decided to pay about $300,000 for anti-spam software. I am on the Bloomington Faculty Council Technology Policy Committee. I was skeptical about the cost, wondering whether the same amount of spam might be stopped at lower cost. I was told that actually the software would do other good things besides stop spam, and am satisfied that this is so. Here, though, let me show why the benefit of anti-spam software far exceeds $300,000. < p> I am told that about 60% of the email traffic that arrives at Indiana University is spam. To look at the benefit of stopping this, however, what is most important is the time cost to users. Let us ignore the benefit from reducing viruses and reducing the temptation of people to succumb to ads that take away their money and immerse them in things such as internet pornography. Instead, suppose everybody just takes a quick look at the spam message titles and deletes them. (Ignore, too, the mistaken deletion of non-spam messages when we're too quick with our delete fingers.)
Suppose each user spends just 30 seconds each day deleting spam. The Indiana Daily
Student wrote an editorial,
IDS "Software a 'pure' waste,
UITS spam program not needed", saying
But we think the problem with spam isn't so much a question of "How can we get rid of
it?" as much as "How much of a problem is it?" With one click of a button, an e-mail is
easily deleted, and users are often warned to never open any e-mails or attachments from
unfamiliar senders. Users can also sign up for alternate e-mail accounts when applying
for things online and should only disclose their school addresses for private matters.
While this might not completely eliminate spam, it decreases the likelihood of receiving
it.
Let's ignore the cost of remembering and checking two different email addresses, and
focus on the 30-second deletion cost. Suppose there are 50,000 users, and each one uses
his email 240 days per year. That means each user spends 2 hours per year deleting spam.
Suppose, too, that the cost of the average user's time is $5/hour. That comes to a spam
cost of $500,000 per year. If the cost of the anti-spam software is only $300,000, the
benefit is clearly far greater than the cost, even under my unrealistically
conservative assumptions.
If you like, you can do the calculation with 5,000 instructors and ignore the other 45,000 students. In that case, let's raise the average time cost to $30/hour (even though the consulting fee rate of someone like me would be more like $200/hour). Then the yearly benefit is only $300,000 per year--but the software still pays for itself in its first year of use.
I'm willing to incur the time cost of writing this because I am so happy the University is willing to spend this money. Universities ordinarily are far more willing to spend the time of their members than straight cash. It is hard for a University administration to force itself to increase its cash expenditure by 0.1% than to reduce its quality by 0.5%, as it would do if it required its members to spend 0.5% of their time on some useles activity rather than on learning, teaching, or research or support of these things. Thus, Indiana University is to be commended for biting the bullet here.
Of course, laws putting spammers in jail would be even better. But we can't wait around till the Indiana statehouse or Congress gets around to doing that. Think of the gain in national productivity when they do!
[ permalink, http://php.indiana.edu/~erasmuse/w/03.12.07c.htm ]
To return to Eric Rasmusen's weblog, click http://php.indiana.edu/~erasmuse/w/0.rasmusen.htm.