products that fail to help reduce operating costs or improve productivity. Customers aggressively pressured vendors to continuously introduce new information-handling tools that were cheaper, faster, and more reliable. Underlying a great deal of that innovative process, usually hidden in the shadows, was a century-long guerrilla war involving the unexciting, yet important, topic of patents. Many of IBM’s customers accepted the company’s combination of “leading edge” developments but also its habit of lagging other innovators. Both patterns existed in pre–World War II IBM. Watson drove his engineers and lawyers hard, without which other vendors could have bypassed IBM because customers always had options for what kind of tabulating equipment to use or to use other types of data processing equipment, neither of which required IBM’s punch cards. These included calculators that were increasingly sophisticated, adding machines, billing equipment, and electric typewriters.

IBM’s legacy of lagging and innovating characterized its behavior from Hollerith’s day to the present but became a purposeful intertwined strategy during the 1920s. The strategy turned on two issues: when an innovation had matured enough to be converted into a product or new feature that could be protected by patents and when the market was ready to accept an innovation. IBM’s engineers did not hesitate to patent myriad innovations, but not all of them surfaced as new features and products. IBM vigorously collected patents and dueled with rivals over control of innovations. Historians of IBM’s early history devoted considerable attention to this issue, focusing more on the actions of the company to protect its innovations with patents than on the timing of new product introductions, the latter always coming along after an innovation first existed.24

Inside IBM, not all heroes were executives and salesmen. Engineers, representing the development and manufacture of products, and salesmen, with their requests for new products reflecting market timing, needed to collaborate. Engineers developed products that enabled corporations and government agencies to grow, function, and leverage data. In the 1920s and 1930s, IBM’s engineers made up a small group, fewer than several dozen souls, but they were productive. Speaking to the 100 Percent Club in January 1928, Watson recalled how quickly engineers had developed new products in months rather than the more customary years, citing events that year and ending with, “I cannot say enough for our engineers and for the wonderful work they have done in producing these machines in such a short time.”25 In a speech delivered at Princeton University that October, he made the critical point that “eighty percent of the business of our company today is done on products that have been developed since 1914.”26

Watson’s praise was not hype. Less than a dozen engineers fundamentally transformed IBM’s product line. They worked on the second generation of punch card equipment, which replaced Hollerith’s original creations. The first generation served IBM’s customers from the late 1880s to the start of the 1920s. They were due for replacement, and Watson and his technical managers assembled a remarkable new team to do that, bringing together talented engineers already in the firm and hiring others. Hollerith hired one of the first who Watson liked, Eugene A. Ford (1866–1948). Ford began his career as an inventor of typewriters, bringing to the company experience with keyboards. He developed a numerical keypunch machine for the Tabulating Machine Company in 1899 and became an employee of the company in 1905. In 1914, Watson put him in charge of a development center located on Sixth Avenue in Manhattan. His small team electrified parts of the tabulating product line. They developed a verifier, used to confirm that cards had the same data punched in them as those typed by a keypunch operator, as an aid to improve data accuracy.

Watson hired Clair D. Lake (1888–1958), who had worked in the automotive industry. He had only attended school through the eighth grade and later a two-year industrial training program but had demonstrated a masterful understanding of engineering principles. A third member of the team, also brought in by Watson, was Fred M. Carroll (1869–1961), who had worked as an inventor at NCR. Watson asked the three to invent a card printer. Meanwhile, when Ford left the company for a while, Lake took charge of tabulator development. In 1917, C-T-R brought on a fourth engineer, James Wares Bryce (1880–1949). Bryce started out as a draftsman and designer at the turn of the century, studied mechanical engineering at New York City College for three years, and then moved to Endicott in the Time Recording Division.

