Seeing For Yourself
A key element of evaluating open source's cost savings is the fact that most open source solutions can be tried as-is without any heavy-duty financial commitment up front. Less so with commercial software, where 30-, 60-, or 120-day trial periods make it difficult to determine what the long-term payoffs will be without paying up. To that end, 451 Group also found that once people make the switch to open source, they can see firsthand how the real long-term benefits aren't in the immediate financial sense.
These other benefits aren't generic. In other words, they'll prove themselves differently to each company depending on the implementation.
The "open source" part of the equation, for those who are relatively clued-in about what open source represents in the first place, means things like less vendor lock-in and inclusive support for open standards. Actually, the question of less lock-in also represents a potential cost issue in the long term as well. The less vendor lock-in you have, the less time and effort (and, yes, money) you'll need to spend wrestling with that issue when and if it becomes a problem.
Mark Driver, a research analyst with the Gartner Group, takes the view that open source works best when community support is "good enough" for the deployment in question. "If you end up paying a commercial vendor for service and support" (whether open source or proprietary) "the role of licensing gets smaller and smaller over time. We believe fully 50% of people deploying OSS aren't actually saving money over closed source alternatives in the long run. They're still getting value, though." He wasn't thrilled by the idea of free software being a one-to-one replacement, though: "Free doesn't mean much if you lower the quality of service (and increase risk)."
The conceit that cost savings is not the only thing you'll get from switching to open source, and maybe not even the most important thing, has been prevalent for a while now. In a survey conducted in 2005 by ComputerEconomics.com, they found the key advantage of open source is not cost savings but flexibility. "Real TCO takes years to gauge," says Driver, so perhaps relying on TCO as the most valuable or immediate metric for whether or not to adopt open source isn't the best idea.
How Much And Where?
The exact amount of cost savings is variable, and debatable, probably because no two companies -- even those that use the exact same software -- are alike. A good deal of this may be because there are often few formal procedures in place for evaluating such costs within an organization, whether open source is involved or not.
Because of this, the numbers themselves can seem counter-intuitive. One survey conducted by SysLab found that smaller companies (10 people or so) only saved about 2% overall, with larger outfits (100+ people) saving about 6% to 7%. "The reason for this lies in the lack of internal know-how of the open source products," they claimed. "The enterprise can only work with expensive external support and the thereby resulting higher costs for training will burden the budget." (These particular numbers, however, carry with them the assumption that a) there is automatically a lack of internal know-how vis--vis using open source, and b) the only way to acquire such know-how is to hire expensive external support.)