Microsoft always has been very smart about pricing of server CALS, causing that the costs of a user login on a MS Server were equivalent to buying an new PC with windows. This fact put every alternative, using server based solutions, on a cost disadvantage.
The mix of Microsoft software which is good enough quality, and the fact that developing alternatives did not offer significant reduction in costs compensating for the risks of user revolt, "Hey my MS-Word icon is not on the right top and it looks weird". Until recent this all worked out well for MS.
Perhaps times changed. First people are now used to smart phones, forcing them to get used to something else than the MS interface, users are more flexible.
A consultant who nowadays still would propose a local 1995 style windows farm implementation to small/mid sized companies is blinded by ancient box shifting paradigms.
The operating costs of a 1995 MS-farm are significant. Many people needed for backup, constant exchange server breast feeding, sql servers crying all the time, and how about rebooting the whole farm once a month for patching the latest set of 9+ CVE's ?.
It can be explained that windows alternatives were not successful in the past, for the future there is a clear potential for alternatives, cloud and thin client based, to gain a significant share on the working place.