19   HARD TIMES, AGAIN, AND ANOTHER TRANSFORMATION

We no longer expect to deliver $20 operating earnings per share in 2015.

—IBM CFO MARTIN SCHROETER, OCTOBER 2014

ON JANUARY 19, 2017, IBM’s CFO explained that for the nineteenth consecutive quarter of Ginni Rometty’s stewardship, IBM’s revenue shrank, and the company’s hardware business had declined even faster. Yet many analysts and stockholders applauded because the source of its revenue showed IBM was transforming away from low-profit businesses to others with larger profit potential (software, analytics, and artificial intelligence), notably cloud computing, which was even better when combined with artificial intelligence. It was essential for a major IT provider to be participating in this area. Since 2012, IBM’s revenue has continued to shrink, and the company has laid off thousands of employees while thousands of others have entered the firm. Employee morale declined, too. Only half the company’s employees would recommend to their friends that they work at IBM.1 In earlier decades, thousands worked to get their children into Big Blue. The media and customers complained loudly in 2013 that things were not going well at IBM. Employees had been making the same point for over a decade. In their defense, senior executives argued that the company was shifting away from hardware to services and software, such as cloud computing and analytics, later into security and mobility, counseling patience and understanding. Analysts applauded these moves but complained about execution: hardware sales were dropping faster than revenues were increasing from the new “imperatives.” They wanted results in the current quarter, while executives understood that big changes took years to accomplish. The question was could they shift—in their words “transform”—quickly enough to keep up, then set the pace in changing the buying practices of their customers and changing information technology?

Industry observers complained that IBM was late to the party, particularly to cloud computing. Roadmaps and their targets predominated for too long as the company’s operational mantra. Finally, in the fall of 2014, the company dropped its roadmap strategies, a half decade after complaints from thousands of employees that they perverted IBM’s ability to succeed, diverted attention from focusing on customers and product innovations, and encouraged excessive cost cutting and large layoffs. IBM was under siege, playing catch-up with changing competitors, emerging technologies, and hence what customers were buying, while mad ex-employees were becoming customers, possibly future poison pills for the company. The timing of changes to new technological regimes is also of great interest to business historians, and not just to economic historians interested in technological innovations and their impact. While much work has been done in studying, say, R&D practices, such as by Margaret B. W. Graham, David C. Mowery, and Richard Nelson, to mention some of the most obvious, fewer studies have addressed how individual senior management teams respond, so this chapter expands the aperture on the issue of change management.2

By 2014, even the culture and practices of the Gerstner period were barely recognizable at IBM, with its sales culture seriously undermined and senior leaders relying too much on financial and accounting strategies as they plodded toward new sources of revenue. They bailed water out of the blue boat to maintain high cash flows from service contracts and to sustain profits and the value of the company’s stock. Senior leaders focused on quarter-to-quarter management of the business out of the Corporate, services, and software parts of the business. This chapter will have much to say on each of these points, drawing on statements by executives and employees. It seemed like the financial and human resources staffs were dictating what everyone had to do. Layoffs were mandated, even for fully engaged service personnel billing their time, all to drive down expenses. In Canada and the United States in 2016, employees were reporting layoffs of salespeople responsible for promoting Watson computing, one of IBM’s focus areas. Growing numbers of employees seemed to express a collective quizzical look. IBM’s history reminds us that Big Blue had overcome adversities before, that it was frequently late to a new technological revelry, something Watson Jr. reminded in his memoirs, but once in a market, it was often a dominant player. Each time, it was not clear whether the company would survive. Too many observers prematurely wrote off the firm. Yet, here we are with a company that generated $80 billion in annual revenue, so while IBM was again in trouble, it was also doing well. IBM’s old culture was gone; a new one that employees, customers, the media, and analysts could see would sustain the company had yet to solidify. The CEO since 2012, Virginia M. “Ginni” Rometty (b.1957), clearly had as rough a time on her watch as John Akers had on his or Thomas Watson Jr. had. In Rometty’s case, she enjoyed less employee support than Watson Jr., or even Akers.3

Understanding recent events at IBM is as important as ever, because IBM continues to play an extraordinary role and draws the attention of scholars, employees, and other business observers. It is involved in shaping national economies and the future of robotics, medicine, artificial intelligence in augmenting human behavior, and computing itself. We face the problem of the company’s future remaining unclear. There is a history lesson to help in our assessment: IBM is so big, powerful, and able to accumulate billions of dollars that it cannot be written off as some technological has-been. To kill the old computer dinosaur would still require an army of dragon slayers. The old dinosaur had evolved into a more advanced form that required different hunters, while naysayers were also transforming into new competitors, such as Google and Amazon.com; new partners, such as Apple; and new investors, such as the Sage of Omaha, Warren Buffett (b. 1930), and his company, Berkshire Hathaway, until he chose to reduce his investment in 2017.4

This chapter focuses on recent events at IBM, even though the story told lacks the clarity made possible by the hindsight of history. IBM remains a company filled with bright, highly educated, experienced employees, even if too many of them are disgruntled, yet cheered on by the market, stock analysts, investors, and customers. What are we to make of this situation? It begins with a failed strategy that took too long to rectify but—and this is an essential quality of IBM—in time management changed and faced the necessity to acknowledge, retire, and replace old practices. I describe the company’s strategy to the point where it needed replacement, assess an enormous problem within the firm of laying off employees and hiring new ones, and assess these actions within the context of IBM’s old Basic Beliefs and the exigencies of current realities. Then I describe IBM’s current business strategy. It is essential to wrap up with a discussion of artificial intelligence—IBM’s Watson system—as that is the central technological and marketing focus of the firm today. While the events described are still unfolding as this book goes to press, the description is done in the language of the historian. Those familiar with the work of Alfred D. Chandler Jr., Louis Galambos, and Geoffrey Jones, among historians, and such industry watchers as Gideon Gartner in the 1970s and 1980s and Kevin Kelly in the 1990s and the next decade, will see that they shaped my discussion of post-9/11 IBM. We see a demonstration of what Martin Campbell-Kelly and Daniel D. Garcia-Swartz demonstrated in their history of the computer industry, that “the persistence of companies is, in an overworked but fundamentally true phrase of our era, constant reinvention,” and only those that “responded to rapidly evolving technologies and markets have been able to survive and thrive.”5 IBM was transforming in a century that continued to see improvements in the lives of people that were made possible by the continued development of information, knowledge, and wealth, which historian of optimism and progress Steven Pinker based on the call “dare to understand!” IBM’s offerings continued to be at the center of that 250-year pursuit of the Enlightenment by humankind.6 That is the central theme of IBM’s history during the twenty-first century. The reader will notice that the tone and depth of detail change in this chapter, as we shift to what management intends, with less emphasis on what IBM did. That is because of the immediacy of unfolding events.