Lake went to work on improving the tabulator product line, simplifying machine operation and reliability. Just after World War I, he introduced the concept of plug boards, similar to those that already existed in the telephone industry, to simplify maintenance. Benjamin M. Durfree (1897–1980), who joined CTR in 1917 to do tabulator maintenance and repair, worked with him to make such improvements. All their efforts came together with the introduction of the Type I printing tabulator in 1921, which customers had wanted for years. It worked so well that in 1924 Durfree went to Europe to promote and install these machines. Meanwhile, Lake began recruiting a new generation of engineers. In 1917, he brought in Ralph E. Page (1896–1991), making him the chief draftsman working on the new tabulator.27 George F. Daly (1903–1983) came in 1920, also as a draftsman working with Lake. Meanwhile Watson promoted Lake to Endicott’s superintendent of tabulating machines.

IBM accumulated relevant patents largely by acquiring companies and their employees. An important acquisition was J. Royden Pierce (1877–1933), a highly regarded inventor in the office appliance industry, and his little engineering company in New York City. Pierce brought along at least two employees. He had completed his degree in engineering at the Stevens Institute of Technology in Hoboken, New Jersey, making him the first worker at IBM to have achieved that distinction. Pierce joined C-T-R in 1922 as an employee, signing over his patents to the company, patents he had developed while he built a tabulator for use in the life insurance industry. Later he developed test-scoring equipment, another that could subtract numbers, and yet another to do multiplication.

So now IBM had two development centers: one in Manhattan with Bryce, Pierce, and a few other employees, and a second in Endicott with Lake and Carroll. Lake worked on tabulators, while Carroll improved the machinery needed to manufacture cards better and more quickly. In 1923, Ford returned to IBM and worked on a new sorter, introduced in 1925 as the Type 80 Sorter. It was also known as the “horizontal sorter,” because all the card pockets were of the same height. The machine was faster, easier to use, and incrementally improved Hollerith’s original design. This was a critical innovation, because sorting subtotals of accounting information was crucial, requiring that cards be tabulated in the correct order. Tabulators in the 1920s and 1930s were also equipped with “accumulators,” which held subtotals. The company introduced incremental changes in its product lineup throughout the 1920s and 1930s. One, for example, was a “summary punch” that made it possible to print subtotals, crucial in competing against mechanical bookkeeping systems from companies like Burroughs and more specialized office appliance firms. That product became possible because of the dialogue that took place between sales and engineering, with the needs of IBM’s sales staff and customers driving engineering’s R&D agenda. These kinds of collaborations became fodder for Watson to use in speeches when opining on the virtues of THINK.

An important development involving the cards themselves came from this small team. All during the 1920s, customers had asked IBM to find a way to put more data on its cards. Lake and Bryce thought that with some changes to the equipment they could squeeze more and narrower columns on a card by punching square holes rather than round ones. They had to develop a card that had more than the 45 columns available at that time. They created a stronger card with 80 columns, which IBM immediately patented and introduced. IBM’s engineers modified all its equipment to handle the new cards. Over time, that new card became known as the “IBM Card,” and it remained in use for the next half century. IBM’s main card rival—Remington Rand—also had a 45-column card, and it responded with a new card that had two rows of 45 columns. Customers loved IBM’s card, and so it became the standard for the industry and the continuing source of much revenue for IBM. The basic changes that appeared in so many products in the 1930s grew out of the work engineers initiated in the 1920s. While the story of IBM’s innovations of the 1930s might later seem to make sense to tell, they were a logical outcome of earlier work.

Recall how IBM introduced innovations in its products but often slowly. How does one explain this apparent contradiction? IBM’s behavior was purposeful, as it chose a lower-risk strategy. The alternative ran the risk of introducing a product that failed; IBM would experience that in the 1980s with minicomputers and some models of the Personal Computer. For most of its history, IBM reacted to requests by customers for new functions rather than these functions being the brainchild of an employee. Innovations were incremental but numerous and were rarely revolutionary. Even the 80-column card grew out of experiences with cards over the previous quarter century. Innovations relied on existing engineering principles. Innovations often took years to develop, but because they arrived continuously, IBM and its customers enjoyed the fruits of new features and functions, and learned from each of them to inform future developments. Incremental introduction of innovations built up momentum, going far toward explaining why so many products came out in the 1930s and not in the 1920s, when a great deal of R&D and patent collecting went on.

