Read this book from the beginning
In 1980, Microsoft was a small software vendor that had built its business primarily on downsizing mainframe programming languages to a point where they could be used to program the desktop computers that were then coming to market. The five year old company had total revenues of $7,520,720, and BASIC, its first product, was still its most successful. By comparison, Apple Computer had already reached sales of $100 million, and the same year launched the largest public offering since the Ford Motor Company had itself gone public some twenty-four years before. Microsoft was therefore far smaller than the company that Steve Jobs and Steve Wozniak had formed a year after Bill Gates and Paul Allen sold their first product.
Moreover, in the years to come, PC-based word processing products like WordStar, and then WordPerfect, would become far more popular than Microsoft’s own first word processing (originally called Multitool Word), providing low-cost alternatives to the proprietary minicomputer based software offerings of vendors like Wang Laboratories. IBM, too, provided a word processing program for the PC called DisplayWriter. That software was based on a similar program that IBM had developed for its mainframe systems customers. More importantly, another program was launched at just the right time to dramatically accelerate the sale of IBM PCs and their clones. That product was the legendary “killer app” of the IBM PC clone market: Lotus 1-2-3, the spreadsheet software upon which Mitch Kapor built the fortunes of his Lotus Development Corporation.
Lotus 1-2-3 was not the first software developed to create and manipulate spreadsheets. That honor belonged to VisiCalc, a program created by desktop software pioneers (and my personal friends) Dan Bricklin and Bob Frankston, the founders of Kapor’s former employer, Software Arts, Inc. Bricklin and Frankston had invented – but famously not patented – the electronic spreadsheet they brought to market in 1979. VisiCalc was successful in the context of the limited number of desktop computers in existence at the time, but Lotus 1-2-3 was explosive in its success. With a copy of the incredibly popular 1-2-3 on board, a PC could provide a business user with more than just a souped-up electric typewriter. Moreover, the Lotus program included other useful, although less often utilized capabilities: graphing and database management (the “2” and “3” in the product suite name). Lotus 1-2-3 yielded $54 million to Kapor’s company in its first year of sales, making Lotus the largest software developer in the world almost overnight.
Affordable, easy to learn and use programs like WordPerfect and Lotus 1-2-3 helped large businesses justify investing in PCs, and made it easier for small businesses to make the leap from electric typewriters to computers as well. This was both good and bad news for Microsoft. On the one hand, the introduction of such popular application software dramatically increased the market for the DOS-based computers that ran these programs, and provided rapidly growing license fees to Microsoft. But on the other hand, the independence of the developers of such software eroded the OS-based advantage that the fledgling software vendor was just then gaining through the development of the clone market. After eight years of operation, the revenues from all products being sold by the company that Gates had built were not much more than half the amount that poured in to Kapor’s new company in the first year of sales of its single program.
Lessons learned: There were two morals for Microsoft to absorb from the lessons of 1983. The first could be found in the huge success of Lotus 1-2-3, which clearly demonstrated that developing the right desktop software applications could be a very lucrative business indeed. Microsoft took this to heart in an ecumenical and pragmatic fashion, developing application software for Apple computers as quickly as it could. Soon, Microsoft was the dominant provider of software for the Apple marketplace, even as Microsoft sought to replace Apple desktops with those running MS-DOS, and later Windows.
Microsoft drew the second moral from its own dramatically increasing OS-based revenues which, conjoined with the fact that it was swiftly taking over the Apple application software market, proved that reaping the full benefits of controlling an OS might not be as easy as first it may have seemed. The success of Lotus also introduced a new problem: If Microsoft wanted to upgrade its operating system to maintain its appeal, it would be dependent on whether Lotus, WordStar and WordPerfect were willing to adapt their products to run on the upgraded OS. And if Microsoft wished to introduce a new OS – like Windows – it could only expect to carry its customers over to its new offering if the most important ISVs followed as well. But that would require them to invest significant extra effort to “port” their offerings to the new OS.
