Network Computingwww.networkcomputing.com

Several years ago, we did a critical review of the IT maintenance budget at Chatham County, Ga., the organization for which I work. We took a look at which services were "Cadillac" and which were not, where those services were necessary and where they weren't.

Process review is always valuable, and during this one we discovered we no longer needed the extended support contract and associated costs for some of our Unix servers. Reviewing the comprehensive server logs kept by our staff, we learned we had placed relatively few service calls to the vendor. Although we had our share of problems with these platforms, they had all been resolved internally.

As a result, we investigated alternative service plans and discovered that the vendor had a per-incident pricing model. Each incident was relatively expensive but affordable based on our low usage pattern. If our needs changed, we could revisit the plan. Ultimately, we dropped our 9x5 software support on many servers and saved tens of thousands of dollars.

Of course, we did not do this with highly critical servers. But for non-life-threatening servers, it wasn't a big gamble to go with this model. This approach got us through some lean times.

If you take this route, you should get buy-in from the staff: Being told that you will have less outside help is quite different from deciding yourself that you don't need as much outside help.

It also pays to involve the staff when you review the IT support budget. During such a review, staff will find all sorts of services they no longer need. For example, when my organization was reviewing maintenance last year, one of the software administrators snickered at a large line item for support of a particular package, saying, "Their support consists of looking up the problem in their Web database and telling you 'Beats me' if it's not there." We dropped the support, started using the Web database ourselves and haven't looked back.

Stepping Off the Upgrade Treadmill Reconsidering your service contracts during lean times is just one way to trim the fat. You might also want to drop "upgrade protection"--paying extra for ongoing maintenance in exchange for automatic upgrades when new releases ship. If you are locked contractually into continuing costs--such as Microsoft licenses that save some bucks early on but that obligate you to pay the same amount for two years afterward--this isn't an option, unless you're willing to invoke your contract's penalty clause. (If you take that route, proceed with a good lawyer and iron undergarments.)

But to some vendors, dropping upgrade protection simply means you don't get new releases. When budgets are tight, this can be a smart move on your part. Obviously, you'll want to involve your user community with this sort of decision (see "Too Many Cooks in the IT Kitchen?"), but people usually agree they can meet business objectives without the newest and shiniest office-productivity software. If there are exceptions, it's a lot cheaper to purchase 10 $200 departmental licenses than it is to stay on a $100,000 annual upgrade-protection plan.