A correct strategy is to build on a differential advantage perceived as unique, defensible and sustainable, and preserve your existing client base until your shiny new strategy has really proven its benefits

"perceived as unique" means something clients really need and feel they can't find from any other vendor;
.Net is more like an expected way for Microsoft to cut development costs (I write 'expected' because there are so few examples of successful software rewrites) rather than an outstanding added value for its customers; I don't mean 'better' or 'worse' than others, just it does not do things so outstanding that others could not do before.

there is an interesting coincidence between the year Microsoft decided to pull the plug of its existing dev clients (2005 - 2006) and the 'peak' denoted in the link from the original poster.

Microsoft could have taken advantage of browser monopoly and 'proprietary' features (remember 'behaviors' in IE 5-6?), especially XMLHTTPrequest, to build advanced web services like Google later did in 2005 - money would have been better invested here than in trying to catch up for years with Google search while it already had 80%+ market share on the consumer market.


>OK, so again, Baller was wrong.
>
>You're implying that there was a correct strategy and he missed it.
>
>What was it?