A couple of years ago I did a piece of consultancy on in-sourcing an IT department that had been outsourced three years before. The COO had realised that his bills for the IT function were substantially higher than when the company had their own IT dept, service levels were down and complaints were persistently hitting his desk. I did the analysis of the project function to see if we could get the PMO back in-house then the ongoing projects.
The whole consultancy piece of work was cancelled when the outsourcing company pointed out that the client had sold their entire IT estate to them in return for a free year of service with no buy-back clause. That included all datacentres, office servers, desktop equipment, software licenses, etc. The outsourcing company, quite bluntly, refused to sell it back. A quick round-table discussion followed by some snap quotes from suppliers gave the rebuild costs as so high that it just wasn't worth the effort, licensing was the most expensive part of the lot.
The old CIO left a year and a bit into the outsourcing deal having apparently been well rewarded for saving an 8 figure sum in the first year. He made himself redundant in the second year with a (gossip-sourced) massive payout. The problems started coming in the latter part of that second year. I'm convinced that most of these outsourcing deals are based on similar scenarios: "How can I get the biggest bonus this year? Next year is irrelevant!"