Microsoft's search case study summary

C-151

Microsoft: froM Gates to satya Nadella

20

IntroductIon

On February 4, 2014, Satya Nadella became CEO of Microsoft. Nadella, a native of Hyderabad, In- dia, was only the third CEO in Microsoft’s 39-year history. Cofounder Bill Gates was CEO from Mi- crosoft’s establishment in April 1975 through Janu- ary 2000 when he passed the reins to Steve Ballmer. Gates remained chairman though until February 2014. The Gates years were characterized by dra- matic growth as Microsoft expanded from a small start-up to become the largest and most dominant software company on the planet, in the process making Gates the world’s richest man. The founda- tions of Microsoft’s success during this period were its two monopolies: the Windows operating system, which at its peak was used on 95% of the world’s personal computers (PCs), and Office, which had a 90% market share in 2012.1

Microsoft continued to expand both revenues and profits under the leadership of Steve Ballmer. During his tenure, revenues expanded from $25 bil- lion to $70 billion while net income grew 215% to $23 billion. One area that did particularly during the Ballmer years was the Windows server business, a division that Nadella ran prior to becoming CEO.

Servers sit at the center of networks of PCs, and are used to perform a variety of functions including da- tabase hosting, file services, Web services, print ser- vices, and applications services. Microsoft makes a version of Windows, Windows Server, which runs servers. The Windows server business was a $20-bil- lion division by 2014. Microsoft gained share from competitors such as IBM, which promoted the rival Linux operating system. By 2014, 75% of servers built around Intel microprocessors used Windows Server as their operating system, as did around 50% of all servers.2 The Linux and Unix operating sys- tems took the number 2 and 3 spots.

Despite impressive growth, Microsoft’s stock price stagnated during the Ballmer era. This reflected a growing concern that Microsoft had lost its leader- ship in the computer industry to three firms, Google, Apple, and Amazon. Google had grown dramati- cally during the 2000s on the back of its dominant Internet search business. Along the way, Google had developed an operating system for smartphones (Android) and laptops (Chrome) that were now chal- lenging Windows on computing devices, a category that had expanded beyond traditional PCs to includ- ed smartphones and tablets. Google was also offering a “cloud-based” suite of productivity tools, Google Docs, which competed directly with Office.

02277_Case20_rev02.indd 151 01/10/15 6:03 PM

C-152 Case 20 Microsoft: From Gates to Satya Nadella

Apple, a firm that was nearly bankrupt in 1997, had done more than any other company to expand the definition of computing devices to include smartphones and tablets. Apple had introduced the first version of its smartphone, the iPhone, in 2007. Differentiated by elegant design and ease of use, two Apple hallmarks, the iPhone was a sensation that re- defined what a smartphone should look like and do. Apple followed the iPhone with the 2010 introduc- tion of the iPad, a tablet device that created an en- tirely new computing category, and one that canni- balized sales of laptop PCs. Both devices ran Apple’s iOS operating system, further reducing the relevance of Windows.

As smartphones and tablets gained popular- ity, more and more computing was being done us- ing these mobile devices—accessing applications and data stored on servers “in the cloud” rather than on a traditional PC. According to Microsoft’s own esti- mate, by mid- 2014, while 90% of traditional desktop and laptop PCs still used Windows, only 14% of all computing devices (a definition which included PCs, smartphones, and tablets), used Windows.3 Although under Ballmer’s leadership Microsoft had tried to grow its share by introducing a Windows smartphone and the Surface tablet, these offerings failed to gain traction. By 2014, Windows Phone had less than 3% of the global smartphone operating system market, while Apple’s iOS held 15.2%, and Android 81.1%.4 In the tablet market, Android had a 65.8% share, Apple’s iOS had 28.4%, and Windows tablets had 5.8%.5 Mi- crosoft was assumed to be losing significant amounts of money on its phone and tablet businesses. To compound matters, after three decades of sustained growth, PC sales were declining: PC sales fell by 4% in 2012 and by 9.8% in 2013, although demand stabilized in 2014.6

Amazon, the world’s largest Internet retailer, was challenging Microsoft from another direction. By the mid-2000s, tens of thousands of servers were being grouped together into “server farms” located in the cloud to host high-traffic Internet websites. Google had built server farms to host its Internet search busi- ness, Microsoft likewise had server farms for its Bing search business and MSNBC Web offerings, and Am- azon had built server farms to host its large online retail business. In 2005, Amazon leveraged the knowl- edge and capacity it had accumulated building server farms to start a new business, Amazon Web Services

(AWS). AWS hosted data, Web services, and applica- tions for paying customer. These data, services, and applications could be accessed from anywhere by a user with a computing device and an appropriate wireless or hardwire connection. By 2014, AWS was viewed as the market-share leader in the emerging cloud-computing business.

Microsoft entered the cloud-computing business in 2010 with Azure (later renamed Microsoft Cloud). Azure was founded within the Windows Server di- vision that Nadella ran prior to becoming CEO. In addition to hosting data and websites, Azure allows clients to build and run applications that reside on Microsoft’s cloud. By 2014, Azure was thought to be number two in the emerging cloud business, with Google and IBM rounding out the top four. Indus- try wide, the cloud-computing business generated $16 billion in sales in 2014, but it was growing very rapidly and was thought by many to represent the future of computing. 7

Commenting on Microsoft’s overall competitive position in 2014, the general manger of one busi- ness unit noted that: “I think we have about 18 to 24 months to get it right. If we don’t, Microsoft is fin- ished.”8 This statement reflected a widespread belief within the company that the computer industry was undergoing a massive paradigm shift, away from the client-server world based on PC architecture in which Microsoft had been so dominant, and toward a world of mobile devices and cloud computing in which Microsoft faced significant competitive challenges. Nadella was as cognizant of this as anyone. By March 2014, he had already honed his vision for the com- pany. Microsoft, he said, was competing in a “mobile- first, cloud-first” world.9 The task facing Nadella was deciding what actions to take to ensure that Microsoft survived and prospered in this brave new mobile-first, cloud-first world. He knew he had to act fast.

BIll Gates and the early hIstory of MIcrosoft

Bill Gates and Paul Allen established Microsoft in 1975. Gates was a 19-year-old Harvard dropout.10 Allen, who was 22, had dropped out of Washington

02277_Case20_rev02.indd 152 01/10/15 6:03 PM

C-153Case 20 Microsoft: From Gates to Satya Nadella

State University to work as a programmer at Honey- well in Boston. Gates and Allen had both attended Seattle’s elite Lakeside high school, where they had bonded over their common interest in computers.

By all accounts, the young Bill Gates was extremely intelligent, hypercompetitive, ambitious, hardworking, and a gifted programmer. One of his former teachers at Lakeside described him as the most intelligent stu- dent she had ever had. He could also be dismissive of people who lacked his technical acumen, abrasive, and hypercritical. One story widely circulated in Microsoft is that if he disagreed with the technical or product presentations of Microsoft employees, he would inter- ject with sharp comments along the lines of “that’s the stupidest thing I have ever heard” or that the idea was “brain damaged.” Legend has it that on more than one occasion Gates reduced a presenter to tears, al- though Gates would argue that it was never the person he criticized, just the idea. Gates respected people who were smart and hardworking like him, who marshaled their facts, and who stuck to their guns when chal- lenged by him if they knew they had the facts on their side. Gates ultimately relied upon such people to lead projects and businesses within Microsoft.

