According to San Francisco-based Ferris Research's newest report, spam's cost is primarily in lost worker time as employees filter spam, deal with false positives, and query corporate help desks for assistance with the plague.
"But the costs could be a lot worse," said Richi Jennings, one of the Ferris analysts who authored the report. "We haven't seen as much of a spike in costs as in spam volume," he added, "because more organizations are putting in better anti-spam technology."
Since 2003, for example, spam volume hitting U.S. enterprises has jumped five fold, while costs have not even doubled. In 2003, Ferris estimated spam costs to U.S. organizations at $10 billion, with 2005's lost money amounting to $17 billion.
With such great sums of money at stake -- 0.17 percent of the U.S. gross national income -- it's easy for companies to make a business case for anti-spam technology, particularly in developed countries where labor costs are higher (thus raising the dollars lost to productivity drains) and spam volumes larger.
"For developed countries, deploying competent spam filtering software makes good business sense," said Jennings. "The business case in emerging economies is less clear-cut."
In countries such as India and China, where spam volumes are relatively low (so far) and labor costs are rock-bottom, costs are less than a third of those in the U.S. per mailbox. "If all you're considering is the direct measurable cost and nothing else, you could say that part of these people's jobs is to delete spam," said Jennings. "It's simply because labor is so much cheaper."
But in higher-priced economies -- such as the U.S., Europe, and Japan -- costs skyrocket, and vary to a surprising degree.
In the U.S., for instance, spam's annual per-mailbox cost to businesses is $170. In Germany, however, that number jumps to $241 a year. Germany's labor costs are higher, noted Jennings, because of fewer work days and high health care and pension costs borne by businesses.
Nor is every method of filtering spam equal in economy, he added. Server-based filtering is considerably cheaper than desktop-based anti-spam solutions: the former typically costs $132 per year per user, while the latter runs $217.
Doing nothing costs even more, since manually filtering drives up the spam price tag to a whopping $718 per year per user.
"There are very few scenarios where we would recommend desktop filtering," said Jennings, "unless it was a 'belt and braces' approach with multiple layers at, say, the gateway, inside the network, and on the desktop. But I don't see any trend of this."
It's hard to make a financial case for desktop filtering, he added, since the IT management costs are so much greater for that tactic compared to filtering at the server. "Anytime an IT department has to roll out software to everyone's desktop, you're talking serious money," he said. Sometimes people try to make the case that desktop-only filtering is more economical for small companies or for those organizations where a minority of the workers receive most of the spam, but "even then, a hosted spam filtering service from a third-party service provider will probably be more economical and effective," Jennings said.