In the 1930s, IBM completed its conversion of the entire tabulator product line to 80-column cards while introducing a continuous flow of new devices that could perform more accounting functions, such as alphabetical accounting, and a multiplying punch (later called the Type 600), both introduced in 1931 in the middle of the Great Depression. The Type 600 was replaced two years later with the Type 601, to provide what accountants called “footing” and “cross-footing,” used to sum tables of numbers by column and row. Now users could subtract and add, making possible even more accounting functions than before.

Improvements out of the two engineering shops made it possible to sell more tabulating applications, so the two U.S. engineering centers merged into a new building in Endicott in 1933, with the Manhattan workshop being closed down. Known as the North Street Laboratory, the Endicott facility housed seven engineers, including Ford, Lake, and Carroll, a number of draftsmen, one patent attorney, various other people to run the office, space to handle special machinery, one metallurgist, and an industrial designer. The building included a blueprint room, a machine shop, and storage for parts. Now all could collaborate more frequently and work faster—essential in order to respond to market demands in the 1930s and 1940s and to meet the immediate needs that soon emerged from the U.S. Social Security Administration and other U.S. and state government agencies.

So, what products could a customer obtain from IBM by the early to mid-1930s? An IBM system included three groups of leased devices. First there was the punch, also known as a keypunch, which looked like a typewriter and was used to punch holes in cards fed to the keyboard. Second was the sorter, used to shuffle decks by columns, alphabetically, or from one number to another. The cards would fall into a bin in the order sorted. Then there was the processor itself, known as the tabulator and, increasingly in the 1930s, as an accounting machine. That machine printed reports or documents (such as checks) and performed basic mathematics (adding, subtracting, multiplying). In larger installations, one might also have a verifier attached to keypunch machines to ensure that cards punched were accurate. Since inputting—punching cards—was labor intensive and involved a great deal of information, it was normal to have many keypunch machines and staff (largely women) to type all those little square holes into cards.

Keypunch machines could be located all over a company or agency, sending boxes of punched cards to other keypunch operators for processing using verifiers, after which they would be sorted. These bigger machines were housed at a central site, often called the “IBM Room,” later a data processing center. That practice of having operating equipment in their own center served as the model for data processing centers after the arrival of digital computing. Many of these centers continued to use keypunch machines for decades, while all other tabulating functions became embedded in computers.

Responding to customers’ suggestions became crucial to IBM’s success. By the late 1920s, some customers had been using tabulating equipment for over a quarter century, so they were both dependent on it and deeply knowledgeable about what they needed next. In 1940, a young engineer, John C. McPherson (1908–1999), became head of a new organization, the Future Demands Department. A graduate of Princeton University (class of ’29), he joined IBM in 1930 as an engineer and later served as a salesman in the railroad industry for three years. Being both an engineer and a salesman, he brought to his new assignment a perfect match of technology/engineering and sales/customer knowledge. The timing of his new assignment was propitious, because as the number of customers increased, their needs became more diverse, often clustering around industry-specific requirements. For example, banks needed accounting tools that differed from those used in, say, inventory management processing in railroads or manufacturing. McPherson’s appointment reflected the culmination of a process long under development at IBM.

Product requirements had to be defined, coordinated, and worked on by the growing engineering community. Customers wanted to reduce the amount of “setup” time needed to run different sets of calculations (known as applications) on their equipment, which called for using removable plug boards. Table 3.2 lists many of the improvements made during the interwar period in response to their requirements. Market needs had to be collected by the sales branch offices and then transmitted to the engineering community. In the 1920s, this was often done in an ad hoc manner, in the form of letters requesting changes, but by 1940 a more formal manner had finally been adopted. It was not enough to write to the “Old Man” for help. IBM and McPherson created a product development process that proved effective for decades. Formalizing that and linking it to market and customer intelligence gathering, although not unique to IBM, was still novel in the 1930s. GE, Burroughs, and a few other major firms had moved in that direction in the 1920s, but since most slowed product development in the 1930s, IBM’s lunge forward during the Great Depression proved fortuitous.