In 1975, Allen persuaded Gates to drop out of Harvard and start Microsoft to write a version of the computer programming language, BASIC, to run on the world’s first commercially available PC, the MITS Altair 8800, which used an Intel 8080 microprocessor. Gates and Allen met with the founder of MITS and demonstrated their version of BASIC for the Altair 8800. This resulted in a deal under which MITS dis- tributed Microsoft BASIC for the Altair 8800, mak- ing Microsoft the first company to sell software to run on a personal computer. Microsoft subsequently wrote versions of Microsoft BASIC that ran on other PCs of the time, including Apple’s first successful of- fering, the legendary Apple II, introduced in 1979.

In June 1980, Steve Ballmer joined Microsoft. Ballmer had been a friend of Gates at Harvard, and was the only person who had outscored Gates on mathematics and microeconomics classes at Harvard. Ballmer had worked at Procter & Gamble after Har- vard, and then moved on to Stanford Business School. Gates persuaded Ballmer to drop out of Stanford and manage business operations at Microsoft. He was em- ployee number 30.

In July 1980, IBM approached Microsoft about using a version of Microsoft BASIC for the IBM

PC, which was then in development. Gates persuad- ed IBM to adopt a 16-bit Intel processor (originally, IBM had been considering a less-powerful, 8-bit pro- cessor). Gates was also instrumental is pushing IBM to adopt an open architecture, arguing that IBM would benefit from the software and peripherals that other companies could make.

Initially, IBM was intent on licensing the CP/M op- erating system, produced by Digital Research, for the IBM PC. However, the current version of CP/M was designed to work on an 8-bit processor, and Gates had persuaded IBM that it needed a 16-bit processor. In a series of quick moves, Gates purchased a 16-bit op- erating system from a nearby company, Seattle Com- puter, for $50,000. Gates then hired the designer of the operating system, Tim Paterson, renamed the system MS-DOS, and offered to license it to IBM. In what turned out to be a masterstroke, Gates persuaded IBM to accept a nonexclusive license for MS-DOS (which IBM called PC-DOS). MS-DOS had a command-line, text-based interface and could only run one program at a time, but, for 1981, it was state of the art.

To drive sales, IBM commissioned developers to build a number of applications for the IBM PC. In addition to Microsoft Basic, these included a version of VisiCalc, an early spreadsheet that was a popular application for the Apple II, a word processor, Easy- Writer, and a well-known series of business programs from Peachtree Software. Introduced in August 1981, the IBM PC was an instant success. Over the next 2 years, IBM would sell more than 500,000 PCs, seiz- ing leadership from Apple, which had dominated the PC market with the Apple II. IBM had what Apple lacked—an ability to sell into corporate America. As sales of the IBM PC mounted, more independent software developers started to write programs to run on the IBM PC. These included two applications that drove adoptions of the IBM PC: word process- ing (WordStar and WordPerfect) and a spreadsheet (Lotus 1-2-3).

The success of IBM gave birth to clone manufac- turers who made “IBM-compatible” PCs that also uti- lized an Intel microprocessor and Microsoft’s MS-DOS operating system. The “clone” industry was born when engineers at Compaq Computer reverse engineered the BIOS chip in the original IBM PC. The BIOS chip converted the operating system into machine language, and was integral to the operation of the PC. It was the only key component of the IBM PC that IBM had not

02277_Case20_rev02.indd 153 01/10/15 6:03 PM

C-154 Case 20 Microsoft: From Gates to Satya Nadella

bought off the shelf from other manufacturers. Com- paq’s BIOS chip was functionally equivalent to the chip in the IBM PC, but used different code and thus did not violate IBM’s copyright. Other PC companies soon fol- lowed Compaq’s lead, including Tandy, Zenith, Lead- ing Edge, and Dell. The birth of the clone industry was a huge boon to Microsoft. By virtue of its nonexclusive license with IBM, Microsoft had the ability to sell MS- DOS to a growing number of clone makers.

In 1983, Microsoft expanded its product offering with the introduction of Word for MS-DOS, the com- pany’s first word processor. Word was differentiated from other word processors at the time by being the first to use a mouse. In 1985, Microsoft introduced a version of Word to run on Apple’s latest machine, the Macintosh. In 1985, Microsoft released the first version of Excel, the company’s spreadsheet offering, which competed with the bestselling Lotus 1-2-3. In 1987, Microsoft purchased a start-up company that had developed presentation software for the Macin- tosh. This product ultimately became PowerPoint, the first version of which was introduced in 1990.

The lead developer for Word and Excel was Charles Simonyi, a key hire at Microsoft who had for- merly worked at PARC, Xerox’s legendary research center, which had pioneered the development of the computer mouse, on-screen icons, a graphical user interface (GUI), object-oriented programming, and the laser printer. In a quirk of business history, senior management at Xerox had passed on the opportunity to commercialize these innovations, which opened the doors to Apple and Microsoft to pick up the ideas and run with them.

In 1982, with business booming, Paul Allen was diagnosed with Hodgkin’s lymphoma. His cancer was successfully treated with radiation therapy, but he took an extended leave and never again held an oper- ating position at Microsoft. In 2000, he resigned from the company’s board of directors.

BuIldInG the douBle Monopoly

By the mid-1980s, Microsoft was doing very well. It became apparent that the MS-DOS business had some compelling economics. While Microsoft bore

the costs of developing successive versions of MS- DOS, the incremental or marginal costs of producing individual copies of MS-DOS were very low. In the case of new PCs, Microsoft simply gave the master code to the manufacturer, who installed MS-DOS on every machine built, and paid Microsoft a licens- ing fee per machine. This resulted in gross margins as high as 90%. In contrast, the gross margins of PC makers at the time were closer to 40%.

the development of Windows and office In 1986, Microsoft went public. The IPO raised $61 million and valued Microsoft at $650 million. Micro- soft now had over 700 employees. The company’s po- sition, however, was not secure. Although MS-DOS was the most widely used operating system for PCs, Apple had shown what the future looked like in 1984 when it introduced the Macintosh. Borrowing many ideas from Xerox PARC, the Mac had a graphical user interface (GUI) which displayed programs as on screen icons. It also used a computer mouse with its point-and-click methodology for selecting tasks. This intuitive interface was a big improvement in usabil- ity over the clunky command-line interface of MS- DOS, which could be intimidating for people without a computing background.

Gates realized that a GUI interface was the future. Microsoft worked closely with Apple to develop the first version of Word for the Mac, which took full advantage of the Mac’s GUI interface and mouse capabilities. Word for the Mac soon became one of the bestselling Mac applications. At the same time, Microsoft took what it learnt from Apple and used it to start developing its own GUI interface, which was christened Windows.

Apple inadvertently helped Microsoft in two ways. First, it licensed its “visual displays” to Microsoft in 1985, enabling Microsoft to legally develop a GUI that had a similar look and feel to the Mac. Second, it was difficult to develop applications for the Mac. Apple did a poor job of providing tools to help third- party software developers write programs for it. In contrast, Gates often said that the most important strategic business unit at Microsoft was its tools busi- ness. Microsoft invested heavily in the development of tools to boost developer productivity. This made

02277_Case20_rev02.indd 154 01/10/15 6:03 PM

C-155Case 20 Microsoft: From Gates to Satya Nadella

it easy for third-party developers to write applications for MD-DOS, and later Windows, and drove adop- tion of Microsoft’s operating system offerings.

The first version of Windows was introduced in November 1985. It was a GUI shell that displayed programs as on-screen icons allowed for multitasking (using more than one program at a time). Windows sat on top of MS-DOS. It was commercial failure. Many users lacked sufficiently powerful hardware to run Windows, and there were few programs avail- able that took advantage of its features. Nevertheless, Microsoft continued development work on Windows.

