There's no doubt that the concept of TCO - total cost of ownership - is becoming increasingly popular. If nothing else, it reminds people that things such as cars and computers are very much more expensive than they like to think.
Of course, it does have drawbacks. The most obvious is that the TCO figures are more or less worthless unless they are based on specific groups of people doing specific tasks in specific geographical locations. (Staffing levels, skills and salaries vary.) On the other hand, it can lead to more sensible IT spending.
Currently, many of the decisions about IT systems are made by IT people. They look at the costs of acquisition, installation, support and so on - the ones that come out of the IT budget. But in badly designed systems, all these costs are dwarfed by the real cost to the company, from (lack of) user productivity. That does not come out of the IT budget but, in the end, someone has to pay for it.
The research and advisory company Gartner made a similar point recently about the TCO of obsolete PCs. If you keep PCs in service for four to six years instead of three, it doesn't add much to the TCO. However, as Gartner points out, it just shifts the costs from the IT organisation to the end user.
I'm always amazed by the number of PCs still running antique DOS-based versions of Windows, when faster and far more reliable NT-based versions have been available for a decade. In a few cases, this must be because these machines are still doing exactly the same job they were bought to do, and there are no benefits to be gained from updating them. More often, it's people pouring money straight down the drain.
Gartner has recently produced TCO figures for a variety of desktop systems of topical interest. For example, it reckons that for knowledge workers, a Windows XP machine costs $5,908 a year, while a Linux desktop costs $6,101 a year. The similarity should not surprise anyone. Since you can buy a perfectly good PC for $600, or $200 a year, the price of the operating system is an insignificant proportion of the TCO.
If you really want to lower TCO, invest in systems management software that enables you to get rid of expensive support staff.
Gartner's figures are not of practical use, because they don't include the cost of application development and the use of company-specific software. Still, they help put the "free" software argument in a more sensible context.
But if you want real TCO figures, you have to do proper time and motion studies with real members of staff running real applications. You cannot assume, for example, that one word processor or mail program is the same as any other. It just ain't so.
I have suffered from one particularly painful example: a notoriously awful email client was costing me at least 30 seconds per outgoing message in wasted time, or about 60 hours a year. Multiply that by the cost of employing your average knowledge worker and it dwarfs the cost of the PC.
· If you'd like to comment on any aspect of Online, send your emails to online.feedback@theguardian.com. Please include your address and telephone number. If you do not want your email address published, please say so.
Useful links
Dublin Core Metadata Element Set
http://dublincore.org/documents/dces
Adobe XMP
www.adobe.com/products/xmp/main.html
JITI Builder
www.hisoftware.com/Franklin/JITIBLD.htm
Gartner: Re-evaluating PC TCO
www4.gartner.com/pages/story.php.id.9423.s.8.jsp
Evaluating Total Cost of Ownership for Software Platforms by Alan MacCormack
http://aei-brookings.org/admin/pdffiles/phpdI.pdf