Table 3.2

Tabulating technology evolution, 1920s–1930s

C-T-R

Six hardware products in 1920

Automatic sorting machine

Two keypunch machines

Quick set gang punch

Hand set gang punch

Tabulators

45-column cards

IBM

32 hardware products in 1933

Seventeen keypunch machines and one verifier (several models for various sizes of cards)

Five sorters (several models for various sizes of cards)

Nine tabulators (in addition to several models within each type)

45-column cards

80-column cards

IBM’s Technological Innovations

1930

Numeric interpreter to print on cards

1931

Automatic multiplying punch and duplicating summary punch

1932

Alphabetic printing tabulating device; verifier with automatic feed and eject

1933

Alphabetic keypunch (Type 3) and reproducing punch

1934

Automatic carriage for printing tabulator

1936

Alphabetic verifier, alphabetic interpreter, collator

1938

Transfer posting machine, reproducing gang summary punch

1939

Mark sensing equipment

Remington Rand, IBM’s leading rival in this market, and Powers-Samas in Europe introduced an equal number of innovations in this period. See James W. Cortada, Before the Computer: IBM, NCR, Burroughs, and Remington Rand and the Industry They Created, 1865–1956 (Princeton, NJ: Princeton University Press, 1993), 111.

IBM’s crowning technological achievement came in 1934, when its salesmen were able to sell the Type 405 accounting machine (figure 3.2). One group of IBM engineers later wrote that it was “an outstanding success” that “remained the flagship of IBM’s product line until after World War II.”28 It could perform many accounting functions electronically and led IBM to rename its tabulating business the Electric Accounting Machine (EAM) Division.

Figure 3.2

The IBM Type 405 was sold as a “system,” not as one product, and was introduced just as IBM’s customers’ volume and complexity of work expanded dramatically. Photo courtesy of IBM Corporate Archives.

In the process of introducing so many new machines and making hundreds of incremental changes, IBM dominated the world’s patent portfolio for tabulating equipment. Bryce, Carroll, Ford, Lake, and Pierce accumulated over 500 patents in their lifetimes; Bryce personally had 244. But they were not alone. C-T-R/IBM cumulatively acquired over 1,200 patents between 1911 and 1940, and another 900 by 1950.29 The bulk of IBM’s patents came between 1926 and 1940, when 969 were granted, but it took a long time for ideas and then patents to emerge as products, a process that Watson and his executive team understood. Hollerith and the executives in ITR and the Scales Division did not fully appreciate that reality, which helps explain why C-T-R obtained only 11 patents between 1911 and 1915. In the next five years, the company garnered 65 patents, adding 143 more between 1921 and the end of 1925.30 Comparing that performance to the products listed in table 3.2, we see the results of Watson’s faith in the future of the data processing market. Add in the constant feedback from the field and customers, and one can see why IBM’s investments were not as bold or irrational as many thought in the 1930s. Rather, IBMers were listening to the market, responding prudently in thoughtful ways.

CUSTOMERS AND HOW THEY USED IBM’S EQUIPMENT

Who were IBM’s customers in the 1920s, and what were they doing with these machines? The 1920s and 1930s was a period when the world was still discovering the power of such machines as they were being invented and sold to them. IBM’s customers, and others that used Remington Rand’s products and those of various office appliance vendors, became the earliest users of computers. Why? Because, by the 1950s, they already understood the power of data processing. The majority of IBM’s business still came from the United States, but it had expanded across Western Europe and to a lesser extent into Latin America and Asia. Because their heritage was largely punch card equipment, knowing what they did about this technology in the half century preceding the arrival of the computer is crucial in understanding the relationship customers had with IBM over the rest of the century.