IBM, too, saw the importance of a GUI interface. IBM was losing market share to the clone makers, so it decided to replace MS-DOS with its own GUI op- erating system, OS/2. IBM contracted with Microsoft to develop OS/2. However, the arrangement was a dif- ficult one. IBM resented the fact that Microsoft had facilitated the emergence of the clone businesses by licensing MS-DOS to IBM’s competitors. IBM was also concerned that Microsoft continued to work on Windows even while it developed OS/2. For its part, Microsoft knew that IBM was also investing in the UNIX operating systems, and had licensed a UNIX based PC operating system, NeXTSTEP, from NeXT, a PC company that Apple founder Steve Jobs estab- lished after he left Apple in 1985. Microsoft knew it would be in trouble if IBM scrapped OS/2 in favor of a UNIX alternative. The pivotal event was IBM’s announcement that it would release two versions of OS/2, a powerful version that would be exclusive to IBM machines, and a basic version for other PC mak- ers. That wasn’t news that Microsoft wanted to hear. Gates decided to sever links with IBM and go for broke on Windows.

The fruit of this effort, Windows 3.0, was intro- duced in 1990. Windows 3.0 was a big improvement over earlier versions. It was well reviewed and became a major commercial success. IBM’s OS/2, meanwhile, garnered mixed reviews and limited market traction. PC manufacturers, seeing a chance to deliver a body blow to IBM, which after all was a direct competitor, adopted Windows 3.0, bundling it with most new PCs. Market momentum toward Windows was also helped by the introduction versions of Microsoft’s increas- ingly popular applications products, Word, Excel, and PowerPoint, for Windows 3.0. At the time, each of these products was number 2 in its market space (Word was behind WordPerfect, Excel behind Lotus

1-2-3, and PowerPoint trailed Harvard Graphics). Microsoft’s rivals, however, were slow to introduce versions of their products for Windows, resulting in big market-share gains for Microsoft’s offerings. To further drive adoption of Windows, Microsoft re- doubled its efforts to provide developers with the best tools, and to persuade them that Windows was best platform for which to develop applications.

In 1992, Microsoft combined its three leading ap- plication programs—Word, Excel, and PowerPoint— into a single offering for Windows, which it called Office. Office was priced slightly below the com- bined price of each individual offering. Microsoft also promised interoperability between the three pro- grams, although this took several versions to perfect. Microsoft’s rivals, including most notably WordPer- fect and Lotus, lacked a comparable suite of offerings and were unable to match Office. From this point on, Office became the dominant suite of productivity pro- grams for information workers.

During the late 1980s, Microsoft started an operating system development project targeted primar- ily at servers. Servers were specialized PCs that sat at the heart of corporate networks of “client” PCs and “served” those “clients,” holding shared files and appli- cations programs used by many machines, such as email systems. Dubbed Windows NT, this was a powerful, 32-bit operating system that could run on servers. Un- like Windows 3.0, it was not DOS based (“NT” stood for new technology). To develop Windows NT, Microsoft hired a team of software developers led by Dave Cutler from Digital Equipment Corporation (DEC). Cutler’s team drew on their prior experience developing 32-bit systems for DEC to develop Windows NT.

The move into the server OS business represented recognition by Microsoft of the growing importance of client-server systems within large enterprises. The development of Windows NT constituted a strate- gic shift by Microsoft toward the enterprise market, where the primary demand for client-server systems resided. Windows NT was an attempt to make secure, stable software that could run “mission-critical” ap- plications within enterprises. Client-server networks were taking business away from the mainframes and minicomputers sold by the likes of IBM, DEC, and Hewlett Packard. Microsoft wanted a piece of this business, and with Windows NT it intended to get it. Introduced in 1993, Windows NT was a solid, sta- ble, secure system that gained increasing acceptance

02277_Case20_rev02.indd 155 01/10/15 6:03 PM

C-156 Case 20 Microsoft: From Gates to Satya Nadella

within enterprises. Windows NT marked the begin- ning of Microsoft’s server business.

To gain further enterprise business, Microsoft added an email client to its Office suite, Outlook, which could connect with corporate email hosted on severs. By the time Windows NT was introduced, Microsoft was also selling a relational database offer- ing, Microsoft SQL Server. A relational database is a product whose primary function is to store and re- trieve data as requested by other software applications, be they on the same computer or running on another computer across a network. Microsoft SQL Server was the company’s entry into the enterprise-level database market, and it pitted Microsoft against Oracle and IBM, both of which had relational database offerings.

The 32-bit technology underlying Windows NT was subsequently incorporated into the next two re- leases of Windows for PCs, Windows 95, and then Windows XP (introduced in 1995 and 1998, respec- tively). Increasingly, this made Windows more than just a GUI that sat on top of MS-DOS. Windows was becoming a fully-fledged operating system in its own right. By the time Windows 2000 was introduced, Windows had effectively shed it DOS heritage.

Windows 95 was a landmark release. Its enhanced graphics effectively closed the gap between Windows and Apple’s Macintosh. Since the introduction of the IBM PC, Apple had been a niche player in the PC business, focused primarily on the education, graph- ic artist, and desktop publishing markets, where its graphic displays and ease of use gave it maximum ad- vantage. With Windows 95, however, the differential appeal of the Mac all but vanished. By 1997, Apple was facing bankruptcy.

the internet tidal Wave One other event occurred during the 1990s that helped to cement the dominance of Microsoft: the explosive growth of the World Wide Web (WWW). Tim Bern- ers Lee, a British researcher at CERN in Europe, in- vented the Web during the early 1990s. The WWW sits on top of the Internet, which itself had been developed by American researchers during the 1960s and 1970s. As Berners Lee conceived it, the Web used hypertext markup language (HTML) and hypertext transfer pro- tocol (HTTP) to enable links to be made to information anywhere the Internet, thereby creating an enormous “web” of information. In 1993, a team at the University

of Illinois led by a 22-year-old student, Marc Andrees- sen, developed the Mosaic Web browser. Mosaic could display information on the Web graphically. This was the beginning of the enormous growth of the WWW. After graduation, Andreessen joined up with Jim Clark, the former CEO of Silicon Graphics, to form Netscape. Netscape further developed the Mosaic Browser, releasing its version, Netscape Navigator, in November 1994. Netscape Navigator quickly became the domi- nant Web browser. In August 1995, Netscape held an IPO. The stock was offered at $28 a share, but closed its first day at $75, valuing Netscape at $2.9 billion.

Prior to the explosive growth of the Web, Microsoft’s Internet strategy involved the creation of a dial-up online service, MSN, which was developed to be included with Windows 95. MSN was similar in conception to early versions of AOL, with email capa- bilities, message boards, chat rooms, and some news and weather offerings. The first version of MSN did not have a Web browser and users could not connect to the Internet. With MSN and Windows 95 in late development, Gates became aware of the rapid growth of the Web. Microsoft legend has it that the WWW was brought to the attention of Gates by memos from two junior engineers, Steve Sinofsky and Jay Allard. Gates immediately saw its strategic significance. In May 1995, Gates wrote a memo to his executive staff and direct reports, calling the growth of the Internet a “tidal wave.” Gates wrote that the Internet “is crucial to every part of our business” and “the most impor- tant single development to come along since the IBM PC was introduced in 1981”. In his memo, Gates went on to say that Netscape was a “new competitor”, and that Microsoft’s strategy should be to make it clear that “Windows machines are the best choice for the Internet.”11

To fulfill Gates’s vision, Microsoft acted rapidly. It licensed a version of the Mosaic Web browser from a company called Spyglass, improved on it and released it as Internet Explorer (IE) version 1.0 in August 1995. IE 1.0 was bundled with Windows 95 and appeared as an icon on the start screen. Although it was too late to change MSN in time for the release on Windows 95, MSN was reworked to utilize HTML and HTTP and give users access to the Web. In late 1996, the new ver- sion of MSN, MSN 2.0, was released. Microsoft also quickly added the ability to insert hypertext links into Office documents, allowing readers of those docu- ments to navigate away to websites.

