Just one clarification. Clusters and infrastructure belong to the same IT organization, so the clusters CIOs have a dotted line into every deputy, and I do their performance appraisals. We are one organization, and that's a very important thing.
However, to go to your specific service question, what we set out to do on day one was we wanted to at least match the service we had when we went through transition. That's enormously difficult to do, as you know, coming from IBM. We worked a lot with IBM and a lot with HP on their transition and what they had done. One thing we had been warned about was we would see a dip in service, so fight against it, but it will happen, and be ready to respond to it.
I would say probably a really great example of what we focused on was through our help desk. Part of the new structure consolidation was with our help desk. We probably took a huge step up in our measurements and our metrics on what was actually happening at the help desk, because we knew our calls were the most important thing telling us how we were doing.
One incredible advantage of having consolidated together was the fact that we could actually see what that picture was. We put a classic big screen up on the wall. We were reasonably well automated on our help desk. We were seeing the number of calls coming in, how many calls were waiting, what kinds of problems they were having. So through that consolidation, we had a single window into how we were impacting the 65,000 to 67,000 employees of the OPS.
I think I also mentioned our upgrade from XP and Office to the new version. That would have been an enormous undertaking seven or eight years ago. It's still a big undertaking, but it's nowhere near as big as it was.