FROM ROADMAP 2015 TO ROADKILL 2015

In the first decade of the new millennium, the company’s most senior executives deemphasized Gerstner’s focus on the needs of customers, concentrating greater attention on delivering “shareholder value” as their prime mission, trumpeting less the needs of other stakeholders, notably employees. While many other large, publicly traded U.S. corporations embraced similar priorities, IBM seemed to be an extreme case, the poster child for financial engineering. That is not to say IBM turned its back on its customers—far from it. It seemed to employees that the firm lowered its appreciation for their contributions as well. Nothing so dramatically demonstrated the shift from earlier priorities than Roadmap 2010, and more so Roadmap 2015. The first one worked, the second one caused declines in revenue. IBM’s productivity shrank and its market shares shrank for nearly half a decade in such key segments as cloud and mobile computing. The magnitude of the effects of the two Roadmap strategies on IBM’s performance is comparable to the effects of those initiated by John Opel in the early 1980s. In the case of the roadmaps, the remnants of Watson-era culture deteriorated sufficiently to again raise questions about the long-term viability of the company. So, how did the roadmap strategy come to IBM?

Recall that in 2006 Sam Palmisano announced Roadmap 2010, which set as a financial target increasing earnings per share (EPS) by 2010, which IBM achieved largely through a combination of layoffs, the sale of poorly performing parts of the company, buybacks of its stock, and routine culling of wasteful and redundant expenses. In the process, the company became addicted to layoffs and increased use of financial strategies, which IBMers and stock analysts increasingly called gimmicks. Efforts to grow IBM with sales and new offerings slowed. Palmisano explained his rationale for that first roadmap in mid-2015, when employees were attacking the second one, joined by a growing number of market analysts, industry watchers, and investors worried whether IBM had driven toward its goal with such force that it reduced its ability to compete. He argued that while investors demanded quarter-to-quarter improvements of all companies, this attitude made no sense when managing a firm. He admitted, “However, that investors want some direction as to where you’ll be so they can measure you and decide whether to invest in you or not. That’s a very fair request.” He continued, “So, we came up with something we felt we could live with and that made sense, which was a 2010 road map to go from $6 to at least $10 a share.” He denied it was a financial target, arguing it was “a long-term perspective of where the IBM Company could be in four to five years. It was a way to be shareholder friendly, but not quarterly driven.” But, put a number in front of an IBMer and they will chase it like a dog after a rabbit.7 The roadmap became the target driving many decisions and actions from 2006 through 2010.

The value of IBM’s stock remained elevated, as did profitable cash flows. IBM continued to pay generous dividends, and its senior executives enjoyed growth in their stock options, but IBM paid a hefty price for these results: a slowing response to new market conditions, such as its delayed entry into cloud computing.8 In the past, IBM’s ability to respond to changing market circumstances was normally at the rate it needed to optimize sales and profits or to be slightly ahead of its competitors, as in the 1960s and 1970s. In fairness to Palmisano, the company acquired PwC’s consulting arm, which was a true investment in the future of the firm, and worldwide manufacturing and service delivery systems were implemented. These actions worked.

Wall Street loved the boldness and confidence with which Roadmap 2010 was announced and carried out, because it promised predictable returns on IBM stock, but employees became hostile toward it, as they came to link it with layoffs of expensive U.S. employees to support the strategy almost from the time it was announced. Customers at first did not see any impact on their service from IBM, but by 2010 that was no longer true. Meanwhile, the press marveled at how well IBM seemed to be run, a pleasant circumstance as much attributable to their naive understanding of what was happening inside IBM as it was to excellent marketing and media management. As one Wall Street insider put it, “No public company we can think of does a better job of schmoozing Wall Street’s Finest and convincing them that there’s a there, when in fact the there is not quite as there as it might seem.”9 Armed with a detailed plan for how they were going to make “the numbers,” the strategists and financial community in Armonk recommended to Palmisano a new roadmap: $20 EPS by 2015. He and his board embraced it for the same reasons they had the first one, with the added incentive that Roadmap 2010 worked, overachieving its targets (reaching $11.52) and making its original goal a year early.10 The decision to announce this second roadmap involved many senior executives, but Palmisano took credit for it.

In 2009, when work on the next roadmap was under way, the executive in charge of IBM’s strategy was Ginni Rometty, then senior vice president and group executive for sales, marketing and strategy. In other words, she and her organization were primary authors, along with Mark Loughridge (b. 1953), who had worked in IBM’s financial communities since joining the company in 1977. Outside of IBM, he was credited as the creator of both roadmaps, with Palmisano’s full support. Loughridge prepared various detailed presentations on how IBM would implement its roadmap. To make the target, IBM had to expand business into new parts of the world and with new products and services, which is where Rometty and her team came in. In the end, buybacks and layoffs contributed successfully to the first couple of years of the second plan, while her team turned in a less stellar performance. It is important to call out her role because when named president and CEO of IBM in October 2011, she had fully embraced Roadmap 2015. One might argue that Palmisano was still around as chairman so she had no real choice but to continue supporting it until he turned over the chairmanship to her in late 2012, but she continued to support Roadmap 2015 for another two years. She really owned it.

By then, financial objectives dominated Armonk’s activities. The first line in Rometty’s first chairman’s letter, in the annual report for 2012, makes it obvious: “I am pleased to report that in 2012, IBM achieved record operating earnings per share, record free cash flow and record profit margins, with revenues that were flat at constant currency.” She continued, “Operating earnings per share were up 13 percent, putting us well on track to our 2015 Road Map objective of at least $20 of operating earnings per share. Importantly, we continued to deliver value to you, our owners.”11 That year, IBM generated $104.5 billion in revenue. It was perfect; just keep doing what Palmisano had done. Investors loved it.

One analyst summed up investor views: “I’ve never heard anyone within IBM articulate the value of the company the way she does.”12 But the roadmap represented more than a “goal” (her term). Rometty explained:

As before, the road map is not simply a list of targets, but a management model built on exploiting multiple ways to create value. Operating leverage will come from our continuing shift to higher-margin businesses and improving enterprise productivity—expected to be $8 billion over this five-year period. We will create value for shareholders through an anticipated $50 billion in share repurchases and $20 billion in dividends.13

Cutting the number of available shares would increase earnings per share (EPS). Rometty anticipated generating $20 billion from “growth areas” by 2015. This growth would come from new markets where local economies were expanding, business analytics, “cloud computing,” and an array of uses of computing to improve human activities such as managing supply chains, farming, water, sewer systems, and urban operations, all under the clever marketing title of “smarter planet.” Charts detailed how EPS would be achieved year by year, with plans being developed, implemented, and updated quarterly until IBM abandoned Roadmap 2015 in October 2014. EPS of $20 had become the target before all others.