02277_Case20_rev02.indd 156 01/10/15 6:03 PM

C-157Case 20 Microsoft: From Gates to Satya Nadella

antitrust issues All of these moves were successful for Microsoft. However, the bundling of IE with Windows bought Microsoft to the attention of the U.S. Department of Justice (DOJ). The DOJ argued that the bundling strategy put Netscape at a competitive disadvantage and was a deliberate attempt on Microsoft’s part to “squash” their rival. Whereas Netscape charged con- sumers for their browser, IE was perceived as being a “free” product. Moreover, the DOJ contented that Microsoft configured the Windows code such that it was slow and difficult for users to download Netscape Navigator and install it on the Windows desktop. For its part, Microsoft claimed that IE was part of the op- erating system and that users expected it to be there.

In the end, the DOJ prevailed. The judge in the case ruled that Microsoft was a monopoly, and that the bundling strategy represented an abuse of Micro- soft’s monopoly power. In 2002, Microsoft and the DOJ reached a settlement that required Microsoft to share its application programming interfaces (APIs) with third-party companies, so that they could write programs that worked well with Windows. Micro- soft, however, was allowed to continue bundling IE and other products with Windows. For Netscape, this was a Pyrrhic victory. The company continued to lose market share against IE, and was not helped by re- ports that its products were inferior in quality to IE. In 1999, Netscape was sold to AOL for $10 billion, a price tag that left many scratching their heads. AOL discontinued the Netscape browser in 2008. At the time it had less than a 1% share of the browser mar- ket, down from over 90% in 1995.

ManaGInG the coMpany

From the outset, Gates made a point of hiring people who were like him—young, bright, driven, competi- tive, technically sharp, and able to argue effectively for what they believed in. A small but influential num- ber of these hires came from Xerox PARC, including Charles Simonyi, who led the development of the first versions of Word and Excel. Ballmer hired some of sales personnel. One of these was an aggressive sales- man named Vern Raeburn. Gates had insisted that

Microsoft should not sell directly to end-users, but Raeburn marshaled his arguments and persuaded Gates to change course. Raeburn quickly pulled to- gether a team to market and sell Microsoft’s products to consumers.

This was the genesis of a split within Microsoft into two distinct functions that persist to this day: an engineering function that develops products, and a sales and marketing function that sells them. For years, Gates was the de facto head of engineering with responsibility for product development, where- as Ballmer was responsible for sales and marketing. Although Microsoft went on to create different busi- ness units—Windows, Windows Server & Tools, and Office all had their own business units, for example— the engineering and sales and marketing functions would cut across these units, creating a loose, matrix organization. Finance, legal, and human relations functions also cut across business units.

To motivate key employees and encourage them to work long hours and commit to the company, Mi- crosoft gave them stock options. When the company did well, and the stock price rose, these employees made substantial sums of money. As the stock price surged after the IPO in 1986, Microsoft stock options became a major draw, enabling Microsoft to hire the best and the brightest. By 2000, it is estimated that the surging stock price had created over 10,000 million- aires among Microsoft employees.12 Paradoxically, by the mid-1990s, some early employees were so se- cure financially that their competitive edge had been blunted. Some were said to have retired on the job. Many other key employees simply left the company to pursue other interests.

Another notable feature of Microsoft that emerged over time was the tendency for people to circulate within the company. It was not unusual for people to change jobs every 18 months, and move from business to business.

formalizing Management Processes As the company’s growth began to accelerate in the early 1980s, Gates brought in people with business experience to help take the load of his shoulders and manage the day-to-day operations and finance side of the business, leaving him to focus on product develop- ment, technology, and strategy, and Ballmer to focus on sales. A key early hire was John Shirley. Shirley

02277_Case20_rev02.indd 157 01/10/15 6:03 PM

C-158 Case 20 Microsoft: From Gates to Satya Nadella

worked for Tandy Corporation, the parent company of Radio Shack. Shirley joined Microsoft as president in 1983 and stayed through 1990. He remained on the board until 2008. People within the company would joke that Shirley was there to provide some adult su- pervision.

In 1994, Gates hired Bob Herbold as chief operat- ing officer (COO). Herbold had a PhD in computer science and had worked at Procter & Gamble for 25 years, where he was responsible for P&Gs worldwide marketing and brand management. Herbold stayed at Microsoft until 2001. Another “adult” in charge of day-to-day operational issues, Herbold saw it as his job to bring discipline to the company without under- mining the characteristics that had made it competi- tive. Herbold describes arriving at a company that was chaotic: “Incompatible systems and divergent practic- es companywide were causing all kinds of problems. Bills from suppliers weren’t being paid on time. We never knew precisely how many people worked for the company. Business units set projections using incom- patible frameworks and measures that prevented a comparison of their performance.”13

Much of this chaos was the result of rapid rev- enue growth often exceeding 30% a year. Herbold notes “a balkanized system had grown up because, for years, Bill had focused on product development and Steve had focused on sales. Meanwhile, business and geographical units had relatively free rein to cre- ate local functional staffs, set business practices, and build stand-alone information systems. They weren’t particularly interested in giving up their autonomy.”14

Herbold moved fast to standardized, basic business- es processes at Microsoft, including financial reporting, vendor payments, and human resources policies. He also found a company with no formal strategic plan- ning process in place. Herbold developed a rolling, 3-year planning process based on a standardized format that included historic and future projections of market share, revenues, costs, and profits. The process distin- guished between established products, such as Windows and Office, and new products where there was a much greater degree of uncertainty. The plans were modified and streamlined every year based on new data.

Herbold also formalized a human resources per- formance-appraisal process that had originally been developed by Gates. The appraisal process required managers to evaluate their direct reports, and it uti- lized a forced curve, such that some members of a

team would always end up being classified as star performers, and others as poor performers. The star performers would get big pay increases, whereas the underperformers would be “encouraged to find a job outside of the company” if they couldn’t bring up their rating over time. Critics of this system, known as stack ranking, noted that it pitted employees on a team against each other, encouraged backstab- bing, and created a real problem for managers who had built strong teams, because they were forced to classify some of their team as underperformers, even though in an absolute sense they might be good.15

the Product development Process Given the nature of Microsoft’s business, a key as- pect of the company’s organization and management structure relates to the way it formalizes development of its software products. In the early years “superpro- grammers” such as Simonyi and Gates drove the vi- sion for products. Gates came to the realization that this model would not scale well. Superprogrammers were in short supply, had little interest in updating a product once it had been created, might not under- stand the market well, and were prone to clash with other superprogrammers. In response a formal sys- tem for developing, testing, and releasing products emerged in the mid-980s.16

The process starts with a program manager, who is responsible for specifying the vision of the product, its key features, development schedule, development process, and implementation tradeoffs. The program manager works closely with senior software devel- opers and with product managers in marketing to achieve all of this. His or her role, in other words, is to coordinate engineering and marketing and distill out of this what the product should do, what its key features should be, and a schedule to achieve these. The program manger is then responsible for manag- ing the overall development effort, and must make the call on features to add or cut in order to hit goals such as schedule. On complex products such as Windows and Office there is a hierarchy of program managers. For example, while there may be an overall program manager for a new version of Office, there will also be program managers for each constituent program— Word, Excel, PowerPoint, and so on.