This was a serious issue, since only by upgrading could Microsoft and other PC software vendors earn additional revenues from customers in connection with products they had already bought and paid for. And, as significantly, only by upgrading could Microsoft keep up with the innovative Mr. Jobs. Thus, while Microsoft was growing at a healthy rate, it was still vulnerably small in comparison to its principal competitors.
|Lessons applied: Microsoft not only learned these lessons, but applied them with a vengeance. In the years to come, it would turn its control of the OS from a liability to perhaps its greatest asset, limiting early access to the “APIs” (application program interfaces) needed to easily adapt third party software to run on Microsoft OS and application software to its friends, and withholding them from its enemies. As Microsoft’s market power increased, the incentives to stay on its good side increased proportionately to the damage that a fall from grace could wreak on anyone that refused to play ball.
Have YOU Discovered the Alexandria Project?
Microsoft was also not above misleading companies that thought they were in Microsoft’s good graces. Lotus fell to that ploy, when it (along with many other vendors) believed Microsoft when it assured them that it intended to continue to support OS/2, an operating system it had agreed to promote with IBM. Based on what came to be called Microsoft’s “head fake,” Lotus dedicated its resources to porting 1-2-3 to OS/2 before turning its attention to Microsofts’ struggling Windows OS. When Microsoft abandoned OS/2 and released a new and much improved version of Windows – with Excel already adapted to run on top of it – Lotus was caught flat-footed and months behind. Sales of the till-then slow selling Excel took off, and steadily closed on, and then surpassed, 1-2-3 from that point forward.
Microsoft would use the same tactics, and more, against WordPerfect in the years to follow. Novell, which later bought WordPerfect, brought a suit against Microsoft, detailing at great length the “anticompetitive practices” that it claimed Microsoft had employed to achieve a monopoly in word processing software. In the complaint Novell filed in November of 2004, Novell laid out the impact on WordPerfect as follows:
By reason of Microsoft’s anticompetitive acts, WordPerfect’s share of the word processing market, which was nearly 50 percent in 1990, fell to approximately 30 percent in 1994, and to less than 10 percent by the time Novell sold WordPerfect and the related applications in 1996. Over the same period of time, and due to the same anticompetitive acts, Microsoft Word’s share of the word processing market rose from less than 20 percent prior to 1990 to a monopoly share of approximately 90 percent by 1996. As a result, Novell suffered lost profits and goodwill during the period in which it owned the rights to and related office productivity applications, and the value of these assets declined by approximately $1 billion between May 1994 and their sale in March 1996.
As Novell noted over and over in the same complaint, Microsoft had been found in the federal district court prosecution that detailed Microsoft’s efforts to defeat Netscape (United States v. Microsoft Corp., 87 F. Supp. 2d 30 (D.D.C. 2000), aff’d in part, rev’d in part, 253 F.3d 34 (D.C. Cir.), cert. denied, 534 U.S. 952 (2001) to have engaged in a consistent, and chilling, pattern of such behavior, including the following:
i) Microsoft chose to forego the short term benefits of having more applications available to run on Windows and, instead, chose to create incompatibilities that obstructed the development of certain applications that otherwise might run on both Windows and other platforms, because such applications threatened the applications barrier to entry.
(j) Through its conduct toward competitors and OEMs, “Microsoft has demonstrated that it will use its prodigious market power and immense profits to harm any firm that insists on pursuing initiatives that could intensify competition against one of Microsoft’s core products,” such as Windows, Office, Word, and Excel.
(k) “[Microsoft] charges different OEMs different prices for Windows, depending on the degree to which the individual OEMs comply with Microsoft’s wishes.
(1) OEMs lack a commercially viable alternative to licensing Windows for pre-installation on their PCs.
(m) Microsoft used inducements such as reductions in the royalty price of Windows to entice OEMs not to pre-install competitors’ applications.