Every annual report from this period, and many speeches by Palmisano and Rometty to employees and the public, made it clear that Roadmap 2015 was the focus. One participant in quarterly IBM analyst conference calls recalled that these were “run by the CFO and focused strictly on the numbers: on the revenue number, on the cash flow number, on the share buyback number, and on the earnings number. In particular, the per share earnings number.” He added, “The CEO never graces the call, and no actual business operators discuss their business. No success stories are told, no customers highlighted. It is all about margins and currencies and so-called one-time charges and so-called one-time gains, and tax rates, and how all those things added up to the earnings per share that quarter. And, most especially, what that earnings per share means for The EPS Roadmap” (analyst’s emphasis).14

Rometty was the longest-serving champion of Roadmap 2015. A graduate of Northwestern University with a science degree, she joined IBM in 1981 as a systems engineer in a Detroit, Michigan, branch office. Articulate, bright, and filled with a confident energy, she quickly moved up the organization in the 1980s and 1990s, largely in services. In 2002, she became the executive most responsible for the acquisition of PwC’s consulting arm, continuing to ride IBM’s march into services and consulting. Colleagues noticed in the 1980s and early 1990s that she sported a large ego and drive. Once she entered management, she tended to use the word “I” instead of “we” to claim credit for accomplishments by her organization and developed a reputation for not being loyal to her employees. At IBM, it was more customary to be inclusive and in public to be humble about one’s achievements. In one internal presentation, Rometty chastised her services executives and managers for being more interested in their families than in IBM’s business. The next day, she sent the attendees an e-mail that may have been the closest to an apology she ever made in public. Her language and behavior dogged her throughout her career, and she increasingly drew criticism because she “never covers your back.” One manager to whom she reported found her “not warm,” lacking a strong “people leadership style.”15 People commented that she did not inspire employees, treating them, to use her word, as “resources,” a term that irritated many IBMers. As she approached Palmisano’s position as head of the company, she had rivals to contend with. One, Steven A. Mills, senior vice president of the profitable software business, was reaching the age of retirement. Another, Robert Moffat (b. 1956), Mills’s peer responsible for the systems and technology (hardware) group, got caught up in a sex for insider trading scandal in 2009, was fired, convicted, and served six months in federal prison.16 A close confidant of Palmisano, he was also responsible for many thousands of layoffs and was therefore unpopular among wide swaths of IBM’s U.S. manufacturing workforce.

Rometty was qualified and faced no rivals, so it made sense to give her the CEO job. It helped that she had a charisma that worked well outside of IBM, so essential in dealing with customers, investors, and the media. Within the services business, she had a reputation for cutting expenses and laying off people to make her targets and for being weaker when it came to implementing revenue growth strategies. She focused more on financial strategies than Palmisano had, far less on expanding the business or nurturing the sales culture that had long worked for IBM. She embraced the move to software and services. She had tied her destiny to Palmisano’s. On the occasion of her appointment as CEO in 2011, she announced that, “What you’ll see is an unfolding of the strategy we have in place,” and according to the New York Times took credit for having “a hand in creating it.”17 She proved more of a technocrat than an inspirational leader, more in the mold of Palmisano and Akers but more blunt.

Figure 19.1

Virginia Marie “Ginni” Rometty served as IBM’s chairman, president, and CEO beginning in 2012, during a difficult period in IBM’s history. Photo courtesy of IBM Archives.

The centerpiece of Rometty’s leadership rested in managing financial results and shareholder value. In April 2012, she hosted her first shareholder meeting, in Charleston, South Carolina. She reported “weak” earnings for the first quarter—unfortunately for IBM a harbinger of things to come—but doubled down on the financial strategy already in play for a decade by reminding her audience of the value of the stock and the $15 billion spent on stock buybacks and $3.5 billion in dividends, up 15 percent over the previous year. She proudly announced that IBM had spent $6.3 billion on R&D, which seemed stingy in comparison to the other numbers she bandied about. She argued that, “Since the beginning of 2000, we have returned $133 billion to you in the form of dividends and share repurchases, while investing $81 billion in capital expenditures and acquisitions, and spending nearly $70 billion on R&D.”18 The global economy was in various states of recession or weakness. At this annual stockholders meeting, she reiterated her commitment to her “Roadmap to the Future.”

By July, IBM’s revenue fell only 3 percent, despite the continuing deterioration in the global economy, while net income rose by 6 percent, the latter made possible by stringent controls on expenses. Employees complained that Rometty made this possible by laying off colleagues. By October, it seemed that IBM was settling into a pattern of disappointments. Phil Guido, in charge of North American sales, admitted to his employees that “we had a disappointing quarter,” because they had not grown revenue. He blamed customers for putting off decisions to buy while they waited for anticipated new products. He also blamed poor performance by sales teams, using phrases like “we had execution issues.” The bad news kept coming in. First quarter 2013 saw no revenue growth, despite continued acquisitions of software firms by IBM. Already shrinking service revenue declined by another 4 percent.

Then a psychological bomb hit IBM HQ. The blast was eerily similar to the one that hit IBM in the early 1950s when its first customer, the U.S. Bureau of the Census, from which Hollerith had come with his tabulating equipment, which became the core of early IBM, chose to obtain its first computer from Univac, not IBM. If one had to pick one event more than any other that caused Watson Jr. to move quickly into computers, that was it. Now it seemed it was happening again, involving “cloud computing”—a form of mainframe outsourced processing. The U.S. Central Intelligence Agency (CIA), a massive user of computing, was moving to the cloud and put out a request for bids for that business. IBM made the short list of two potential suppliers. On February 14, 2013, the agency awarded the business to Amazon.com, the large online retailer, for which cloud services was not even core to its business. Amazon.com had made the case that because e-commerce was very much like cloud computing, involving millions of individuals accessing server farms to process data (orders), it could support the CIA’s large number of employees. What made the award more shocking was that Amazon’s bid cost a third more than IBM’s. IBM protested, since it was the low-cost bidder, but government auditors responded that CIA officials had “grave” doubts that IBM could provide reliable technology and cloud computing, rating its capability “marginal.” IBM was now, incredibly, too “risky” to choose. As one reporter put it, “The CIA butt kicking is a microcosm of larger problems IBM is having as it struggles to adapt to the cloud era,” instead focusing too much on what the writer referred to, using an internal IBM employee term, as “Roadkill.”19

Complaints leaked to the public. On the Internet, bloggers accused Rometty of not having a vision, of she and IBM executives being “out of touch” in going after the cloud opportunity, and of maintaining too many layers of management. One industry commentator, Steven Zolman, argued that “IBM’s sales culture is poison” because so many in sales did not understand the products and services they sold, obsessing instead on closing higher-margin sales. Moreover, executive compensation was tied directly to whether the company made the earnings per share targets of Roadmap 2015.20 Others worried that exporting services to other countries—offshoring—reduced the quality of IBM’s performance in North America and Western Europe. Success at IBM had to start “with a strategy that’s designed to impress more than just Wall Street investors”; instead Rometty was providing “more of the same” rather than a response to changing conditions in the market.21 Customers were buying fewer machines and software housed on their own premises and more computing services through subscriptions to cloud-based offerings.