It is important to understand that many of the ideas for a product’s features come from developers

02277_Case20_rev02.indd 158 01/10/15 6:03 PM

C-159Case 20 Microsoft: From Gates to Satya Nadella

and marketers. Program managers are leaders and fa- cilitators of the process, rather than bosses, and they must work through persuasion and negotiation. In part, this may be due to the high status that developers in particular have within Microsoft’s culture, some- thing that can be traced back to Gates and Simonyi. Indeed, most program managers were themselves star developers who rose through the engineering ranks.

Once the product vision, key features, schedules, and the like have been mapped out, it is up to soft- ware developers to implement the vision and features. Developers write the code. Typically, a small team of senior developers and program managers will take charge of the product architecture, and developer leads (first-line managers) will provide detailed guid- ance to their teams of programmers. While developers may be the source of ideas for new features, they are required to clarify what each feature accomplishes, and to help program managers decide what to in- clude in a product, and what to cut in order to stay on schedule.

Testing the code is the responsibility of developers and testers. Developers are meant to test their own code frequently (typically every day). They also work in pairs with testers and are required to hand their code over to a tester for testing before adding their work to the “official build.” The goal in this process is to reduce the bug count to zero. Microsoft also has a specially trained group of people who perform final tests on a completed product to see if it is ready for shipment. As part of this process, beta versions of the product will be released to key customers for feedback, and the product will be tested in a usabil- ity lab. Microsoft has approximately one tester for every developer, an unusually high ratio but one that is consistent with the goal of producing stable, secure software that can run mission-critical applications for enterprises.

Over time, Microsoft routinized this process, with offering such as Windows, Office, and SQL Server go- ing through 3- to 5-year definition, development, test, and release schedules. As these products have grown in complexity and features, there was a tendency for the process to become more bureaucratic and harder to manage. This was made more challenging by the fact that many program managers, senior develop- ers, and development leads were people who had ex- celled in a technical sense but had little management training or experience. In the mid-2000s, this led to

serious issues when Microsoft ran over budget and over schedule while trying to develop Windows Vista (discussed later in the case).

the BallMer years

When Bill Gates handed the CEO role over to Steve Ballmer in February 2000, the company was at the top of its game. Windows and Office dominated their respective markets, generating prodigious amounts of free cash flow. The stock price had hit an all-time high of $58.72 on December 23, 1999. Microsoft was the most valuable company on the planet, and Gates the world’s richest man. Gates continued to work full time at Microsoft until 2008, assuming the role of chief software architect, with primary oversight for product development. He also remained chairman of the company.

During the Ballmer era, revenues increased 280%, to $70 billion, while net profit expanded by 215%, to $23 billion. The stock price, however, dropped below $40 a share in mid-2000, and did not break through that level again until 2014, after Ballmer had resigned. The failure of the stock price to advance despite grow- ing top and bottom lines reflected a widely held belief among investors that Microsoft had lost its leadership position in the industry. Moreover, critics believed that the company was destroying economic value by investing in businesses that did not generate a positive ROI. These included the Xbox videogame business, Internet search, and the device businesses that encom- passed the Zune music player, smartphones, and tab- let computers. By the end of the Ballmer era, it was widely believed that the shift to a world characterized by mobile devices and cloud computing presented an existential threat to Microsoft’s core operating system business.

One of the first problems that Ballmer had to con- front was the risk of a slowdown in the rate of growth of both Windows and Office. The markets for both products were now mature in most developed nations, implying that revenues would increasingly come from replacement rather than first-time demand. Although there was still plenty of room for growth in develop- ing nations, those markets were also characterized by extremely high levels of piracy—as much as 90% in markets like China and Vietnam. Indeed, even in

02277_Case20_rev02.indd 159 01/10/15 6:03 PM

C-160 Case 20 Microsoft: From Gates to Satya Nadella

developed markets such as the United States, piracy rates for software products are as high as 20 to 25%.17

Two trends helped Microsoft weather the matu- ration of its two primary product offerings. First, a significant number of consumers in developed mar- kets purchased multiple devices: a laptop and a desk- top for example. Second, Microsoft continued to grow its share of the enterprise markets for Windows Server and SQL Server, taking business from UNIX, LINUX, Oracle, and IBM. Microsoft’s success in the enterprise space reflected the fact that to a consider- able extent, the company had succeeded in building stable, secure software that could run mission-critical applications in enterprises. Given that Windows for the client and Office were also widely used within en- terprises, Microsoft was increasingly focused on its enterprise business. Indeed, by the early 2000s, Mi- crosoft was more of an enterprise company than a consumer company.

Product diversification: Xbox Under Ballmer, Microsoft continued to diversify its product offerings, entering into new markets. The first was the videogame market. By the late 1990s, Sony dominated this market with its PlayStation console and related game offerings. The market was worth $20 billion globally and was growing. Microsoft saw the PlayStation as a threat. The PlayStation was a specialized computer that ran a non-Microsoft oper- ating system, and could theoretically be connected to the Internet via a TV cable. Moreover, the PlayStation was often located in the living room. Bill Gates had long dreamed of having Internet-enabled computing devices in the living room that operated interactive TV, and could also be used for Web browsing, playing games, and online shopping; but Gates wanted those devices to run Windows.

Microsoft had capabilities that persuaded man- agement that the gaming market was a viable target. Microsoft had produced one of the bestselling PC games of all time—Microsoft Flight Simulator—and had published another, Age of Empires. Through MSN it also had the world’s largest online gaming site, MSN Gaming Zone, which had 12 million sub- scribers in the early 2000s. Moreover, Microsoft in- tended to use a customized version of the Windows operating system to power Xbox. This would save de- velopment costs and make it easier for developers to

write games for the Xbox, because many were already familiar with Windows programming APIs and tools.

Microsoft lacked the ability to produce hardware, so it decided to outsource this to a contract manu- facturer, Flextronics. Microsoft’s strategy was to price the Xbox at or below cost to drive adoption, and then make money on the sales of games, either directly in the case of games developed in house, or from royalty fees in the case of games developed by third parties. For this strategy to work, it had to guarantee Flextronics a profit margin, which meant paying Flextronics a sub- sidy on every machine manufactured.

Xbox was introduced in late 2001, after $1.5 bil- lion in development costs. The company faced tough competition from Sony’s new offering, PlayStation 2 (PS2). To drive adoption, it cut prices for hardware ag- gressively. By 2003, Microsoft was thought to be losing $100 on every Xbox it sold. To make that back and turn a profit, Microsoft reportedly had to sell six to nine games per Xbox.18 By late 2004, Xbox was still a distant second to PS2 in the videogame market, hav- ing sold 14 million consoles against Sony’s 70 million. While Sony was making good money from the business, Microsoft was registering losses. Microsoft’s home and entertainment division, of which Xbox was a part, lost $4 billion between the launch of Xbox and mid-2006.

In November 2005, Microsoft introduced its next- generation console, Xbox 360. Again, contract manu- facturers made the machine, and again Microsoft paid them a subsidy to ensure their profit margins. Sony fol- lowed a year later with its PS3 console, as did Ninten- do with the Wii console. The Wii was a less powerful machine than either Xbox 360 or PS3, but it came with a motion sensor controller than changed the way play- ers interacted with games. The Wii bought a new gen- eration of casual gamers into the market and turned into a surprise hit for Nintendo. Meanwhile, Micro- soft and Sony slugged it out in the hard core gaming market. Demand for Xbox was helped by Microsoft’s enormously popular Halo franchise. As the market ex- panded, all three companies were able to make profit on an operating basis in the business. However, both Microsoft and Sony hurt themselves with quality problems and component shortages early in the prod- uct cycle (Microsoft had to take a $1.05 billion write off in 2007 for replacing poor-quality consoles).

