Microsoft is preparing to announce Windows version 3.0, a major rewrite with many improved features and a much more attractive visual interface. At the same time, the hardware needed to run Windows is decreasing in price rapidly, making it affordable to many more customers. We think the combination of reduced price and improved features could make Windows much more successful than it has been in the past.
The introduction of Windows 3 is both a big opportunity and a big challenge to Apple. The challenge of Windows is that it could seriously reduce the perceived differentiation of Macintosh, if Microsoft is allowed to define the terms of the debate. On the other hand, if Apple takes the initiative, Windows 3 can be positioned as what it really is: old technology and a poor investment relative to Macintosh. By focusing public attention on interface issues, Windows 3 gives Apple a chance to articulate the full range of Macintosh superiority over PCs.
This document is designed to help you plan for, and out-compete, Windows 3.
-------------------------
Situation: Intensified Competition
Converging trends in PC hardware and software. Trends in PC-compatible hardware and software will converge in 1990 to increase the attractiveness to many customers of IBM and compatible PCs. On the software side, the new 3.0 version of Windows will be a much better-looking and more functional computing environment. On the hardware side, the type of computer needed to make Windows work properly has decreased dramatically in price. As a result, Windows is becoming more useful and less expensive at the same time.
Improvements to Windows
Windows 3.0, which we think will be released sometime between March and May 1990, will be a substantial improvement over the current version. It is not as powerful or feature-complete as Macintosh, but it is a major step forward. Changes in Windows 3 include the following:
• A much more attractive interface. The overall look of Windows has been upgraded tremendously. Color schemes have been harmonized, interface elements have been redrawn and countershaded, and a set of color icons has been designed for use in the file manager. In addition, color pictures and patterns can be displayed on the screen background. Taken together, these changes make Windows look slick and snappy—a dramatic contrast to previous versions, which were always awkward and dumpy. For the first time, Windows looks nice. Apple should not underestimate the psychological effects this change will have on analysts and customers.
• A better file manager. For the first time, Windows has an icon-based file manager, allowing users to manage files and start programs by clicking and dragging icons with the mouse. (The Windows 3 file manager is very similar to the one in Presentation Manager. We expect the two environments to look almost identical by the end of 1990.)
• Better use of memory. Programs written for Windows 3 will be able to access much more memory than they could in previous versions of Windows, which still has major problems with the DOS 640K barrier. This change will create some compatibility problems for old Windows programs, but once they are rewritten, the benefits will be faster performance and more functionality. Memory management problems are today probably the single biggest hindrance to application design for Windows.
• Multitasking. The current version of Windows contains multitasking features similar to those in Apple’s Multifinder. The new version apparently includes support for a scheduler when used on 386 systems, with the ability to specify the priority for background and foreground tasks, and to control the minimum allowable timeslice.
• Printer drivers. Preliminary versions of Windows 3 include driver support for 164 third-party printers, from 28 manufacturers, including Apple, Epson, HP, IBM, NEC, and Okidata. Windows 2 came with 24 drivers. (The increase may not be as dramatic as it sounds; many of the newly-supported printers share the same PostScript or PCL driver. Nevertheless, the change is impressive).
• 24-bit color support. Windows 3 supports 24-bit color. The ability to use 24 bits has always been built into Windows, but was not easily accessible before. Now the user can specify colors in 24 bits, and they will automatically be dithered to the closest representation possible on the user’s display. Presumably, true 24-bit images could be displayed on the appropriate hardware, but we haven’t yet been able to test that.
• One version. Currently, Microsoft sells different versions of Windows for the 286 and 386 processors. The 386 version has more features, including the ability to run multiple DOS sessions simultaneously. Windows 3 will apparently come in just one version, which will run on all PCs (although with poor performance on anything less than a 12 MHz 286). When run on 286-based computers, its 386-specific features will be disabled.
• Macro recorder. A new macro recorder function allows the user to record a series of keyboard clicks and mouse movements, and then play them back.
• Other improvements include a vastly improved color painting program, which replaces the old Windows Paint, and built-in support for 13 European languages, including French Canadian, Portuguese, and Icelandic.
• New applications. We expect to see new applications announced along with Windows. Windows PowerPoint will probably be released by Microsoft, along with a new version of Excel. New versions from Micrografx may be shown, and we may also see the premiere of an interface-creation product (resembling NeXT’s Interface Builder) from Asymetrix.
The changes in the file manager, memory use, and print drivers respond to the three most common customer and developer complaints about Windows.