Headlines became uglier, such as “IBM’s Sales Slump Turns Stock into Dow’s Lone Loser of 2013,” despite increases in dividends. Total sales declined in 2013 because of sluggish sales of hardware that could not be made up by growing revenues for software and services. Rometty was blamed for trying to cover these gaps by relying on job cuts, tax gains, and sales of assets in order to increase EPS. One portfolio manager, Todd Lowenstein, complained in late December 2013 that, “We don’t think investors are going to be paying up for financially engineered EPS.” They “want to see top-line growth, and IBM is just not part of that trend.”22 IBM’s supporters among reporters, bloggers, and industry watchers were now making suggestions on how IBM could fix its problems.23

IBM’s customers found a growing number of cloud service providers. IBM did not see growth in overall sales as measured by revenue for the next several years. On January 21, 2014, after releasing the previous year’s disappointing numbers, Rometty announced that the executive team would “forgo our personal annual incentive payments for 2013.”24 But she was sticking with Roadmap 2015. It began to appear that customers were voting with their budgets for something IBM was having difficulty offering.

IBM’s circumstances deteriorated more rapidly in 2014. Stockbrokers were tiring of IBM’s focus on earnings and began paying more attention to its shrinking revenue:

Even though IBM delivered 11% EPS growth in FY13, revenues declined, and the company saw a sharp decline in hardware profitability.… While IBM was able to offset much of the decline through cost savings and margin expansion in software and services, IBM also excluded a $1B restructuring charge from non GAAP EPS, a stark change from previous years when restructuring expenses … were included in the P&L. Moreover, a lower tax rate … generated over 85% of the company’s EPS growth.25

Translated into nontechnical language, IBM used an alternative set of measures along with the required GAAP (generally accepted accounting principles) to emphasize its cash flow.26 Another criticism of Roadmap 2015 was that, “It’s a bit of a mystery why IBM is sticking to the EPS roadmap when it appears less and less relevant.”27 One stockbroker forecast that IBM would “repurchase $15B or more of its shares this year [2014], and rebalance its workforce by at least 13,000 heads. Many investors believe IBM should be investing in its future by buying new technologies and investing in building capability, rather than rightsizing the organization.” The same commentator offered up the theory “that IBM’s financial incentives (senior executives; incentive programs are based 60% on operating net income near-term and 80% of operating EPS long term) make the choice simple.”28 More sinister reports began circulating that IBM had stashed assets and employees in subsidiaries in low-tax countries, such as Ireland and the Netherlands: “The Dutch group [IBM] had three employees in 2008, a number that has since swelled to about 205,000 as of the end of 2012—only 2 percent of whom actually work in the Netherlands.”29 The press began reporting that IBM had accumulated $44 billion in profits outside the United States for which it paid no U.S. taxes. More analysts and industry watchers criticized IBM’s fixation on earnings that spring and summer.

In first quarter 2014, revenue shrank again by 4 percent year over year. Rometty shifted her message to growth: “In the first quarter, we continued to take actions to transform part of the business and to shift aggressively to our strategic growth areas including cloud, Big Data analytics, social, mobile and security.” She added, “As we move through 2014, we will begin to see the benefits from these actions.”30 Meanwhile, IBM’s stock declined in value by roughly 20 percent. According to one critic, the market’s support for IBM so far came “with [a] 25% reduction in IBM’s workforce, most of which will be shed from the company’s hardware business. When has that ever been good news? The company called it ‘rebalancing its workforce’ so that it can better focus on its new priorities including analytics, cloud and cognitive computing.” The same critic, however, said the “company has finally come to terms with its history of underinvestment.”31 CNNMoney put the problem more succinctly: “Two years into Ginni Rometty’s tenure as CEO, the company faces a double dilemma: Revenue is shrinking while the company is having trouble hitting its ambitious EPS targets.”32 Wall Street was becoming nervous about her priorities and ability to execute. Forbes got personal, saying, “IBM leadership appears to have lost its way,” adding that “managing earnings is not managing for long-term success” and concluding that “the real problems—R&D cuts, higher debt, massive stock buybacks” were dampening optimism for the company. That other firms were doing massive buybacks did not matter; IBM was now seen as being in real trouble.33

Reports of “fed up” IBMers seeped into the press: protests by IBMers in France over working conditions in June, earlier complaints at various IBM plants in the United States and Latin America, downsizing reports out of India. In September, employees in various parts of the company were informed that their skills were not up to snuff and therefore they would have to take training to stay in the company and, while doing so, would be docked 10 percent of their salary. People were shocked because normally an employer paid for training, not employees. This action was seen as yet another act of penny-pinching financial engineering to drive down costs in the face of inadequate growth in revenue. People began asking, what is going on? Where is the board of directors when you need it? What does such an action say about IBM’s values? Was IBM misleading customers when it said it had the skills needed to serve them, justifying IBM’s pricing, yet accusing employees of not having those skills? One reporter wondered about Rometty’s role. Was this a random act by an executive making “a boneheaded move without informing anyone?” The editorial conclusion to that question advised IBM’s management: “Don’t be jerks.”34 Headlines became hostile, even from sources historically friendly to IBM, such as The Motley Fool, which ran one that read, “Should the Dow Jones Get Rid of International Business Machines Corp?”35 It all came to a head on October 20, 2014.

It was not going to be the usual quarterly conference call hosted by IBM’s CFO to answer analysts’ questions about earnings past and anticipated. For one thing, Rometty turned up on the call along with her CFO, Martin Schroeter. On the call were analysts from Goldman Sachs, Morgan Stanley, Sanford Bernstein, Barclays, UBS, Citi, and other institutions, including Cantor Fitzgerald, now back in business after losing almost its entire staff in the destruction of the World Trade Center towers on 9/11. Schroeter first went through the normal recitation of the numbers: a 4 percent decline in revenue this quarter compared to 2014’s third quarter, and net income of $3.7 billion, also down from expectations. Services revenue was flat, margins declined, the company faced a “tax headwind,” and it generated about the same cash as before ($2.2 billion). Hardware sales dropped precipitously. Schroeter blamed weaker software revenue, again on “some sales execution issues,” and volatility in world currencies. He reported “strong growth” in revenue from IBM’s focus areas, its “imperatives.” A piece of news he wanted to emphasize but that the press paid hardly any attention to was IBM’s sale of its microelectronics business to GlobalFoundries, the firm that would provide IBM with its future semiconductor technology. Schroeter said that IBM would move more quickly into higher-growth businesses. Schroeter said IBM would use more overseas “global delivery skills,” code language for cheaper workers. Because of divestitures of slower businesses, he cautioned to expect revenue to be down $7 billion, with pretax losses of a half billion dollars. So far, this was a routine call about declining IBM performance and a strong defense of what it would do next.

As Schroeter was ending his prepared remarks, it became clear to everyone on the call why Rometty was on it as well. The CFO dropped a bombshell so quietly, so briefly that it took even IBMers several days to realize what had just happened. Schroeter announced, “Given our third quarter performance, the actions we’re taking and with only 15 months till the end of 2015, we no longer expect to deliver $20 operating earnings per share in 2015.” IBM had just announced that it was abandoning Roadmap 2015.