(n) Microsoft punished OEMs that pre-installed office productivity applications competing with Microsoft’s applications, by charging them higher prices for Windows and withholding technical and marketing support.
(o) “Because of the importance of ‘time-to-market’ in the software industry, ISVs developing software to run on Windows products seek to obtain beta releases and other technical information relating to Windows as early and as consistently as possible. Since Microsoft decides which ISVs receive betas and other technical support and when they will receive it, the ability of an ISV to compete in the marketplace for software running on Windows products is highly dependent on Microsoft’s cooperation.
(p) Microsoft withheld crucial information regarding Windows as part of its strategy to injure firms that threatened to weaken the “applications barrier to entry.” Id. ¶¶ 90-93.
(q) Microsoft employed a strategy of giving away its software products for free.
(r) Microsoft entered into anticompetitive arrangements with OEMs that foreclosed competing products from the OEM distribution channel. Id. ¶¶ 144-241.
(s) Microsoft used Microsoft Office to maintain the applications barrier to entry that protected its operating systems monopoly. Id. ¶¶ 341-356.
In the 1980s, Microsoft was beginning to learn how to play that game as it faced the decision of how to upgrade its OS without losing its customers. As early as 1983, Gates gave a private demonstration to IBM of an early version of what was to become Windows, promising its hardware partner to release the product within a year (Windows 1.0 did not in fact reach the market until 1985, and was then an add-on that operated on top of DOS rather than a stand alone OS). When Windows 1.0 did at last reach the marketplace, sales were poor, in part due to a lack of application software, and in part because Windows 1.0 paled in comparison to the elegant Macintosh GUI released by Steve Jobs the year before. Microsoft was therefore in danger of losing to Apple the market share it had so recently acquired, because some products (like spreadsheets) were far better adapted for use with GUI-based systems than on DOS-based machines.
The way forward was therefore clear: expanding Microsoft’s existing line of application software would provide the type of revenue differentiation that would please Wall Street (Microsoft was already preparing for its own IPO, which it launched with great success in March of 1986). And developing and controlling its own versions of the most popular business software applications would also would make it more likely that Microsoft could carry its OS-based market advantages forward from one product generation on to the next.
Windows makes the scene: 1987 was the turn around year for Windows. In January, an ISV called Aldus ported its popular PageMaker desktop publishing program to the new Windows platform. PageMaker exploited the “point and click” capabilities of both a GUI and a mouse, and was already a hit on the Macintosh. And Microsoft released Windows 2.0 – a version that took much greater advantage of the GUI-enabled possibilities that Apple had so ably demonstrated (enough advantage, in fact, to inspire the Apple copyright lawsuit that plagued Microsoft for several years). When Microsoft launched Excel later the same year on Windows, it wasn’t a buggy version 1.0 – it had been introduced by Microsoft for use on the Macintosh in 1985, and had evolved from a more primitive spreadsheet called Multiplan that Microsoft had first offered in 1982.
Critically, Lotus lagged on porting 1-2-3 to Windows due to Microsoft’s head fake, a decision that could have hampered sales of Windows 2.0, but served instead to help Excel gain traction. By 1995, sales of Excel had passed those of 1-2-3. Perhaps happily for Lotus (and its stockholders), IBM acquired Lotus later the same year through a hostile takeover, and at a healthy premium, notwithstanding the fact that none of Lotus’s subsequent products ever achieved close to the success of its first.
Microsoft later out-pointed WordPerfect as well, although again the progress was slow. Microsoft Word (initially called Multitool Word) was originally developed for sale on Unix-based systems, but Microsoft ported it to DOS in late 1983. Word was in some ways ahead of its competition, employing both a mouse as well as enhanced screen display features that made it something of a transitional step between typical plain text DOS programs and those that exploited a GUI-based interface. Word was released on the Macintosh in 1984, and after a rocky release of version 3.0 on that platform in 1987, eventually made it to the Windows OS in 1989. Once again, Microsoft’s tactics worked, when WordPerfect fell to the same head fake that had taken Lotus in, and failed to promptly release a version of its flagship product for Windows. By the time Word 2.0 was released, Microsoft had taken the sales lead in word processing software as well.
