The buildup to NT began after the incredibly successful launch of Windows 3.0 in 1990. For the next 3 years, Microsoft spent considerable time proclaiming that this new version of the product, once known as OS/2 3.0, would be the 32-bit successor to the 16-bit Windows 3.x product line. But as NT neared completion, complaints began to surface that the product was too big and resource-hungry to fit the existing desktop profile. Microsoft had heard these complaints before with other products, but Moore's Law which, roughly paraphrased, states that computing capacity doubles every 18 months, had always bailed out the company in the past. In a rare case of Microsoft losing its nerve, NT was quickly hustled offstage and repositioned as a local area network (LAN) alternative to Novell's NetWare where, with Novell's unwitting assistance, it enjoyed tremendous success.

Microsoft then cobbled together a DOS-based 32-bit hybrid that would eventually be known as Windows 95 and switched promotional gears, telling everyone that this product was in fact the desktop upgrade Microsoft had been promising. Windows 3.x's huge installed base, IBM's ineptitude in marketing the competing OS/2, and a massive promotional campaign all contributed to Windows 95's tremendous sales success. But over time, the positioning problem grew in the critical desktop arena. Windows NT, then 2000 (the more things change …), had always been available in a "workstation" version that directly competed with the Windows 9x family. After all, both product lines were called Windows. They were both 32-bit operating systems. The desktop versions were comparably priced. They even looked alike. So, which to buy?

Microsoft tried to help customers make the decision via a classically bad 1996 ad campaign many referred to as "Two Nags Racing." A two-page spread, it featured a picture of two horses running neck-and-neck with the caption "You See a Horse Race. We See Two Thoroughbreds." Apparently no one at Microsoft had realized that, well, yes, but the horses are racing. And as we all know, only one horse can win. So, which customer is going to ride the losing steed? Faced with such a choice, corporate America paused (and the ad was quickly yanked). Two years after the release of Windows 95, over 60 percent of the U.S. corporate market was still using Windows 3.x. This didn't seem to particularly bother Microsoft; after all, businesses would have to upgrade sooner or later and they had only one choice. A Microsoft choice. Right? Right.