Events in 2015 and beyond would determine whether that was what Rometty and her team were doing. They already had a new strategy, one that many IBMers, investors, and customers liked—the “imperatives.”36 In response to a question, Rometty doubled down on the importance of IBM’s cloud strategy, which included data and analytics, saying that IBM was going to move more quickly in transforming the company to pursue new growth areas. Asked about declines in the number of employees, the CFO acknowledged some occurred as a result of selling businesses but also because of relying on “global delivery centers,” in which IBM did work remotely for clients. Rometty reinforced the need for speed, use of mobile and other technologies, and reducing layers of management. Toward the end of the call, Rometty repeated that, “We no longer expect to deliver the 2015 EPS objective.” She admitted that Q3 of 2014 was bad, yet the company was “fundamentally better positioned than it was a few years ago.”37

Once IBM released its numbers, the stock market responded quickly. IBM’s stock dropped to a three-year low. The press picked up on the message about canceling the roadmap. Rometty sent a note to all employees announcing the weak performance, then pivoted to a reminder that the company would implement a three-part plan: establish a dedicated cloud business, introduce more flexible software offerings for customers and transform how services were delivered, and continue to streamline the company’s operations to improve its speed and agility. Mark Cuban, an outspoken billionaire, blurted out his disgust: “IBM is no longer a tech company. They have no vision. What they’ve evolved into is a company that does [arbitrage] on acquisitions. It’s stock buybacks. Who is IBM anymore?” He would “absolutely not invest in IBM,” saying it was no longer a computer company but instead “they specialize in financial engineering.” Just as bad in Cuban’s opinion, “They are an amalgamation of different companies that they are trying to ar[bitrage] on Wall Street, and I’m not a fan of that at all.”38 He was expressing how many people reacted after the announcement. One reporter put it simply: “And now Rometty is left with no clear, articulated turnaround strategy.”39

Andrew Ross Sorkin at the New York Times came partially to Rometty’s defense by claiming that she was stuck with the prior strategy launched by Palmisano, but he was also tough on her: “All these ‘shareholder friendly’ maneuvers have been masking an ugly truth: IBM’s success in recent years has been tied more to financial engineering than actual performance.” Probably emboldened by the 7 percent drop in the price of the stock to $169.10, he criticized IBM’s practice of buying back so many of its shares, citing $138 billion spent on share buybacks and dividends, while spending only $59 billion on its own operations (capital expenditures and $32 billion in acquisitions). These actions led him to conclude that “IBM has arguably been spending its money on the wrong things: shareholders, rather than building its own business.”40 He was not alone in this view. David A. Stockman, in an earlier life President Ronald Reagan’s director of the Office of Management and Budget and a banker, posted on his website earlier in 2014 that “IBM is a buyback machine on steroids that has been a huge stock-market winner by virtue of massaging, medicating and manipulating” its earnings.41 Sorkin argued that Rometty was “late” in reinventing the company, and shareholders should have pressured the company to implement needed changes instead of basking in the mirage of rising stock values and strong dividends.42 Could Rometty turn the company around in time?

IBMers and ex-employees went to the Internet to opine: “Anybody who watched quarter after quarter of revenue declines while expecting the $20 target to happen must believe in magic”; “Give the company back to technology, taking it away from financial engineers … and treat their star employees well”; “Build organic products that companies and people want”; “Under the deal with GlobalFoundries, IBM will pay the semiconductor company $1.5 billion to take over chip manufacturing operations [true] so the business was managed to the point it had a negative value?”; “Buybacks are, in my view, a negative leading indicator that signals no other productive uses of capital are available.” Some thought Rometty inherited a bad situation from Palmisano, that he had cleaned out much of the fat, leaving her to cut muscle to make the target. But a more widely held opinion was that, “She has had almost 3 years to make things better at IBM and has failed convincingly. When is the Board of Directors going to take some action? Is anybody at the helm or is the ship going down with the crew?”43 Concerns about layoffs dominated employee comments: “Sure lay off all employees here in the US and offshore every job. See how fast this ship sinks then”; “If layoffs are necessary, so be it. But Rometty should be the first to go”; “Same old story at IBM, layoffs, layoffs.… They have no other tricks.” The bottom line was that, “IBM has simply lost its edge.” Harking back to her reputation for lacking empathy, one employee wrote, “She speaks to staff and customers as if they were toddlers in kindergarten.… It breaks my heart after 40 years to see the changes in IBM. They forgot their strength was their people.”44

When Rometty reported to investors in the annual report for 2014, published in early 2015, she no longer led off with Roadmap 2015 but rather with the argument that IBM was continuing to move to higher-value businesses, with a growing consensus of approval among IBM’s investors and business media. It was also her strongest set of statements about what would soon be called IBM’s “imperatives,” such as the move to cloud, analytics, and Big Data. Revenue had shrunk again, this time by 5.7 percent, to $92.8 billion. Now consolidated gross profits had shrunk as well, by 4.7 percent. Global Services, supposedly the growth part of the business, shrank by 3.5 percent.45 The first quarter was marginally better, with EPS up over first quarter 2014 by 9 percent and net income up 4 percent on flat revenue. Of that revenue, the contribution from strategic imperatives rose by 30 percent. Rometty reported things were unfolding as planned. Meanwhile, she hired consultants to help develop a strategy to fight off activist investors bent on ousting her. The activists failed, but they caught the attention of Wall Street.46 The year 2015 was turning out to be another bad year.

It proved far worse for employees who were laid off. “Resource actions,” “RAs” as they were known at IBM, attracted increased attention from the media.47 By late spring 2015, such stories were linked to IBM’s overall performance. Instead of having 134,000 employees in the United States, as it did in 2005, it now employed roughly 78,000—Corporate no longer reported how many it had anywhere in the world, just the grand total. Analysts speculated on how many more would go based on charges to restructuring reported by IBM, leading to estimates that IBM had disposed of 15,000 U.S. employees in 2014 and might “only” eliminate 12,000 in 2015. As one analyst put it, “That’s what it’s come to at IBM. Cutting the jobs of 12,000 people gets shrugged off as business as usual.”48

It was becoming difficult to find optimists. In October 2014, the board authorized another round of stock buybacks, while Rometty declared that, “We will continue to make the investments and changes necessary to manage our business for the long term and to shift to higher-value offerings.” But, she also doubled down on the financial intent of remaining “fully committed to returning significant value to shareholders,” hence the buybacks, dividend payments, and continued reduction in the cost of employees. Pondering all of this, one American analyst wondered “whether IBM will have enough cash flow to fund these buybacks organically, or if it will need to issue even more debt,” which could cause rating agencies at some point to “downgrade IBM to junk.”49

Revenue dropped by over $11 billion to $81.7 billion. Net income increased by a billion dollars, although income from continuing operations shrank by $2 billion. Net cash remained strong, at $17 billion; cash on hand was closer to $8 billion. Buybacks continued at nearly half the rate of earlier years. The board approved a slight uptick in dividends, but the value of the stock had dropped from $160.44 at the end of 2014 to $137.62 on December 31, 2015. IBM was positioned to lose many supporters. The story was becoming monotonous: investments in growth areas, continued layoffs, excuses about currency issues, and IBM’s calls for patience from investors who had been extraordinarily patient during Rometty’s tenure, probably because of consistent dividends and buybacks.