The advent of Office: Part of Microsoft’s advantage lay in its savvy decision to offer a bundled version of most of what we know today as Microsoft Office at a very attractive price, compared with à la carte purchases of the stand-alone products of its competitors. As usual, Microsoft first offered the suite to the Macintosh customers it did not yet have (in 1989), and only made the new software available to its own, existing Windows customers in the following year. Buying the whole package from the same vendor had advantages beyond price, however: because Microsoft developed each of the component products, they could work easily and seamlessly together. As a result, when a customer used Microsoft’s Word as well as its Excel programs, data entered into one could be easily transferred into the other. In fact, an entire spreadsheet could be embedded in a Word document, providing a great benefit to a harried middle manager rushing to complete a report in time for a meeting with her boss.
Moreover, because Microsoft now had useful application products as well as OS software to sell, it could offer OS/application packages with bundled pricing as well, making it easy to practically give away new products it wished to promote while still turning a profit. As a result, it could make it particularly attractive for both existing as well as new customers to make the transition from their old application software to Microsoft’s competing – and (initially) low priced – offerings. Once installed, of course Microsoft could sell upgrades of the same software to the same customers – at full price – for many years to come.
The dawn of the age of collaboration: Microsoft had another advantage, due to the dramatic growth that it enjoyed during the 1980s. With its wildly successful public offering behind it and ample cash in the bank, it was a more formidable and credible competitor. Timing was in its favor as well, because as Ethernet-based networking became easier and client-server architecture more pervasive, the practice of collaborating on documents was becoming more common in the business world. Even beyond the local area network, documents began to change hands in a new way: instead of marking changes on a draft by hand and mailing or faxing these hard to read amendments back to the original author to be keyed in, a party to a transaction (for example) could make the changes directly on her own computer, and then send the revised document back and forth overnight on disk to the other parties that might want to suggest changes, using another comparatively recent innovation – affordable, overnight, national couriers like Federal Express. Eventually another new innovation called “email” allowed you to skip the overnight delay entirely, and simply trade documents back and forth in real time attached to electronic messages.
Once it became possible to swap documents in electronic form, of course, it became important for both sides of the exchange to use the same software. At first, one product couldn’t open a document created by the other at all. But even after competing vendors began to work that out to a degree, the results weren’t perfect. Fonts might change, along with margins and other formatting, and footnotes might no longer be footnotes, appearing instead in strange, unintended ways or places. Valuable time was lost in resolving such issues every time a document changed hands. As more and more business owners began using Word, their law firms, accountants and other service providers felt the pressure to switch to the same software that their customers used – and did switch. As the pool of Word users became larger and larger, the pressure to join the club became irresistible, whether you liked Word better than what you were already using or not. The result was devastating for WordPerfect, and helped seal the fate of Wang Laboratories as well, which had never transitioned away from the dying, minicomputer-based world it had helped to create. The same need for a common program, of course, applied to spreadsheets as well. The result was that not just Word, but Office, became the “de facto” standard of the marketplace – the product that everyone used, because it made life so much easier if you used the same product that everyone else did.
The power of the network: Economists have a word for this aggregating process, calling it the “network effect.” Succinctly stated, the more people (or businesses, or interoperable databases) become part of a network, the more valuable the network becomes to everyone and everything that is connected to it. The most familiar example is the telephone system. As the number of people you know that own a telephone increases (not to mention the number of dry cleaners, pizza shops that deliver to your door, and so on), the more worth your while it becomes to buy a telephone yourself.
