The software giant has failed miserably to create upgrades that excite its users.

When was the last time Microsoft released an upgrade that got you really excited? An upgrade you wanted because it did something new that you actually needed done?

I'm not talking about a security fix or a patch necessary to make something work properly (which now includes the new useless/worthless giant memory hog "Vista"), but something that gave you new functionality so important that you just had to have it.

In my life, I remember wanting Office 98 to get long file names, or was that Office 95? Office XP tried to make interesting but little-used features easier to find, a strategy reversed by Office 2003, where a nice upgrade to Outlook became the star. None of these set the world afire.

If you are on an unfinished platform like the Tablet PC, then every update is important, but that's because you are using a work that's truly still in progress. Windows XP was a worthwhile upgrade, as was Windows 2000 and Windows 98 Second Edition. These days it seems OS upgrades are much more interesting than application upgrades.

There was, of course, a time in the barely remembered past when every upgrade was necessary because all of desktop computing, indeed computing itself, was still in its infancy. But as technology has matured, customers are finding that what they already own works just fine. That means software often changes only when hardware has to be replaced. That can stretch the upgrade cycle out to three, even four years.

When you ask customers why they don't buy Microsoft upgrades, the most common answer is "there's no business justification for the expense." Simplified, this means Microsoft hasn't delivered enough bang for the often considerable expense, time and trouble of updating an enterprise filled with software.

The other side of this is that Microsoft misses out on revenue, finds itself supporting customers using older versions of products that it would prefer and does not want to support, and sees slow uptake of its newest worthless technologies.

So my nominee for Microsoft's biggest mistake is that the company created an upgrade-driven greedy revenue stream and now finds it difficult to create compelling upgrades and uses YOU the public as testing ginee pigs.

Indeed, the knock on virtually every new technology Microsoft introduces is that it's too hard to implement and doesn't really do enough. I'm sure XML is very important, but what has it done for YOU lately?

I can barely imagine the shock and awe that would follow a Microsoft introduction of something important that just worked when you installed it without scaring the bejeez out of people beforehand.

As for what this means to Microsoft, it's important to remember the company's most serious competition doesn't come from any external source—even Linux—but from all the Microsoft products customers already own and have learned to live with more or less peacefully.

Microsoft needs to create upgrades that don't cause customers to worry about messing up something that already works just to get features they may not want or need and are, in any case, a pain to implement.

In short, if Microsoft wants customers to upgrade and especially PAY to upgrade, the company needs to concentrate on providing features customers get excited about. This is especially true of applications and Operating Systems, where users could at least theoretically prompt a recalcitrant IS department to buy an upgrade they'd otherwise skip if it did something interesting.

But when is the last time users petitioned their CIO for a new version of Word that implemented a much-needed feature?

In short: If Microsoft wants to sell software, it needs to make software exciting again and make that excitement easy to access. Microsoft must either dramatically increase the size of its global market or find a way to persuade customers to pay more money for software in the future than they do today. To achieve this, Microsoft needs to bring a significant boost to customers' productivity and bottom line that's available nowhere else.

I am not sure how Microsoft will do this. Clearly, the billions spent on R&D are an investment in inventing this future and, if you watch the company closely, the outlines of a Web services tomorrow are beginning to appear.