Don't let the title fool you, there's no "upgrade" from XP to Windows 7--a clean install is required, as is a complete reinstall/testing of all your essential programs, and a migration of your data. This process, as noted in my previous IT Bulletin entitled "Win 7 Redux" is tremendously time consuming and costly. In spite of these obstacles, we are strongly recommending adoption of Windows 7 within Q1 2010 for all of our clients using Windows XP.
One key question that has arisen is "Should I upgrade my old desktop/laptop or buy a new one with Win 7 pre-installed?":
In managed IT, we spend a lot of time looking at TCO and other metrics to answer these questions. I won't bore you with the analysis, so here is the short form answer in two parts for those of you on the fast track:
If you have at least 18 months of next-day parts warranty or better, or you live/work in a region that has same-day high quality parts availability, you should retain your desktop/laptop and incur the cost of an "upgrade".
If on the other hand, you have less than 18 months of next-day parts warranty or better, or you do not live/work in a region that has same-day high quality parts availability, you should look carefuly at purchasing a new system with Win 7 pre-installed and a new 3 year next-day parts warranty.
For those of you who really need the train of thought:
After 26 years carefully observing this business, maintaining intense relationships with component vendors, and running extensive per-client TCO analysis both annually and every three years, this is what I have observed--
On a three year turn (average length of desktop/laptop warranty from major vendors), it is implicit that a user will have an opportunity to upgrade the OS at least once. Additionally, for Windows platforms, at least one disaster rebuild instance due to virus/spyware, OS corruption, or hardware failure must be built into any TCO assumptions. Statistically, rate of above failures will increase toward the end of any warranty period, because the underlying component vendors will not build hardware that will regularly last geometrically longer than the warranty period + 6 months (we have analyzed the fiscal incentive/deincentive mechanism between component vendor and consolidator and its impact on empiric MTBF at great length--more on this in another bulletin). These assumptions coupled with the CBA of approx $1000 to "upgrade" versus the ability to purchase new (including migration) for 2x the $1000 "upgrade" cost and to also recieve latest higher-performing technology with a new three year warranty points heavily in the purchase new direction--no one wants to spend a $1,000 to "upgrade" and then contend with the costs of a hardware failure within the next six months. This indication is especially true if the subject system is a heavily traveled mid-cost laptop. Note that the end of update support for XP increases the likelyhood of emergent problems/opportunistic exploitation and that no prediction can be made as to the cluster operation that such problems/exploitation will produce for TCO analysis.