Although Microsoft did achieve profitability on an operating basis for Xbox business by late in the Xbox 360 cycle, on a cumulative basis the return on

02277_Case20_rev02.indd 160 01/10/15 6:03 PM

C-161Case 20 Microsoft: From Gates to Satya Nadella

investment was still believed to be negative. One bright spot for Microsoft was the growth of its online game subscription service, Xbox Live. Introduced in 2002, by mid-2013, Xbox Live had around 45 million paying sub- scribers who used it for everything from playing multi- player games online to streaming movies from Netflix and browsing Facebook. At the time, Microsoft was thought to be generating annual revenues in excess of $3 billion from Xbox Live.19 Microsoft also garnered strong reviews and sales for its Kinect motion sensor controller. Introduced in late 2010 for the Xbox 360, Kinect was developed as a response to Nintendo’s Wii controller.

In late 2013, Microsoft launched its third-gener- ation game console, Xbox One. Sony matched with the launch of its PS4 system. At launch, Microsoft positioned Xbox One as an all-purpose entertainment system for the living room, controlling TV, music, and film streaming services through the Kinect mo- tion and voice sensor, in addition to being a game console. Sony focused its marketing for the PS4 on the core gaming market. By mid-2014, Sony was believed to have sold 7 million PS4 consoles, versus 5 million Xbox One consoles. With Satya Nadella now in charge, Microsoft changed the marketing strategy for Xbox One, emphasizing its capabilities as a gam- ing machine and co- promoting it with new iterations of its popular Halo and Call of Duty franchises.

Product diversification: internet search Another hallmark of the Ballmer era was Microsoft’s expansion into Internet search. Microsoft had long had primitive Internet search functionality on its MSN service, but it had never seen search as a cen- tral feature. This changed with the rise of Google, a company that didn’t even exist until 1998. At the core of Google’s rise was a search algorithm that cleverly ranked the relevance of a page for a search query ac- cording to the number of pages that linked into that page. Google went to great lengths to make sure that its search results were “pure.” It did not mix organic and paid search results, thereby improving relevance to the user (paid search results were originally placed on the right hand side of a search page, separate from organic search results).

What made Google a valuable company was its combination of highly relevant search results with a business model that made money out of search activity—lots of money. This was the ‘pay-for-click”

model, where advertisers paid Google every time someone clicked on an advertiser’s link. From a standing start in 2001, by 2014 Google had grown into a colossus with $68 billion in revenues, almost $21 billion in net profits, 67% of the market for Inter- net search in the United States, and an estimated 70% of worldwide search marketing spend.

Along the way, Google had moved aggressively into Microsoft’s turf. Reasoning that with the growth of smartphones, ever more search would come through mobile devices, Google had pushed into the smart- phone business with its Android operating system, which it licensed to hardware manufacturers for free. The economic logic was that Google would be the de- fault search engine on Android phones, so every time someone search for something on an Android phone, and clicked on an advertising link, Google would make money. Google also developed its own Web browser, Chrome, which it distributed for free. The economic reasoning was similar. Since search is conducted with- in a Web-browser environment, and Google was the default search engine on Chrome, Google would cap- ture more search based advertising collars if its own browser were widely used. Both of these products were phenomenally successful. By mid 2014, Android was found on 85% of the world’s smartphones, and Chrome was the browser of choice on 46% of all desktops and tablets (relegating Microsoft’s Internet Explorer, the long time market leader, to second place with 20%).20

Microsoft tried to counter Google’s rise in the Internet search business, but its success was limited,— and very expensive. Microsoft adopted Google’s pay for click search model, and developed a similar search algorithm, but was unable to gain much market trac- tion and its market share remained stuck under 10%. Part of the problem was brand confusion. Microsoft’s search feature was initially known as MSN Search, sounded dull and uninspiring next to Google. In 2006, MSN search was rebranded as Windows Live Search, and given some new features. A year later, the name was changed again to Live Search. Ultimately, Micro- soft came to the realization that the “Live” brand was not resonating with consumers, who found it confus- ing. In June 2009, Microsoft’s search engine was re- branded Bing. Microsoft supported the Bing launch with a $100-million ad campaign.

In 2008, in an attempt to grow its share of the U.S. paid-search market, Microsoft launched an unsolic- ited takeover bid for Yahoo. Yahoo was number 2 in

02277_Case20_rev02.indd 161 01/10/15 6:03 PM

C-162 Case 20 Microsoft: From Gates to Satya Nadella

the US search market. Microsoft was number 3. Ya- hoo rejected the bid. A year later, however, following management changes at Yahoo, Microsoft and Yahoo entered into a 10-year partnership under which Bing would be the exclusive search platform on Yahoo. Although the precise terms of the deal were not made public, it is known that Microsoft pays Yahoo for search traffic. In 2013, 31% of Yahoo’s revenue ap- parently came from Microsoft payments.21

Product diversification: smartphones and tablets Microsoft was an early leader in the smartphone busi- ness. It first offered an operating system for smart- phones, Windows Mobile, in 2002. By 2007, 42% of all smartphones used the Windows Mobile operating system. Smartphone manufactures such as Motoro- la and HTC paid a licensing fee to Microsoft to use Windows Mobile. As was normal at the time, Win- dows Mobile powered smartphones had a physical keyboard and a small screen. The devices were pri- marily sold to enterprise customers, who used the phones for email, appointments, text messaging, and Web browsing.

In 2007, Apple introduced the first iPhone, which revolutionized the smartphone market and signifi- cantly expanded demand (see Table 1). The com- bination of a touch screen, virtual keyboard, larger

screen size, elegant design, and ease of use made the iPhone a huge hit in the consumer marketplace. Busi- ness people too, bought iPhones in droves, leading many companies to adopt a policy of “bring your own device” with regard to smartphones. Growth of the iPhone got a further boost from the development of third-party applications and the opening of the Apple App store in 2008, which made it easy for us- ers to find and download apps onto their phones. The supply of Apps was facilitated by efforts on Apple’s part to make it easy for third-party developers to write Apps for the iPhone. In 2010, Apple introduced its tablet offering, the iPad. The iPad used the same iOS operating system as the iPhone and had most of the same attributes, including elegant design, a touch screen, and access to the App store through wireless connectivity. All of this helped drive rapid growth in consumer demand.

When Apple released the iPhone, Google already had its own operating systems for a touch screen phone in development. Google had acquired the original de- veloper, Android Inc., in 2005. The first smartphones running on Android appeared in 2008. Google’s busi- ness model was to offer Android for free and make money from advertising linked to mobile search. By 2013, Android was the dominant smartphone OS, fol- lowed by Apple’s iOS (see Table 1). Tablets that ran on Android stated to appear soon after the launch of the iPad in 2010, and by 2014 Android was also dominat- ing the tablet OS market (see Table 2).

year android ios (apple) Microsoft BlackBerry nokia

2007 3 15 12 78

2008 11 17 23 73

2009 7 25 15 34 81

2010 67 47 12 47 112

2011 220 89 9 52 93

2012 451 130 17 34 0

2013 759 151 31 19 0

source: Gartner.com, various press releases.

table 1 Global smartphone sales (millions) 2007–2013

02277_Case20_rev02.indd 162 01/10/15 6:03 PM

C-163Case 20 Microsoft: From Gates to Satya Nadella

The introduction of the iPhone, and then Android phones, decimated Microsoft’s market share (see Table 1). By 2011, Microsoft’s OS was found on just 9 million smartphones shipped that year. Android was on 220 million phones, and Apple sold 89 million iOS phones. The situation in the tablet market was no better, where Microsoft was caught completely flat- footed by the introduction of the iPad. Google, on the other hand, adapted very quickly and soon gained market leadership.

