Courtesy of Sadagopan I saw this Forrester analysis saying captive IT or BPO shops many corportaions and tech vendors have set up around the world are "failing".
If they are failing, so are other buy versus build decisions - custom development of software versus buying packaged software, decisions to in-source outsourcing deals that have not worked out or indeed decisions to not outsource certain functions in the first place.
Having talked to a few captive unit managers, the single biggest driver for them is the high 40% gross margins the large Indian vendors target. That makes a juicy target to try and squeeze out. Others have IP protection reasons to do it. Still others want to emulate the payback that GE and others have seen from well implemented captives (not just from lowered operating costs, but also from capital gains upon sale of the captive to an outsourcing firm. Firms like Symphony are actively calling technology vendors and offering to buy captives).
Personally, I am more of a buy versus build guy, but I understand why companies try to make captives work. Not a "failing" strategy by a long shot if you understand the economics, before you sit down and negotiate to buy.