Microsoft, the Positive Feedback Loop, and Ginger

Last week's column about software development in a vacuum garnered some great responses. Many of you speculated—correctly—that Microsoft must get amazing amounts of user feedback and use that feedback to improve its products. This thinking makes sense because Microsoft has enormous resources and the largest software user base in existence. This week, I explain why I think the company's methodology is flawed and give a relevant example of an innovator who used a different product-design methodology.

Before signing on with Windows NT Magazine in late 1999, I had participated in dozens of Microsoft beta tests, so I've seen how various teams at Microsoft receive and use (and often don't use) feedback from users and testers. Since that time, I've worked with Microsoft on a different level. I've spoken with product and program managers, and I've used this experience to gain a separate and unique perspective about the company's activities. Microsoft generally gets most user feedback after a product ships, especially with its dominant Windows and Office products. So the company thinks in terms of these monolithic platforms when developing solutions and often overlooks more efficient solutions that don't fit into its Windows or Office model of the world.

I received an email about Internet Information Services (IIS) 6.0 about 10 days ago that touched off last week's UPDATE topic. Microsoft had published an overview of IIS 6.0 for Windows .NET Server testers, and the introduction notes that "many of the new features and changes IIS 6.0 introduces are driven by the requirements of you, our customers." This statement suggests that customers' needs didn't drive at least some of the changes, and the cynical might note that maybe Microsoft's needs drove those changes. The notion that Microsoft doesn't develop products solely to meet the needs of its customers is a better indication of a monopoly than anything else, isn't it? After all, a company that doesn't dominate the market must meet its customers' needs.

Here's a better example of monopolistic practices. Microsoft originally planned to ship Windows 98 during the 1997 Christmas season, but Microsoft executives artificially delayed the release so that the company could further integrate Internet Explorer (IE) into Windows. Microsoft made this decision, as court proceedings later proved, solely to harm Netscape, although the company was able to provide technical justification for the change after the fact. (But as any former NT 4.0 user will tell you, nothing made that previously solid OS more wobbly than adding IE to the mix). During the antitrust trial, PC makers—Microsoft's partners—described how they lost millions of dollars in late 1997 because of Microsoft's perceived need to protect its Windows monopoly. Although it's an extreme example, it shows how far a company will go to change a product without regard for the end user.

Microsoft gets feedback on its dominant products through a model called the positive feedback loop, which works like this: After you define a standard and dominate a given market (e.g., desktop OSs and Office suites), the benefits multiply exponentially because more platform users result in more developers building solutions for the platform, and more solutions lead to more users. Windows and Office have become so valuable that they're stuck in this perpetual positive feedback loop, although maybe "stuck" isn't the right word: It's a lucrative business.

After you have a winning solution, you don't radically change it, because it becomes more expensive for the company and its users to change than it is to maintain the status quo (and support "legacy" users). That's why NT took almost 10 years to go mainstream: Users don't jump on a new bandwagon, no matter how much better it is. And that's why competing products in the Linux world often are exact duplicates of Windows applications and why improvements to Windows and Office are incremental, rather than revolutionary.

In this development model, users lose out on innovative solutions that address their needs. But once in a while, someone thinks unconventionally and comes up with something brilliant. Coincidentally, that happened this week, when inventor Dean Kamen introduced his Segway scooter (previously codenamed Ginger—see link below). Instead of melding an entrenched solution—the automobile—with incremental improvements such as pollution-reducing technologies and smaller, composite materials, Kamen came up with a brilliant and unique solution for overcrowded urban areas. GM or Ford would never have released such a product because it impinges on their existing sales and costs much less. It's equally inconceivable that Microsoft would promote any technology that harms Windows or Office, which is why the company killed its Net Docs project. The project would have replaced Office with small, light, .NET-savvy Web services. The fact that the company spends billions of dollars on R&D and user feedback doesn't really matter if all we get is simply a better Office and Windows. Round and round we go.

Resources
Genius Inventor Announces Ginger