TCOs and ROI Analysis

Another IT manager puts a different spin on things: "It's easy to tell if your process changes are decreasing storage costs," he says. "If I can lay off a storage tech-support person, my improvements are making a difference." Ouch. But indeed, the preponderance of storage costs--upward of 60 percent, by some analysts' estimates--are labor costs associated with poorly managed storage infrastructure.

Critics say TCO may not be the best tool for guiding financial-management objectives because it takes into account only the costs of IT infrastructure, not the benefits. They prefer ROI analysis, which compares the costs and variances anticipated from alternative storage strategies to determine a projected IRR (internal rate of return). ROI is expressed as the amount of money an investment will return over a fixed period--7 percent over nine months, for example.

Performing ROI analyses of several storage options produces a set of IRRs you can use for financial comparisons, so ROI can be used to justify technology acquisitions in terms the corporate bean counters will understand. But IT managers often confuse ROI with payback analysis, which focuses on the cost savings that will accrue to determine how much time must pass before the solution "pays for itself." A payback analysis describes value as a statement--the solution will pay for itself within X number of days, weeks, months or years.

Neither ROI nor payback analysis provides a business case for new technology; each describes only the assumed financial value of an option. It's up to the IT or storage manager to identify the business problem and its technological root cause. If cost reduction is the only motivation for change (as it is for my bank VP buddy), say so. But a full-fledged business case is based on cost efficiency, risk reduction and business enablement, and only proposed technology acquisitions that provide value in all three categories will likely get the nod.

A final caution: ROI and payback analysis both depend on solution performance, so test products in your own environment before you bet on any vendor's proposed solution.