Declining Prices for Windows Hardware
Windows-compatible hardware was once expensive. Performance has always been a problem for Windows. It will run on low-cost 8086-based machines, but is too slow to be viable for most users. The minimum hardware for Windows is probably a 12 MHz 286-based machine, but it doesn’t really shine until it is used on a 16 MHz machine, preferably a 386sx. That configuration gives it good performance, plus the improved memory management of the 386. However, until recently 386-based machines were also relatively expensive, giving Macintosh a substantial advantage among price-sensitive customers who wanted to run graphically-based software.
Price declines for the 386sx. In the second half of 1989, the prices of 386sx-based machines declined rapidly. At the beginning of the year, a generic 386sx system with a hard disk and VGA monitor had cost about $3,000. By the end of the year, that same sort of system sold for about $1,800. The price decline was not due to 386sx price cuts by Intel—the chip went from about $98 to about $78. Instead, we think the change was driven by the entry of low-margin PC clone-makers into the 386sx market, combined with declines in other component costs. As a result, the cost of a full-function Windows system dropped about 40% in 1989. We expect that price to continue to drop in 1990, but at a slower rate.
Name-brand vendors like IBM and Compaq have been slow to respond to the changes in 386sx pricing. As a result, their 386sx systems are much more expensive than the generics, and are not selling well. In 1990, we expect them to reprice their systems to bring them closer to the competition.
Scenario: A Good Year for Windows
In the computing industry, when new functionality is added to a product, its price usually increases, or at least stays flat. Windows is not following that pattern—its price is declining at the same time that its functionality is increasing. We think this combination will greatly increase the attractiveness of Windows to many customers and industry observers, and could push up Windows sales. Here’s how we think it will happen:
Windows will be positioned as the PC interface for the rest of us (probability 0.9). When Windows 3 is released, we expect Microsoft to reposition it as the environment of choice for price-sensitive users and anyone who can’t afford a 386-based system with lots of RAM and a large hard disk—in other words, anyone outside of large business. We expect an especially strong push toward small businesses, home users, and education.
It’s not clear if IBM will be an ally or an antagonist of this positioning. In the past, it has been reluctant to say anything nice about Windows, for fear of hurting OS/2. If IBM does anything with Windows, we think it will try to niche the product as an education and single-user environment, a positioning that would help to keep large businesses away from it.
OS/2 acceptance delayed further (probability 0.7). Windows 3 and OS/2 version 2 (due out at midyear) will have the same look and feel. Both will be able to run multiple applications at the same time. Both will be able to run multiple DOS sessions at the same time, if used on a 386 processor. Windows will have more applications, and will run on less expensive hardware. Confronted with this situation, we think many customers who might have switched to OS/2 will instead use Windows for a few years, and wait for OS/2 to mature (and for its cost to come down). Some customers, primarily in large business, will definitely choose OS/2 because it has more features, and we still think it will be the long-term winner in the PC operating system wars. But the new version of Windows is likely to lengthen the battle.
Strong competition in $2,000 to $3,000 price range (probability 0.9). In 1988 and 1989, the most intense competition in the PC industry was focused on systems over $5,000 in price. We think the focus will change in 1990. The entry of Windows 3 and the decline of its price will bring new levels of functionality to the $2,000-$3,000 street price range (the segment that historically has generated the largest unit volume). We think this will reignite demand from price-sensitive customers, who have not seen dramatic product changes in a couple of years. In particular, people who use computers in K-12, home, and small business may respond very well to the new products.
Sales Implications: Put Microsoft on the Defensive Now
We think the new version of Windows will give Apple a strong challenge, but also a big marketing opportunity. Apple’s actions today will do a lot to determine the success, or failure, of Windows 3.
What Microsoft Wants to Do
If Microsoft is allowed to take the initiative, it will position Windows as a product that delivers everything good about the Macintosh, but is also DOS-compatible and runs on low-cost PC clones. Most PC users today think that Macintosh is just a graphical interface machine, so if Apple does nothing, Microsoft will probably succeed. This would not take away the installed base of Macintosh users; most of them know better. But it would make it much more difficult for Apple to convert additional users to Macintosh.
What Apple Can Do
Set the agenda. Apple in the past has set the agenda for the PC industry, and can do so again, by aggressively explaining the advantages of Macintosh, and Apple’s view of personal computing. This will force Microsoft to explain why it hasn’t matched Macintosh, instead of crowing about the improvements to Windows.
Position Windows 3 before it is released. Apple can also proactively position Windows 3 (even before it is released) as old, inappropriate, limited technology. By using the contrast between Windows and Macintosh, Apple can clarify the scope of the advantages of Macintosh. It is easier to explain what Macintosh does well when you have something to compare it against.
Here are some of the themes Apple can use in that effort:
How Apple can set the Agenda
Talk about Apple’s approach to personal computing. Apple is generally recognized as the leader in personal computing. Apple understands personal computer users, and designs its systems from the start to make them more productive and more satisfied. One important outcome of this approach is Macintosh’s unified architecture, in which the hardware, operating system, peripherals, and applications all work together to produce maximum performance (more on that below). In contrast, the PC-standard architecture is divided and poorly-coordinated. The elements of it don’t fit together well, and innovation is very slow.