In response to the rapid emergence of Apple and Android, Microsoft developed a new operating sys- tem for touch screen smartphones, Windows Phone. Windows Phone had an active-tile, “Metro” interface, rather than the on-screen icons used by Apple and Android. The first Windows Phones started to appear in late 2010. Microsoft also established its own App store, the Windows Phone Store.

In early 2011, Microsoft entered into an alliance with Nokia to jointly develop Windows Phones. Like Microsoft, Nokia had been caught off guard by the emergence of the iPhone and had seen its market share slide. Nokia had used its own Symbian operat- ing system in its smartphones. Like Windows Mobile, Symbian was a primitive, first-generation smartphone OS that lacked the full features and functions of Android and iOS, including touch screen capability, a virtual keyboard, and a supply of third-party apps that could be downloaded onto the device. Under the alliance, Nokia agreed to phase out Symbian and switch to Microsoft’s Windows Phone OS. The first products of this alliance, Nokia’s Lumina phones, were introduced in late 2011.

Despite some favorable reviews, the Lumina phones grew more slowly than the market, and

Microsoft’s market share remained in the low sin- gle digits in most nations. Reasons given to explain this included the lack of appeal of the Metro in- terface and relative paucity of third-party apps for Windows Phone.

In September 2013, Microsoft announced its in- tention to acquire Nokia’s mobile phone business for $7 billion. In justifying the acquisition, Ballmer argued that merging the two companies would streamline product development processes, lower costs, and re- sult in better phones and higher gross profit margins.22 It was also noted that Nokia was the only company still willing to make Windows phones, and if Nokia pulled out, what would happen to Microsoft’s phone business? Critics wondered whether an acquisition that made Microsoft a phone maker might not alien- ate other phone makers, such as HTC, who would now see Microsoft as a direct competitor.

In addition to the phone business, Microsoft entered the tablet business with its Surface offering. The Surface was positioned as a cross over between a conventional laptop and a tablet. Introduced in late 2012, it used a Windows 8.1 operating system, which by that time was also being used for Windows Phone. Like the Windows Phone, the Surface garnered some favorable reviews, but sales were slow to pick up and the product initially failed to make a dent in the dominance of Android and iOS in the tablet market. However, following the introduction of the Surface Pro 3 in mid 2014, sales ap- peared to be accelerating. In the last 6 months of 2014, Microsoft sold $2 billion worth of Surface tablets.

Windows offerings Under Ballmer Windows Vista was the first version of Windows com- pletely developed under Ballmer’s leadership (although Bill Gates oversaw the project). Vista started out as a more ambitious project with the code name of Long- horn, but when that ran into difficulties, it was recast as Vista. A primary goal for Vista was to increase secu- rity. Released in January 2007, more than 5 years after its predecessor, Windows XP, it was not well received by the marketplace. Vista took 2 years longer than ex- pected to develop and it was several billion dollars over budget. It was a huge program—with 50 million plus lines of code—and utilized a lot of computer memory to run, resulting in unacceptably slow performance for many users. It quickly drained battery life on laptop

year android ios (apple) Microsoft

2011 17 40 0

2012 53 61 1

2013 121 70 4

2014 160 65 11 source: Gartner.com, various press releases.

table 2 Global tablet sales (millions) 2010–2014

02277_Case20_rev02.indd 163 01/10/15 6:03 PM

C-164 Case 20 Microsoft: From Gates to Satya Nadella

computers. Moreover, it irritated users with constant popup authorization prompts for user account control. Many potential adopters simply stuck with Windows XP rather than switch to the much-maligned Vista. By October 2009, Windows Vista had 19% of the PC operating system marketplace, while Windows XP, an 8-year-old OS, still enjoyed a 63% share.

Many insiders blamed the poor performance of Vista on a development process that got out of hand. One problem was “too many VPs in reporting struc- tures too narrow.” There were 12 layers of manage- ment between Bill Gates and a developer at the base of the Windows organization. As one former Win- dows development lead noted:

“I once sat in a scheduled review meeting with at least six VPs and ten general manag- ers. When that many people have a say, things get confusing. Not to mention, since so many bosses are in the room, there are often negotia- tions between project managers prior to such meetings to make sure no one looks bad . . . In general, Windows suffers from a proclivity for action control, not results control. Instead of clearly stating desired outcomes, there is a penchant for telling people exactly what steps they must take.”23

Other insiders complained about a lack of ac- countability, constant churning of features and speci- fications, with new features often being added without adequate testing, leading to system crashes and fur- ther development delays. Several people also noted that with Gates heading Vista, CEO Steve Ballmer was unwilling to step in and resolve problems that were resulting in delays and cost overruns.

Once Vista shipped, many of the top engineers on the project retired. Steve Sinofsky, who had been running Office, was bought in the run Windows. At Office, Sinofsky had run a very tight ship, releasing new versions on schedule like clockwork. The Office organization was also much flatter than the Windows division, with only four levels of management.

Sinofsky flattened the Windows organization, re- ducing the number of levels of management from 12 to 5. He pushed developers to get Vista’s successor, Windows 7, to market quickly. Originally conceived as an incremental update to Vista, Windows 7 was a more streamlined program that fixed many of the perfor- mance problems and irritations with Vista. Introduced

in 2009, reviewers saw Windows 7 as a big improve- ment over Vista, and the operating system sold well.

Once Windows 7 shipped, Sinofsky and his team turned their attention to Windows 8. Released in 2012, Windows 8 was positioned as an operating system for the new era of digital devices. Windows 8 used the same Metro-style, tile-based interface that had first been used on Windows Phone. Despite the lukewarm reception of the Metro interface on the phone, Microsoft saw the interface as an important differentiator. Sinofsky was a major advocate of the Metro interface, going so far as to push Microsoft to kill a competing interface for tablets being devel- oped within the company because it was inconsistent with the Metro theme that he wanted on all devices.24 Known as the Courier, the tablet was the brainchild of a group within Microsoft’s Entertainment and Devices Division, headed by Jay Allard. The Courier was on track to hit the market in 2010, just months behind the iPad. Widely admired within Microsoft, Allard was the force behind the creation of the Xbox business and was instrumental in pushing Microsoft to embrace the Web back in 1995.

The Courier was a two-screen tablet that folded like a book and had a touch screens. Early prototypes had elicited rave reviews from outsiders who had seen it, some preferring what they saw to prototypes of the iPad, which was then under development. But the Courier used a modified version of Windows as its operating system, and the interface departed substan- tially from the Windows norm.

When the Courier dispute surfaced, Ballmer found himself in the position of having to choose be- tween two of his best managers. Unable to make up his mind, he brought Gates into the decision. Gates, who by now had given up all day-to-day operating re- sponsibility, met with Allard and his team. His criti- cism was that Courier didn’t align with Microsoft’s key Windows and Office franchises. Not only did it use a customized version of Windows, and a nonstan- dard interface, but it also did not include an Outlook email client (Allard pointed out that users could get email through an onboard Web browser). For Gates, this was a fatal flaw, and the Courier was cancelled. Within months, Allard had left Microsoft along with his boss, Robbie Bach. It would be another 2 years before Microsoft had a tablet offering, the Surface.

In addition to the Metro interface, Windows 8 also supported touch screen technology and could be used

02277_Case20_rev02.indd 164 01/10/15 6:03 PM

C-165Case 20 Microsoft: From Gates to Satya Nadella

on a tablet in addition to desktop and laptop PCs. Released in 2012, Windows 8 received decidedly mixed reviews. Although reaction toward its perfor- mance improvements, security enhancements, and improved support for touchscreen devices was posi- tive, the new user interface of the operating system was widely criticized for being potentially confusing and difficult to learn. Many users particularly disliked the fact that Microsoft had removed the start menu.