The process of becoming a customer of IBM, Remington Rand, or Burroughs was not simple. First, an organization had to collect, analyze, and use large amounts of information. Such businesses included big insurance companies, banks, railroads, and large data collecting agencies, for example a census bureau or a national pension department. It helped if one had thousands (or millions) of people to track, such as an automotive manufacturer or an army. Second, that largeness meant that without using IBM’s expensive and complicated equipment, one had to collect and use information manually or rely on simple desktop calculators. Either option often proved more expensive than deploying IBM’s machines. For scientific applications, calculating by hand might be too tedious, slow, or impossible, such as in defining rotations of planets. Third, the equipment had to perform functions that made sense to a customer, which is why accounting, census taking, and inventory control attracted early users.31

Fourth, a customer had to have a capable staff. These included keypunch operators, others who could use peripheral equipment and “program” tabulators to do their bidding, and still others to manage the flow of cards coming in and out of a data processing center and to deliver reports to where they were needed according to some fixed schedule. Such managerial requirements meant accounts had to employ one or more managers able to work with IBM to acquire equipment, help shape how it would be used, and decide who would be responsible for day-to-day interactions with salesmen, maintenance people, and those collecting bills for IBM. If a customer did not have one or more of these prerequisites, IBM still had an answer for them: its services bureau.

In the 1920s, dozens of companies and government agencies that were C-T-R’s customers before 1924 grew into hundreds of organizations, and by the late 1940s several thousand.32 As a corporation became big by the standards of its day, an IBM salesman would show up, in the beginning more as an expert on routine accounting and the machines but later also knowledgeable about the data processing needs unique to an industry. While census taking and accounting were the first wave of applications fairly regularized across organizations, every customer was nonetheless unique, so IBMers tailored use of their equipment to their user’s operations. That took months of study, design, and negotiation, so acquiring a new account, or a new user within an existing account, took time, sometimes more than a year or two. This also meant customers were reluctant to swap out office machine vendors.

When the installation of IBM equipment was completed and the equipment was functioning as needed, customers became dependent on IBM’s products for decades, binding them to their supplier for equipment and cards. Over time, IBMers sought new ways for customers to use their equipment and, as IBM’s revenues demonstrate, customers did more data processing, evolving into information-intensive enterprises. They learned from each other and from IBMers how best to use this equipment. How IBM did something therefore often became how their customers did it, too, whether in processing data in accounting, inventory control, or sales. Through individual adoptions of data processing, IBMers influenced how companies and agencies worked. Watson captured the essence of that influence in 1930 when he said, “But as fast as one company after another demonstrated that the tabulating machine was a success, it became possible to apply the same use throughout a whole industry.”33

Arming accounts with more information was not the only by-product of IBM’s activities. At least as important was the expansion in the number of people working in offices, particularly women, as typists, clerks, secretaries, and keypunch operators. Women who entered this world began as keypunch operators, a job almost as easy to learn as using a typewriter, which they had been doing since the 1880s. Some keypunch operators moved on to operate IBM’s data processing machines. IBM was willing to train a customer’s personnel to operate its equipment at no cost, in the process creating a small army of workers who knew how to use only IBM products. They could be counted on to favor these products over those of an unfamiliar competitor. By the end of the 1930s, one could even see in business such formal titles as the “IBM Department” or “IBM operators,” much like the term “xeroxing” meant photocopying in the 1970s. IBM intentionally promoted the branding of professions after its name.

Users of IBM’s equipment clearly understood what they were doing. Saving on operating costs proved important, as one user reaffirmed: “Tabulating our labor costs by hand would, of course, be impractical as it would require at least six more people,” while the IBM machine was “paying for itself every week.”34 Others focused on efficiency, keeping “a running check on exactly where they stand and to maintain a more nearly continuous contact with customers and prospects by frequent dunning and soliciting.”35 Data accuracy was always critical, described by one observer as “the elimination of errors,” because IBM’s machine “saves confusion and time wasted in checking besides promoting better organization and efficient management.”36 Users appropriated new functions as they became available, such as using tabulators to do payroll accounting and printing paychecks, tracking these transactions, reducing errors, and resolving disputes with workers. A typical example is Ford Motor using the equipment to pay its thousands of workers.

Over time, scientists and engineers began to use IBM’s products. In the 1930s, it became possible to use these machines for complex mathematical calculations, such as in astronomy or in mechanized testing and grading. Watson personally led the development of a strong relationship with nearby Columbia University, a tie that lasted for decades. The Statistical Bureau at Columbia pioneered uses of punch cards in science and mathematics, providing IBM with positive press and goodwill among academics and engineers. One astronomer in particular, Wallace John Eckert (1902–1971), was an effective evangelist. During the 1930s, he persuaded Watson to donate equipment to his laboratory, donations that IBM’s leader concluded would burnish his company’s reputation while creating new demand for his machines. By 1937, Eckert had established the Thomas J. Watson Astronomical Computing Bureau, involving Columbia University, IBM, and the American Astronomical Society.37 Working with scientists, IBM’s engineers, other scientists and engineers, and customers learned how to calculate statistics using tabulators, an application of increasing importance to government agencies.

