If the battery in your iPhone goes dead or the carriage drive belt snaps in your printer, the procedure for replacing either might be less complicated than the assembly instructions for a particle board TV stand. If you're a tinkerer and you don't mind reducing a product to a pile of hobby parts, go to town. Some find this sort of activity enjoyable, and there's a fair education to be had as well. But most of us won't bring that "let's crack this puppy open" ethos to work. If a machine breaks in business, it isn't a job for the office's self-identified Mr. Handy. Generally, someone wearing a logo shirt is called in. Repair is never as easy as it looks, and the cost of paying for quick professional repair is lower than the cost of extended downtime or, if a do-it-yourself repair is thoroughly botched, the price of new equipment. That's just common sense.
This is a good time to rewrite common sense, and to start writing and fulfilling service contracts in-house. Only a small percentage of the equipment that an IT operation uses actually needs extended or on-site service coverage. If equipment is chosen well, a service contract is just flushed money, a bet with your money the odds of which heavily favor the vendor. It does not accrue, it does not raise your computing or storage capacity over time. The people you're supposedly paying to keep at the ready in case your server breaks down at 3 a.m. on Sunday are not going to pitch in to help you QA a new site or fly to a client's location. A hardware service contract is not outsourcing. It is insurance, and considering the range of scenarios a hardware service plan doesn't cover, it's not even all that hot as insurance.
[ Looking for a closer shave? See InfoWorld's article "16 ways IT can do less with less." ]
That makes hardware service contracts an easy target for bloodless budget cutting. For starters, I'll wager that most large corporations are spending money to cover repairs on equipment that's not even plugged in. Companies that buy extended or on-site service for server equipment should weigh the long-term cost of a contract against the one-shot cost of spare parts. If a vendor doesn't offer attractively priced spare parts kits, it's time to shop for a new vendor.
At least two smart vendors realize that there's money to be made selling systems to businesses that are willing to do their own repairs and upgrades. This week, you read my review of Apple's Nehalem Mac Pro, a desktop/workstation that's built for easy, tool-free, and risk-free user service and upgrades. Apple's Xserve is constructed to come apart easily as well. AMD exercises a policy of keeping existing Opteron servers and workstations serviceable and upgradable. AMD continues to manufacture CPUs that are replaced by new models, keeping the cost of repair parts low. Most of the time, upgrading to a CPU with a higher clock speed or more cores does not require a new system purchase.
Once you see system repair as an insurance policy, you realize that except for systems of considerable scale, you can handle your own underwriting. Bank the money you'd spend on service contracts and use it instead to fund upgrades later on. Just make sure you're buying upgradable computers. That happens to be my area of expertise, and I'd be glad to help out.