In 2016, IBM’s revenue declined from $81.7 billion in 2015 to $79.9 billion, gross profit from $40.7 billion to $38.3 billion, and the all-important net income from $13.2 billion to nearly $11.9 billion. A combination of stock buybacks and increased dividends caused the stock’s value to rise by over 20 percent, but the legacy business was still shrinking faster than the strategic imperatives were growing. However, imperatives now accounted for just over 40 percent of the company’s revenue, which many felt meant IBM was finally getting traction in its latest transformation.50 While the CFO explained last year’s financial performance, Rometty was in Davos, Switzerland, at the World Economic Forum.

THE PROBLEM OF LAYOFFS

Dogging Armonk was the company’s continual turnover of personnel. As the company shifted out of low-profit businesses and into others requiring new types of skills, IBMers left as part of what was sold off or transformed. Prior to such events, and over a period of years, layoffs occurred on a continuous basis in those parts of the business. No part of the business felt secure from layoffs, and not just the lower ranks. Increasingly, those in higher echelons, at divisional director and vice presidential levels, were laid off also. Manufacturing, software, and services divisions more aggressively shifted work to places where workers were less expensive. The poster child for that strategy was the move of thousands of jobs out of the United States to India.51

Layoffs could not be hidden in IBM’s factory towns of Endicott, Poughkeepsie, and Fishkill, where IBM populations decreased by the thousands. In Endicott, IBM’s presence went from over 10,000 to under 1,000 in less than fifteen years. Weeds grew in the empty parking lot at the closed IBM Country Club; paint peeled off the walls in the clubhouse itself. The Homestead became a privately owned hotel. Poughkeepsie’s IBM population shrank just as dramatically. Similar stories of job losses could be cited in Germany, France, Canada, Great Britain, and Australia, among other countries. The thinning and replacement of IBM’s population occurred in a recurring pattern. Those parts of the business that represented the past, the least profitable, contracted continuously, quarter after quarter, year after year. An employee’s performance rating or skill played a decreasing role compared to what organization they resided in when it came to influencing their fate. By 2010, it seemed employees waited for the ax to fall, to be, in their words, “RA’ed” (meaning a “resource action” occurred). It began largely in those countries where labor laws provided the least resistance to such layoffs, notably nations of the British Commonwealth and the United States. As IBM worked to achieve its Roadmap 2015 goals, layoffs spread across France, the Netherlands, and Germany, despite local regulations making them difficult and expensive.

Four other trends accompanied layoffs. First, between the 1990s and 2017, severance payments shrank. For example, in the United States in the 1990s, someone leaving IBM would receive two weeks’ salary for every year worked at the company up to six months of their current pay, in a one-time payment, plus medical benefits. By 2015, severance was one month’s pay.

Second, benefits declined for those remaining. There was the case in the late 1990s when IBM attempted to replace its traditional pension plan with a new system that put money in an employee’s 401(k) fund that discriminated against workers over the age of 40. A court ruled that illegal. In 2006, the company tried again. Fearful of rising and unpredictable costs for pensions and medical insurance in the United States, IBM announced that it would stop payments to its defined-benefit pension plan and instead make those payments into an employee’s savings plan. In 2012, Corporate announced it would only make a one-time annual payment to the 401(k) accounts of individuals on the payroll as of December 15, so anyone laid off earlier in any year would not receive matching funds for pension purposes. Table 19.1 lists some of the obvious changes. Each step met with anger on the part of workers, who lit up websites with complaints. The layoffs continued.

Table 19.1

Changes in IBM’s U.S. employee benefits, 1999–2014

1999

IBM pension plan replaced with a cash balanced account—later dismissed by court order as being illegal

2005

IBM announces it no longer will have a defined-benefit pension for employees; replaced with a payment to their 401(k) accounts instead

2007

IBM stops contributions to its pension fund, freezes all pension benefits, with no accruals for long-term employees;

IBM announces enhancement of Tax Deferred Savings 401(k) Plan (TDSP)

2013

IBM announces that to receive TDSP matching grants an employee must be on the payroll on December 15, with funds to be deposited on December 31

2014

U.S. retirees on Medicare moved to extended health plan

Source: Peter E. Greulich, A View from Beneath the Dancing Elephant: Rediscovering IBM’s Corporate Constitution (Austin, TX: MBI Concepts, 2014), 161–162.

A third, perverse practice seeped into IBM’s behavior, noticeable by 2008 but blatant by 2010, involving the appraisal system. Every individual received a year-end performance appraisal that after text describing results included a rating—a grade—from “1” (outstanding) to “2+,” “2” (average), “3” (underperforming), and “4” (not doing the job). There had always existed loosely defined guidelines as to what percentage of the workforce in a department, region, plant, or division should receive “1” appraisals, for example, much like an academic grade curve. As the years passed, guidelines and rules of thumb became hard targets imposed to keep labor costs down (a 1 was paid the most, a 2 received no bonuses or salary increases, and so forth). By 2010, even if a manager had an outstanding set of employees, the human resource community mandated that a certain percentage had to be rated a “3.” The percentage to be rated a “3” evolved into specific targets; for example, “You will appraise 5 of your staff a 3.” When the next layoff was announced, those rated a “3” were the first to go. The appraisal system lost credibility. Employees and their managers did not take the exercise seriously as a means to motivate better performance and reward results. It got so bad that by 2015 the company had implemented a new appraisal system, promoted as providing continuous feedback to employees on their behavior and performance. Nothing changed; layoffs continued.

IBM was saddled with an appraisal system that many employees believed only worked to identify who to lay off next and was far from the Basic Beliefs one chose (Watson Sr.’s, Gerstner’s, or Palmisano’s). One employee’s comments were typical, this one from October 2016: “Longtime IBMer. Always Top Performer 1+, 2—never a 3 in my career. Was put on a PIP [performance improvement plan, improve or you are fired] along with untold others. All high level and older. Instead of RA they were told their performance was not up to standard for the half and given incomprehensible targets for the PIP. 30 days severance.”52 By 2011, dismissed IBMers were complaining of age discrimination, largely in the United States. It did not help that Rometty and executives used the term resources to describe employees. The dehumanizing aspect of the word became an expanding affront, particularly to older employees. Thousands of employee Internet postings from 2007 to 2018 complained that morale was “terrible.” From a human resources perspective, IBM was in as serious a period of personnel churn and crisis as it had been in since the late 1980s.