Scientific applications were different from commercial ones, requiring different thinking and uses. Commercial applications were relatively simple to perform but required processing of a great amount of data, such as millions of social security or inventory records. Scientific and engineering applications required calculations more complex than the four basic mathematical functions and far less data. For the former, handling volumes proved important, while for the latter it was speed of calculation.38

Becoming a new customer and staying with IBM seemed easy because of the study IBM would conduct to determine how best to use its equipment and because it would help do the financial analysis to justify its cost. IBM’s practice continued throughout the century, but it was always the hardest step for both sides to take. In addition, by leasing equipment, customers did not need to risk precious capital on some new technology, and machines could be replaced whenever IBM brought out a more attractive product. By the end of the 1920s, there existed a growing pool of workers experienced in using IBM equipment. Both customers and IBM shared a stable cash flow, for customers as a budgeted expense and for IBM as predictable income. Best practices and the use of experienced, trained personnel spread from one account to another, creating the IBM-centric ecosystem that proved so successful that some 80 percent of all tabulating users in the world knew it.

IBM used its growing presence to branch out into other office appliances, most notably electric typewriters. IBM’s engineers knew a great deal about keyboards and electricity, two technologies central to IBM’s products for decades. IBM’s senior management appreciated the strategic business nature of the two linked together. When contemplating new opportunities in the 1920s and 1930s, they explored developing specialized banking systems, developing test equipment for education, and bringing out their own line of typewriters. To add to IBM’s ability to produce electronic keyboards for its tabulating systems, in 1933 the firm acquired Electromatic Typewriters, located in Rochester, New York. It was rare among the many typewriter firms in that it produced electrical machines and held patents on them. While electric typewriters had existed since around 1900, secretaries preferred manual ones, and ultimately they, not data processing managers, were the customers for such machines. Unfortunately, however, Electromatic’s machines cost about $250 each, roughly twice that of a manual typewriter, and so they did not sell well, even after becoming part of IBM. At only about 6,000 units a year, its sales accounted for only 5 percent of the market for typewriters.39 Its real asset was the keyboard technology, which IBM’s engineers appropriated for their keypunch machines. Expertise in keyboards made it possible to produce typewriters and other accounting machine keyboards for decades. Engineers even toyed with the idea of connecting typewriter-like devices to telephone lines to create a teletype tool that a half century later we called e-mail. Because of technological and patent issues, they set aside this idea.

IBM introduced its first commercially successful electric typewriter in 1935. By the end of World War II, its salesmen had gained significant momentum in that market at the expense of Underwood (the largest vendor), Royal, and Remington Rand. In addition to selling well, IBM’s machines proved superior and highly reliable, leading a generation of users to remain loyal to it and subsequent products.40 IBM developed a sales force specializing in selling typewriters. Within IBM they were considered the best individual salesmen in “pitching” a product. For decades, they cultivated typists and secretaries, making it both highly successful and useful to place the IBM logo in so many visible corners of a customer’s enterprise beyond the data processing center, such as immediately outside a manager’s or executive’s office door, where his secretary sat.

TABULATING THE EVOLUTION OF IBM AND THE CULTURE OF THINK

For decades, the central challenge for C-T-R and then for IBM was not just efficient management to improve productivity—Flint’s primary reason for bringing in Watson—or to improve the product line and the quality of sales operations, although both were essential to the company’s long-term prosperity. The central challenge for the company was to create a broader market for its products than envisioned by Hollerith. That effort required IBM to educate potential customers and salesmen about its products and how best to apply them at the individual account level. Watson never quite explained it bluntly, but he hinted at it when he repeatedly argued that only a tiny percentage of the world used tabulating equipment. The closest he got to the central issue was when he used the term “data processing,” which he combined with exhortations to “think,” “learn,” and “take action.”

