At Microsoft, Quality is Version 3.0

John Battelle's Searchblog features an ACM Queue magazine interview with Tim Bray, early search pioneer with OpenText and now director of web technologies at Sun. It's a great history lesson that recalls the noble intentions that people start with, how failing to learn from history leads one to try to fight the same battles over and over again, and the wisdom that often comes from simplicity.

Tim argues that trying to create a program to convert real-world concepts into search-friendly metadata is really, really difficult. I agree. I've argued the challenge of encoding real world concepts like regional tastes in rotisserie chicken is a bit like trying to explain the difference in haute couture between Dallas and Des Moines. Google came up with an elegant way of infering relevance, but since then they've built a nice portfolio of patents that prevents others from replicating their success. Companies aspiring to use the Internet to create value should understand Microsoft's approach.

I like Microsoft because people are constantly underestimating them. Despite their enormous size, they have mastered the art of versioning software. It took Microsoft Word three versions to become a serious threat to WordPerfect. Excel hit v3.0 before Lotus 1-2-3 sensed real trouble. Windows versions 1 and 2 were jokes, but 3.0 proved the Apple Macintosh GUI wasn't the only game in town. The story repeats itself time and again: it came as no surprise that with IE3.0, Microsoft had regained the ground they had lost to Netscape.

Microsoft's disciplined approach helped them recover from the blunder that assumed Office, not kludgy HTML, was destined to be the dominant authoring tool for the Internet. By incorporating versioning into corporate strategy, companies can best ensure their business partners and customers are part of their ongoing value story.