Talk about Apple’s business structure. In addition to structuring Macintosh to produce great personal computing, Apple has built its business model around the same goal.
• No hidden agenda. Unlike most other PC companies, Apple focuses just on personal computing equipment. This means Apple is free to push PC innovation to the limit, without fear of cannibalizing other products. This also means Apple doesn’t have a hidden account-control agenda to prop up sales of minicomputers or mainframes. Macintosh is designed to connect into a wide variety of computing environments. Unlike some other systems, Macintosh will not lock the user into buying other equipment from a single vendor.
• Support for third parties. Third-party hardware and software developers are an integral part of Apple’s business model, and are encouraged and guided by the company. Apple doesn’t sell applications, so leading-edge developers have more incentive to write for Macintosh. That’s one of the reasons that Macintosh software is so good.
• The contrast to Microsoft. Microsoft is also a personal computing leader, but its position is based in large part on its ability to manage alliances (especially with IBM). Microsoft cannot give PCs a unified architecture, because it does not make PC hardware. Microsoft’s culture emphasizes feature count and clever code over user-centered design—for example, Windows was launched in the mid-1980s, but only now is Microsoft settling on a graphical file manager for it. Microsoft sells applications, so it competes with its own developers. That’s one of the reasons there are so few Windows programs.
Make people understand the full range of Macintosh superiority. Apple needs to change the minds of people who think Macintosh is just a graphical interface. Macintosh is a powerful and versatile productivity tool, not just a DTP machine or an easy-to-use box for beginners. This is important for fighting not just Windows, but also Presentation Manager and the Unix interfaces. (For a more thorough discussion of Macintosh differentiation, see our report, “Beyond User Interface.”)
Put the interface in context. Most of the world does not understand the role of a graphical interface in a personal computer system. They don’t understand the differences between interfaces, and they don’t understand how those interfaces make computers easy to use. Here are three key messages they need to receive, in order to understand the superiority of Apple’s approach to ease of use:
• Explain what really produces ease of use. Most of the industry thinks that the key to ease of use in a computer is a graphical interface—any graphical interface. Although the GUI is a key factor in making a computer easy to use, it is not the only element. Apple does a lot of other, often subtle, things to make Macintosh easy to use—everything from the simplicity of LocalTalk networking to the ease of shutting down a Macintosh II. In addition, a well-designed GUI is much more effective than a poorly-designed one. Factors like consistency between applications (which allows the user to learn new programs without a lot of retraining) are very important, and Macintosh delivers them much more effectively than Windows. Customers need to understand that ease of use means a lot more than just throwing a few windows and icons onto a screen.
• Emphasize the importance of metaphors. One of the most important design elements in the Macintosh interface is the use of real-world metaphors. The Finder is a good example. It uses a desktop metaphor, with recognizable and intuitive icons like file folders and a trash can. This helps make Macintosh easy to learn and use. Windows lacks this sort of well thought-out interface design. The most glaring example is the Windows 3 file manager, which uses a tree diagram to represent the file structure on the disk. This sort of design is intuitive only to programmers and DOS hackers.
• Emphasize functionality over flashiness. One of the most striking improvements in Windows 3 is its appearance. The Macintosh interface is still better thought-out than Windows, but it may look less than striking when compared to the vibrant colors and countershading of Windows 3. We saw some of this effect when NeXT announced its machine, but the effect of Windows 3 will probably be much greater (because it will be shipping at introduction). Because customers usually do not fully understand the role of a graphical interface, many will judge it by its superficial attractiveness. It is important to educate them now, before Windows 3 is released, about the difference between flashiness and functionality.
How to Position Windows
Apple reps should not mention Windows by name to customers, until they bring it up. Mentioning Windows too early could just build more customer interest in it. For customers who are already considering Windows, we recommend using the items below as a response. For other customers, Apple reps can raise these issues without mentioning Windows directly.
It’s old technology. Microsoft will want people to think of Windows as exciting and leading-edge, but it really isn’t. It’s still built on top of DOS, and it still has a lot of limitations (more on that below). Apple should position Windows 3 as flashy-looking, but worn-out technology—like a rusted jalopy with a new coat of paint.
It’s a risky investment. There is no longer a single, standard PC operating environment. Customers today are faced by a choice between DOS, extended DOS (in many forms from many vendors), Windows, OS/2 for the 286, OS/2 for the 386, OS/2 Extended Edition, OfficeVision, and so on. This situation is deeply confusing to many customers, who risk being stuck with obsolete equipment if they make the wrong choice.