Once such a network effect becomes noticeable around a new product or service, its popularity can grow rapidly indeed. Of course, once the tipping point is reached, the viability of similar products or services that can’t connect to that network (say, telegraph companies, in the telephone example) plummets, no matter how valuable they may have been before, or how attached their customers may once have been. In the case of word processing software, many former users of WordPerfect greatly preferred that program to Word, and sorely missed their favorite features (such as the “RevealCodes” function, which does not exist in Word). They also found that Microsoft’s developer-focused culture resulted in some complementary features in Word (such as macros) being much more complex and difficult to use.
But Microsoft’s marketing power and the bundled appeal of Office proved superior to WordPerfect’s abilities and the attractiveness of its stand-alone product. Eventually even the most ardent WordPerfect users, such as law firms, converted to Office simply because they had no choice. By the early 1990s, WordPerfect’s market share was evaporating, just as any economist’s model would predict that it should.
But that wasn’t all. In order to not only capture but keep its new customers, Bill Gates took no chances. Along the way, Microsoft’s developers crafted their new software in such a way that users of WordPerfect and other still-popular products could convert their documents into Microsoft Office documents with all of the formatting (e.g., margins, fonts and so on) beautifully intact, making it painless to make the switch to Office. But they made sure that this easy path from other products to their own was a one-way street. Converting a store of existing documents into Word was easy, and the results were flawless. But converting them back later on was not. Soon, it was not possible at all. As Microsoft business customers created more and more documents that they might need to access in the future, leaving Office ceased to be a viable option at all, for all practical purposes. Like a lobster entering a fisherman’s trap, the way in for a customer was all too easy, but the way out was not.
The bundles grow: As Microsoft steadily increased its share of the office application market, it broadened and solidified its monopoly by bundling more and more desktop products and functionalities together into both Windows and Office. Once again, the value of the bundle increased, even if the price did not. But often, the small ISVs that had originally developed the duplicated products – together with all competition and innovation with respect to their product features – simply ceased to exist.
Of course, when business people bought home computers, they wanted to use the same software at home that they used all day at work. And Microsoft made it attractive for new computer users to do just that, crafting deals with PC vendors that made it attractive to buy a computer with Office as well as Windows pre-installed. Students could buy particularly inexpensive versions of Word or Office that were not fully equivalent to commercial versions, but still had more functionalities then they were ever likely to need. In due course, not only business users, but virtually the entire market had become effectively locked in to the combination of Word on Windows. With time, competition in operating systems and in many types of business application software effectively ceased to exist on the desktop.
By this point, the barriers for any competitor to challenge Microsoft in the office productivity software market had become nearly insurmountable, just as they would be for anyone trying to sell a handset that couldn’t call any of the millions of other telephones already in use. Microsoft could now charge almost whatever it wished for Windows, Word and Office without losing its existing customers. Equally troubling, once the competition was wiped out, there was no longer any reason for Microsoft to regularly introduce new and better versions of that product. So major upgrades became less frequent. Higher prices and lower production costs had the predictable and intended result: Microsoft’s profit margins – and stock price – both increased to spectacular heights. And while it still needed to listen to its customers, Microsoft could be selective about what it gave the customer in response. It could also make it more attractive to customers to buy other types of products from Microsoft as well, which more and more of them proceeded to do.
What a difference a decade makes: In 1990, Microsoft revenues had soared to almost $1.2 billion – up an astonishing 32% from just the year before. And in the following year, its sales grew even more quickly, exceeding $1.84 billion. As a result, Microsoft’s resources were formidable and growing rapidly.
Ten years after introducing MS-DOS, life in Redmond was very good indeed – and getting rapidly better. For the next fifteen years, sales of Office would simply grow and grow, without a cloud on the clear, blue horizon.
->Next Chapter: Eric Kriss, Peter Quinn and the ETRM
An open standards tree standard grows in Massachusetts
<-Previous Chapter: Products, Innovation and Market Share
In which Microsoft conquers the desktop operating system marketplace through a combination of luck, a willingness to adopt the innovations of others, and aggressive business practices