A considerable amount of economic activity involves persuading people to take action to buy something, what one scholar called “sweet talking.”41 As products and activities became more complex or expensive, persuasion as an economic activity became more sophisticated. Selling industrial equipment clearly falls into this category. IBM’s selling efforts—persuasion—exemplified that activity, extending the scope and sophistication of such actions far beyond what Hollerith had, and what ITR or the scale business needed. IBM’s customers had to be persuaded what to do. The importance of that function in an economy, or for IBM, accounts for the enormous investment of effort, time, and thought in the development and use of IBM’s sales organization.

IBM had to wrap all its actions and successful “sweet talking” into a set of contractual terms and support services that made it possible for new users to embrace punch card equipment. Even those who had already taken the plunge still needed to learn about new uses of the equipment and how to properly adapt the equipment to their needs. Then IBM’s challenge was to keep them committed to using its products. Watson instinctively understood that the way his company could prosper and grow was to lead in the development of the market for IBM’s machines. As that happened, it was the sales organization that was present at the creation of demand for products at the individual account level. IBM essentially had few rivals. Powers had to do the same thing for this tiny niche market, but it did not do so as effectively as IBM. IBM expanded its market at approximately the rate that it could increase manufacturing, sales, and services.

That ability to keep market size and capacity synchronized resulted in two consequences. First, IBM overwhelmingly continued to dominate a larger market. It did this so well that in the 1910s it already “owned” over 90 percent of the market, and as it expanded, thanks to Powers (later Remington Rand) and a French company called Machines Bull in the 1930s, it actually lost market share, which dropped to nearly 80 percent, where it remained until the end of the punch card era, but in a market far larger than the one inaugurated by Hollerith. The British Empire market represented an exception. BTM and Powers shared a great deal of the market, but IBM split it in half with those two rivals. By controlling the market’s growth, IBM dominated it, executing normal sales, services, and product development and manufacturing.

Second, because IBM was de facto the only supplier of this class of technology for most companies and governments, it became the hub, the center of the tabulating data processing ecosystem, the “go to” place for goods and information, and, later, management practices, for organizations automating their largest information processing activities. Other office appliance vendors did not enjoy that level of influence (they thought of it as control) over their clients, because customers could swap out calculators, billing machines, adding machines, cash registers, and typewriters at will. There were many suppliers and users that also had a half century of experience with these technologies. Customers never became as dependent on one vendor as they did when it came to large data processing projects. Tabulating equipment made larger enterprises possible.

Being able to persuade customers to concur with IBM’s suggested actions and proposals remained a high priority within the firm’s corporate culture. THINK as an idea was about forming an opinion about an issue in a rational manner and being able to persuade others of a point of view. It came back to the notion of “sweet talking,” which makes sense if the rationale is persuasive both to IBM’s colleagues and its customers. The act of persuasion and the desired decision by IBMers and customers to take action became the key economic task of the company. All else followed: product development, advertising, training, installation, maintenance, and use.42

That symbiosis of growing the market as quickly as the company could serve it and becoming the center of that market (its ecosystem) made it possible for IBM to grow and prosper. When computers came along, a new generation of IBM executives understood the need to hop onto the new technology that otherwise would replace their tabulating machines. To his credit, Thomas J. Watson Sr. created the data processing ecosystem that made it possible for IBM to do well in a relatively tiny but lucrative market. In each decade between 1910 and 1940, there were only a few hundred accounts, a few thousand machines, and only a few tens of millions of dollars, but IBM learned how to optimize all these moving parts to make the business profitable. When demand for data processing surged during World War II, IBM essentially had to ride the wave, running its operations as they had evolved them in the 1930s. Not until after World War II did a new generation of managers confront different circumstances that made Watson’s older generation, his inner circle, and now aging tabulating engineers outmoded and in need of replacement.

While most of the world suffered during a nearly decade-long economic depression, IBM had a different experience, one that led it to become what Watson envisioned back in 1914. How that happened is the story told next.

Notes