The incumbent's conundrum
Most incumbents get this completely wrong. Microsoft has now become the latest example, as its own CTO Ray Ozzie let slip in an internal memo he recently sent to Microsoft employees:
"We should've been leaders with all our web properties in harnessing the potential of Ajax, following our pioneering work in OWA (Outlook Web Access)."
Should've, but didn't. At the time, Microsoft was intent on making sure that as few as possible users deserted the desktop environment in favor of web-based applications. Following up that "pioneering work" would have directly contradicted company strategy. Microsoft had no incentive to do anything that would hasten a switch to web-based interfaces — which is why it ignored a succession of insiders who tried to argue the case. The trouble is, other companies had no such incentive, and Microsoft has ended up following them rather than leading.
Now that Microsoft has got itself on the back foot, things will go from bad to worse for the software giant, I'm afraid. With no track record of serious investment in web-centric business models, the company has no pool of internal expertise in that area that it can draw on. That means its efforts to regain lost ground will continue to be stymied, while it will continue to waste money and resources attempting to defend its established products in areas where they have already lost the initiative (for example, attempting to supplant PDF as a format for document exchange).
Microsoft's decline will be much like IBM's from 1985-1995, when Microsoft itself was the principal agent of disruption. You'd think, given that experience, Microsoft would know better, but maybe no company is strong enough to battle its own incumbency.