Change of basic constituent order should be relatively frequent. If two languages share a common ancestor, yet have different basic word orders, then at least one change (and possibly more) has occurred at some point in the course of their descent. When two related languages have the same basic word orders, it is often not possible to tell whether this is so because they are related, or because they are within the same linguistic area, because proximity between two languages is actually a better predictor of whether they will share basic word orders than is relatedness.
See, for example, the map: http://wals.info/feature/81A?tg_format=map&v1=c00d&v2=cd00&v3=cff0&v4=dff0&v5=dd00&v6=d00d&v7=cccc#
Languages form clusters where the basic constituent order is the same, regardless of their genetic affiliation. Although the region containing India and Southeast Asia contains a rather large number of language families, the area can be divided into two contiguous zones where the word order is the same: SOV in India, SVO in SE Asia.
On the importance of language contact in language change AND stasis, see: Dryer, Matthew S. 1989 "Large Linguistic Areas and Language Sampling""Large Linguistic Areas and Language Sampling". Studies in Language 13: 257-292. (a copy is available on the author's website)