Fourth, layoffs were done gruffly, with an increasingly less sensitive approach to the feelings of individual IBMers. Their website postings were filled with accounts of the bluntness of layoffs, and these were now also beginning to filter into the economic and business literature. One early example from 1993 set the tone. An employee in good standing in Poughkeepsie, with 16 years of service, was told he was being laid off and was dismissed out of his building. His formal dismissal letter was both demeaning and reflected a betrayal of trust in his company: “You have been designated a ‘surplus employee’ effective immediately.” This individual was one of 8,000 dismissed by IBM from the Central Hudson Valley area of New York.53 That other large U.S. corporations were doing the same, including cutting back health benefits and retirement pensions, only bolstered senior management’s belief that layoffs were a necessary and effective strategy by which to run the company worldwide. The old social contract had been broken. The chief economist at the U.S. Department of Labor in 2010–2011, Betsey Stevenson, observed that the days of taking care of one’s employees were over, that “profits and efficiency have trumped generosity.”54

With all the churn, it became increasingly difficult to know how many people worked at IBM. It is normal in a large enterprise for people to come and go, for the employee population to ebb and flow, and IBM is no exception. IBM, however, only publishes total population numbers and in 2018 it stated that the firm employed 367,000 worldwide.55 In March 2018, a nonpartisan news outlet, Propublica, published a detailed report documenting age discrimination in how IBM chose who to lay off in the United States and in Western Europe, adding further controversy to its personnel practices.56

FIVE INITIATIVES AND THE EMERGENCE OF A NEW STRATEGY

With the obsessive roadmap focus diminished, the elements of IBM’s strategy underpinning it became more obvious, easier to talk about, and increasingly made more sense to IBMers, analysts, and customers. Time would tell if Corporate had really retired Roadmap 2015 or simply continued it without explicitly discussing it. Rometty restated her financial objectives in ways that aligned more clearly with IBM’s business strategy. For example, in early 2015, she set her sights on having IBM generate over 40 percent of its revenue in 2018 from five markets: analytics, cloud computing, cybersecurity, social networking, and mobile technologies. Older businesses, such as mainframes, would diminish, but IBM’s challenge would be to grow revenues from new markets quickly enough to cover the retreat from the old while sustaining, and at some point growing, total revenue and profits. Units that IBM sold off, such as chip manufacturing, represented the elimination, in Rometty’s words, of “empty calories,” businesses that generated revenue but low or no profits. Hardware was not going away; rather, it was needed in new forms. Rometty said, “We can’t hold on to our past.” Some employees thought she was shedding IBM’s core culture as much as old businesses when making such comments.

IBM got ahead of competitors with analytics software and with processes to manage and analyze massive bodies of data. While IBM arrived late to cloud computing, it slowly extended its presence among large enterprises in 2015 and 2016 with combinations of cloud services. Services involving cybersecurity represented a new market for IBM and most competitors in 2016, while a partnership with Apple Computer offered the promise of portable computing and app development platforms lodged in cloud servers. Social networking involved management consulting as much as software services. IBM’s acquisitions were closely tied to supporting these areas of focus, particularly to catch up in cloud computing. IBM purchased a cloud computing firm called SoftLayer for $2 billion and opened a number of data centers to house cloud services. IBM amassed skills it did not have in cybersecurity, a new field for all IT vendors and customers, purchasing firms with those capabilities. Now Rometty had a strategy that resonated and played to her strengths in communicating their importance.57

Perhaps more interesting, yet still difficult to see, was a different transformation occurring with IBM’s offerings, involving Watson, an artificial intelligence offering combining computing hardware and software. If it worked as engineers at IBM and its senior executives anticipated, it could be as important as System 360. More than a bundle of technologies, it consisted of a series of new services and capacities as well. These included analytics for Big Data, massive computing, and the ability of computers to learn, to teach themselves new facts, and to acquire new insights. It also combined technologies still troubled with start-up issues and much hype. Watson is a computer that answers questions presented to it in a human language. IBMers poured trillions of facts, entire libraries, and every medical article they could get their hands on into this machine’s memory and armed it with tools to rapidly scan this data for an answer. As it responds to questions, it learns from the experience what better answers to offer. This is a simple description for an early, practical form of artificial intelligence converted into a commercial product, a service, and most readers have already heard about it.

In February 2011, on the popular American television quiz show Jeopardy!, two of the show’s previous champions competed against an IBM computer named Watson.58 Contestants were given broad clues to an answer and then had to phrase their response as a question. The humans had rapid command of large amounts of facts and excellent judgment, even instinct, about the right answers. Along with Watson, the contestant who responded the quickest won dollars. Over the course of two nights—February 14 and 15—Watson competed against these two champions. On the first night, it tied with one and got ahead of the second. On the second evening, Watson crushed the humans, winning $77,147 to their $24,000 and $21,600, respectively. The grand prize of $1 million went to Watson; IBM donated the money to charity. In the studio audience were IBM employees, from Palmisano to IBM scientists and engineers who had been working on artificial intelligence and computing for decades. It was not clear if Watson could win, although the engineers had played many games with it, even with the TV producers, who also wanted to know if it had a chance before putting it to the test on national television. It probably did not calm nerves that IBM’s executive royalty showed up for the live contest. When Watson won, chests puffed at IBM for a moment, especially in the Research Division, for it had not enjoyed such a spectacular success in a long time.59 IBM garnered much positive PR and goodwill. Winning was what so many people expected: “It was IBM, what do you expect?”

Why all the fuss at IBM, in the world of computing, and even by the press and the public? While critics would then, as now, point out the limits of Watson’s capabilities, nonetheless something had changed. Almost everyone had known for decades that if you provided a computer with very specific instructions—software—it could carry out tasks, like moving data through the Internet, moving boxes on a conveyor belt, or rapidly calculating accounting numbers. That had been going on since the 1940s, but, as two IBMers commenting on that kind of computing put it, “Computers today are brilliant idiots,” because all they could do was store large quantities of data and perform calculations on it. They could not understand, learn, or adapt the way a human could, and those were the capabilities computer scientists were beginning to bake into this new class of computers. That is why the Jeopardy! victory was so important for them, for IBM, and probably for the world.60 The victory felt like IBM was back where it historically had always found itself—in the middle of the grand challenges of its time, and this one included big computers! As its developers improved Watson’s capabilities, IBM’s executives and computer scientists increasingly warmed to the realization that this technology had the potential of burnishing IBM’s offerings, of answering a question Rometty asked often of her employees: How do we make IBM “essential”?

Beating two quiz champions was an important marker of progress, but the breakthroughs had been inching forward over several decades as engineers and computer scientists found ways to make computers accumulate vast amounts of information—essentially everything, for example, ever published in medicine or weather data going back decades, even centuries—explore possible answers to a question, determine the odds of one being more right than another based on prior encounters with similar questions, and then select the optimal one, all rapidly. Now imagine a doctor or a medical researcher collaborating on, say, diagnosing a cancer and then have the computer suggest to that doctor treatment options that had a high probability of being successful based on prior cases, while feeding researchers case information in support of their own research. That was a new form of collaboration between humans and machines. Tracking its prior transactions—experiencing—with a topic, a question, or an answer, that system would apply learning to its activities, simply repeat steps, and keep track of how often it did them, or how quickly, as in more traditional computing.