Just before Fall Comdex 1989, IBM and Microsoft tried to clarify the operating system situation with a joint press conference. They announced that Windows will not get all the features of OS/2 (specifically mentioned were distributed processing, threads, 32-bit flat memory, and long file names). In addition, they encouraged developers working on new programs to develop for OS/2 before Windows, and they said that new high-performance software should be developed just for OS/2’s 80386-specific version. Finally, after mid-1990, IBM and Microsoft will release new applications on OS/2 before they release it on Windows.
Microsoft will want customers and analysts to forget about this press conference, but Apple shouldn’t let that happen. Customers should understand that there’s no safe port in the storm engulfing the PC industry. In particular, because of the confusion among developers, there’s no way to know which environments will have a good base of software, and which ones will just fade away.
Apple should contrast Macintosh to the confusion in the IBM-standard world. Macintosh System 7 is competitive with OS/2 on features, but doesn’t require any more powerful hardware than Windows (and in some cases less). A Macintosh purchased today is a good investment both for its current capabilities, and for its growth path into the future. This isn’t just a promise—Apple can use its own track record over the last six years to demonstrate it.
It’s only an interface. By contrasting the functionality of Macintosh and Windows, Apple will be able to demonstrate, with examples, the limitations of just tacking a graphical interface onto an old PC (for an example, see the check-mark chart at the end of this document). This should help Apple in two ways: It will clarify the drawbacks of Windows, and it will help reposition Macintosh as a full-function computing system.
It isn’t nearly as good as a Macintosh. Although Windows 3 is an improvement, it still does not match the functionality of Macintosh. Here are some of the continuing Macintosh advantages:
• Macintosh has fewer compatibility problems. System 7 will not “break” most Macintosh applications, but most Windows programs will need to be rewritten in order to take full advantage of Windows 3. The memory management scheme that allows Windows 3 applications to use more memory is incompatible with most popular Windows applications, including Windows Excel. The user can still access the older applications, by starting Windows in a mode that uses the old memory model. This mode allows virtually all existing Windows software to run, but will limit the performance of new Windows applications (we suspect that some new Windows applications won’t even run in this mode). Because of this, we think most Windows users upgrading to version 3 will choose to also upgrade their applications. The upgrade will not be a disastrous problem, but it will be a major inconvenience.
• Macintosh has more applications. Macintosh has thousands of applications. Windows has hundreds. Although developer interest in Windows has increased, most PC software manufacturers are still concentrating on OS/2. Therefore, Windows will never have as rich and varied a software base as Macintosh.
• Macintosh gives the best of both worlds. It is competitive with OS/2 in features, but competitive with Windows in price and hardware requirements. (Although this information is correct from a technical standpoint, it works best on customers who aren’t obsessed with pre-emptive multitasking and memory protection).
---------- Footnotes ----------
* For technical readers: To run most current Windows applications under Windows 3, the user has to start Windows 3 in real mode. In prerelease copies, the user does this by typing "win /r" at the DOS command prompt.
___________________
COMPARISON CHART
Windows vs. Macintosh: Here’s What you Get
Standard features of Windows on an IBM PC, vs. features of a Macintosh running System 7.
Feature Macintosh + System 7 IBM PC + Windows 3
Minimum configuration 68000 & up 80286 or 80386
2 MB RAM 640K (2M for large
applications) [1]
Hard disk recommended Hard disk required
Graphical interface √ √
Multitasking √ √
Easy access to >640K RAM √ √ [2]
Interapplication communication √ √ [3]
Run DOS programs √ [4] √
Thousands of applications √
Group computing
Built-in networking (HW & SW) √
Serverless file-sharing √
Database access manager √
Communications toolbox √
File management
Full-length file names √
Aliases √
Search/Find √
Metaphor-based interface √
Graphics
# of colors standard [5] 256 16
Outline fonts √
Multimedia
Sound manager √
Time manager √
Utilities
HyperCard √
Built-in help √ √
Macro recorder √ √
Bundled paint program √
Bundled text processor √
Other features
Floating-point coprocessor √ [6]
Bundled mouse √
Bundled operating system [7] √
SCSI connector √
Multiple monitor support √
Read DOS, Mac, Apple II disks √
Auto-mount floppy drives √
Soft power √ [8]
---------- Footnotes ----------
1. InfoWorld, 1/8/90.
2. For new or rewritten applications only.
3. Microsoft’s IAC architecture is not as robust as the one in System 7.
4. Optional, through use of third-party software.
5. Standard Apple color board on Macintosh vs. VGA on the PC. Obviously, this applies only to color systems.
6. Standard on all Macintosh II and SE/30 models. Extra-cost option from IBM.
7. A few second-tier manufacturers (not IBM or Compaq) bundle Windows with their systems.