Market take-up of Windows 8 was slower than Mi- crosoft had hoped. Sinofsky abruptly left Microsoft in December 2012. Recognizing that the Metro interface was not resonating with many users, Microsoft an- nounced that it would release an update, Windows 8.1, in October 2013. Windows 8.1 tried to address some of the criticisms, and gave users to ability to dispense with the Metro interface and revert to the traditional start button and menu. Despite this, adoption con- tinued to be slow. By mid-2014, only 12.5% of PCs were using Windows 8 or 8.1. Most consumers and corporation stuck with Windows 7.

the cloud computing initiative By the mid-2000s, serious conversations about cloud computing were taking place within Microsoft. The “cloud” referred to the idea that data, operating systems, and applications could be hosted on server farms comprising of thousands of machines, rather than on servers and PCs within an enterprise. These conversations were based on a realization that in a world where computing device users were always con- nected to the Internet through wired or wireless links, there were compelling economic reasons for moving computing power and programs off servers located within enterprises and onto server farms. Specifically, the cloud could deliver more value to users at a lower cost than traditional client-server networks.

On the value side, there was clearly great utility as- sociated with storing files on the cloud and being able to access those files anywhere anytime through any connected device. The files could be in the form of documents, music, video, or databases. Moreover, by reaching out and accessing programs and data stored on the cloud, users with simple devices such as smart- phones and tablets could theoretically access vast amounts of computing power when they needed it.

On the cost side, it was apparent that moving com- puting resources onto the cloud could save businesses

a lot of money. Corporate IT departments tradition- ally shouldered the costs of buying and maintain- ing computer hardware and software, activities that accounted for almost 90% of all IT costs. Servers, however, often only ran at 5 to 10% capacity, while much of the software installed on corporate servers and PCs was only rarely used. By moving data and applications onto a server farm, demand for comput- ing resources could be aggregated and servers could be run at closer to 90% capacity. This implied signifi- cant economies of scale in the costs of computation. Microsoft’s estimates suggested that, under optimal conditions, shifting to a cloud-based model could re- duce IT costs by as much as 80%.25 Moreover, instead of paying for software that was rarely used, corpora- tions might be able to pay for software only when they used it.

Microsoft proposed to build a cloud-computing business that would host data and applications for corporations, taking the burden of infrastructure and maintenance costs off their hands. In return, corpora- tions would pay a fee for storing data, and either a sub- scription or runtime fee for executing applications. As early as 2006, Steve Ballmer had stated that Microsoft had no choice but to go “all in” on the cloud.26 By 2010, this commitment had developed into Microsoft’s Azure cloud-computing initiative, which was located within the company’s Server & Tools Division.

Cloud computing was seen as comprising of three segments; infrastructure as a service (IaaS), platform as a service (PaaS), and software as a service (SaaS). IaaS refers to basic hosting of data, websites, and the like. Amazon’s AWS is primarily an IaaS offering. PaaS refers to the idea of building a software platform upon which software applications can be built and run. Microsoft’s Azure platform is essentially a PaaS offering that uses Windows Server technology. Think of Azure as Windows for a server farm of 10,000 ma- chines. SaaS is the idea that software applications can be hosted and run on the cloud. Salesforce.com was an early leader in SaaS with its customer relationship management (CRM) software.

By 2011, Microsoft was committed to competing in all three segments. The company would host data for enterprises and consumers (IaaS), it would con- tinue to develop Azure so that enterprises could write applications that would run well on the cloud (PaaS), and it would reposition many of its products such as Office, SQL Server, and Dynamics, as software as a

02277_Case20_rev02.indd 165 01/10/15 6:03 PM

C-166 Case 20 Microsoft: From Gates to Satya Nadella

service (SaaS) offerings. In June 2011, Microsoft in- troduced Office 365 to enterprise users. Office 365 was a cloud version of its bestselling Office suite. In 2013, Office 365 was offered to consumers. Enterprises paid a licensing fee and consumers an annual subscription fee for Office 365. Users could download the pro- gram to multiple devices (for consumers the limit was five). They could also store Office documents on Mi- crosoft’s Cloud using its One Drive storage offering. Microsoft shifted to a rolling—release model for de- veloping Office 365, updating the program on a quar- terly basis—a marked departure from the historic 3- to 5-year development schedule at Microsoft.

One problem Microsoft had to grapple with in shifting toward a cloud-computing model was that it represented a change in the underlying econom- ics of its business. In the traditional model, most of Microsoft’s costs were associated with the fixed costs of developing programs such as Windows and Of- fice. The marginal costs of producing more versions of a program were very low, so that at high volumes Microsoft earned gross margins in the 90% range on Windows and Office. In the cloud-computing model, however, Microsoft had to build and maintain server farms, which could cost anywhere from $500 million to $2 billion each in fixed costs, and which consumed large amounts of electricity. There was a general be- lief that even at high volumes, the gross margins as- sociated with a cloud-computing business would be significantly lower than what Microsoft was accus- tomed to. As people within the company were fond of saying, “in the cloud business we actually have costs of goods sold.”

On the other hand, while cloud computing was still a small business in 2013, generating perhaps less than $10 billion in revenues industrywide, rapid growth was predicted going forward. Industry revenues were projected to balloon to $150 billion by 2020. Clearly Microsoft had to embrace this business.

satay nadella takes charGe

By early 2013, Ballmer was coming under increasing pressure from Microsoft’s board of directors. De- spite robust revenue and earnings growth under his

leadership, Microsoft’s stock price had stagnated. Microsoft had lost its technological leadership in the industry to Apple and Google. The company’s prob- lems with Vista and Windows 8, and its failures in the smartphone, tablet, and Internet search businesses, had led directors to question the direction of the com- pany. Ballmer agreed that it was time for someone else to take the helm, and the board started to look for his successor.

Satay Nadella was picked to succeed Ballmer and took charge on February 4, 2014. Nadella was a native of Hyderabad, India. In 1988, he received an engineering degree from the Manipal Institute of Technology. He then travelled to the United States and earned a Masters in computer science from the University of Wisconsin. Later, while working full time at Microsoft, he earned an Executive MBA from the University of Chicago. Nadella had worked at Microsoft since 1992. He was senior VP of R&D for the Online Service Division from March 2007 until February 2011, when he was appointed president of the Server and Tools Division. This division grew at a healthy pace under his leadership. Moreover, the Azure cloud-computing initiative was based within this division. Nadella was credited for his adept lead- ership of the nascent cloud-computing business.

Nadella moved quickly to put his stamp on Micro- soft. Emphasizing a break with the past, Microsoft, he said, was competing in a “mobile-first, cloud-first” world. In this world, said Nadella, Microsoft must empower people to get things done. By June 2014, he was talking about Microsoft being the premier “pro- ductivity and platform company for the mobile-first, cloud-first world.”

In March 2014, Nadella announced that Micro- soft would offer a version of Office 365 for the iPad. A version had been in the works for some time, but re- lease had been delayed because of fears that it would boost demand for the iPad and hurt Microsoft’s Sur- face tablet. Nadella asserted that in a world where Android and iOS are widely used, Microsoft had to make its applications run on those platforms too. By Fall 2014, Office for the iPad had over 30 million downloads. Also in March 2014, Nadella announced that Windows would be free for devices smaller than 9 inches, meaning smartphones and tablets. Clear- ly this was an attempt to jump-start adoption of Windows on digital devices, and to match Google’s strategy of giving away Android for free.