The technology in Watson had evolved to a level where its users had a machine that exercised learning, or cognition, to be more technical. Once IBM’s marketing staff understood what the company’s engineers had started to create, they came up with a way to promote this evolution with what appeared initially as a clumsy phrase that captured the essence of what was happening: cognitive computing. People would simply have to become comfortable with the term because for years to come this form of computing would characterize how information processing would be done and for what uses. All of a sudden, new possibilities opened up for IBM, its competitors, corporate customers, and everyone else. Now imagine, for example, one’s smartphone connected to such a system answering questions individuals had that go beyond such mundane ones as getting driving directions or answering trivia questions to settle a bar bet. Imagine a doctor in a remote African village asking his smartphone to assess a lump in a woman’s breast to determine if she has cancer, and do it quickly, cheaply, and so accurately that this doctor would want to use his phone regularly in his practice. An IBM-Apple alliance on mobility now made sense. IBM’s early foray into massive computing to analyze and graphically present output of Big Data did, too, and so did cryptanalytics, for similar reasons. Today, with data being collected by sensors in quantities orders of magnitude greater than before, it could run through such systems to make it understandable to humans. As students of this computing put it, “A cognitive computing environment requires sufficient amount of data to discover patterns or anomalies within that data,” enough “that the results of analytics are trustworthy and consistent.” It was about acquiring insights, the thing that highly knowledgeable people did well.61

Kevin Kelly, a founder and editor of Wired magazine, is an unapologetic technoenthusiast. He visited the Watson system around 2015, four years after the Jeopardy! event. Watson’s technology had moved from a purely experimental system to one that was beginning to seep into IBM’s cloud offerings. It was setting up IBM’s cloud offering to be more than a less expensive place where customers could process data without having to buy all of those old “idiot” technologies and pay for staffs to babysit them. Kelly found Watson was being used by customers with whatever tools they had: smartphones, tablets, and so forth. His reaction is worth reading:

This kind of AI can be scaled up or down on demand. Because AI improves as people use it, Watson is always getting smarter; anything it learns in one instance can be quickly transferred to the others. And instead of one single program, it’s an aggregation of diverse software engines—its logic-deduction engine and its language-parsing engine might operate on different code, on different chips, in different locations—all cleverly integrated into a unified stream of intelligence.62

A great deal of progress had been made in a few years. It was also an example of an unintended consequence, because it was difficult to imagine it being so open to different platforms and input equipment if IBM had not made the decision in the late 1990s to embrace open source technology. Making IBM more agnostic about software and hardware used with its products facilitated development of Watson’s technical architecture.

Other IT companies understood there was the same potential in medicine (IBM’s initial focus area for Watson) and in other areas. Staff involved with Watson even published a cookbook based on the system’s analysis of what ingredients are most enjoyed by people, including one for Indian paella.63 Major cloud companies wanted this kind of technology; Amazon used cognitive computing in retail. By 2014, 322 companies were spending in excess of $2 billion to develop similar systems. All the key IT players were in the game: Google, Facebook, LinkedIn, Pinterest, Intel, Twitter, and Chinese firms such as TenCent and Baidu. All hired AI experts, even luring away some of IBM’s Watson people. IBM’s five imperatives were beginning to gain an identity different from the one they began with earlier. They started to dribble out—“escaped,” as Gerstner said of IBM’s laboratories—into concrete forms as machines, software, and specific services. If you have any lingering doubts, recall the history of RISC, which took years to evolve into products. As early as 2010, IBM strategists and technologists understood the potential: “The goal is to have computers start to interact in natural human terms across a range of applications and processes, understanding the questions that humans ask and providing answers that humans can understand and justify.”64

IBM began to invest in commercializing this technology. IBMers reported on its use by big customers after 2012. Of particular interest to Watson’s engineers and management, and to Rometty, was getting Watson working quickly in medicine. Here much patient data already existed in digitized form, where IBM had under way experiments and long-term projects and where the need for diagnostic tools and research was urgent, compelling, and could be financed by governments, universities, foundations, and pharmaceutical firms. IBMers formed collaborative projects with partners ranging from Nuance Communications on clinical decision support functions to doctors at Columbia University in New York to figure out where best to use such technology. A partnership created in September 2011 with WellPoint, an American health care provider, led to the development of early tools to suggest treatment options for doctors. Their initial product focused on lung cancer.

In January 2014, Rometty established the IBM Cognitive Business Group, committing the company to selling Watson-based services, with a staff of 2,000 and revenue targets. While its growth depended on continued evolution of Watson technology, offerings, and market receptivity in the face of growing competition, the group made progress, if more slowly than senior executives wanted. Nonetheless, Rometty bragged that it would generate $10 billion in annual revenue within a decade.65 She possibly had her own technology “big bet.”66 At IBM, a big bet required massive attention and the muscle of the entire firm, such as when Thomas Watson Sr. bet on tabulating equipment, Watson Jr. on System 360, and now Rometty with AI. In each instance, it was not fully obvious at the time when bets were placed that they were, in fact, big bets. The one possible exception was Lou Gerstner’s decision to keep IBM together, which he and everyone around him knew was major, but even this big bet did not have the risks of the others. System 360 did not start out as one but rapidly evolved into IBM’s biggest.

As with all big bets, execution was always the main concern. Critics could always be counted on to question the wisdom of making them. IBM was no exception. For example, in 2016, Tom Austin at Gartner thought Rometty’s “moonshot” would take years to unfold and said that, “It seems like they’re swimming upstream with that.”67 He did not say anything IBMers all over the company did not already understand. IBM could afford to step up and maintain its “constancy of purpose,” a phrase coined by quality guru W. Edwards Deming (1900–1993), whose ideas often reflected IBM’s approach to a business or technology strategy. The market for such AI technologies drew IBM’s attention as much as anyone else’s. In 2016, IDC, a highly respected IT industry watch group, valued the AI market at $8 billion. Its analysts thought the number would rise to $47 billion by 2020. No wonder Rometty was in a hurry. Again, IBM was in a footrace, one that favored her company.68 IDC opined that in 2020 the big winners would be Amazon, Google, Microsoft, and IBM. Although IBM was not sharing information about how much revenue it generated from each imperative, estimates by industry watchers speculated that revenue attributed to Watson technology hovered at $500 million in 2016. It did not matter whether this was a guess, hubris, or how IBMers coded their transactions in their CRM system. It remained their future.

IBM’s strategy involved providing AI platform tools and hardware, but unlike the other major players, it leveraged large bodies of data (such as weather data) and experts (such as weather forecasters and doctors). Medical uses accounted for two-thirds of IBM’s Watson revenue in 2016, and they seemed to be working. Steve Lohr of the New York Times reported:

At the University of North Carolina School of Medicine, Watson was tested on 1,000 cancer diagnoses made by human experts. In 99 percent of them, Watson recommended the same treatment as the oncologists.

In 30 percent of the cases, Watson also found a treatment option the human doctors missed. Some treatments were based on research papers that the doctors had not read—more than 160,000 cancer research papers are published a year. Other treatment options might have surfaced in a new clinical trial the oncologists had not yet seen announced on the web. But Watson read it all.69

Even though IBM’s financial performance remained anemic in 2016 and 2017, Rometty had to be feeling better about her overall strategy, thanks to Watson, as she said, “Digital business is converging with a new kind of digital intelligence—what you will recognize as Watson. We call this Cognitive Business.” You can forgive her hubris when she said, “We can literally build cognition into everything digital.”70 Rometty’s dominant messages were increasingly about cognitive business, a mantra she continued throughout 2018. For better or worse, she may have found her voice. Now she had to execute with a company still troubled.71

Notes