January 06, 2003

Some of you may have noticed that it has been awfully hard to reach me by e-mail of late, that the messages you have been sending to pdoocey@cmp.com have been bouncing back to you at a frequent, and I'm sure quite annoying, rate.

Alas, the problem lies in our new e-mail filtering device. Ostensibly put in place to kill spam, the system has yet to consistently determine which e-mails are legit and which are pitches. The end result, a lot of good mail gets tossed out with the bad.

We are attempting to solve the problem. If you have been trying to reach me and have suffered a bounce back, please e-mail a short message to that effect and I will add you to an exemptions list that should solve the problem.

In the past, I have used this column to point out shortcomings in bank or vendor technology strategies in the hope that a lesson could be imparted. So I would be remiss, and somewhat hypocritical, if I did not cast the same eye on our e-mail filtering experience, in hope that the mistakes made here will not be repeated elsewhere.

Indeed, the original intention of the e-mail filter was to make both the employees and systems at CMP Media more efficient. E-mail is a critical channel of communication for us, and the influx of unwanted spam was slowing the flow of information and causing storage problems. Eliminate the spam and improve the flow of information, the easier for work to be accomplished.

So an e-mail filter was set up that keys on common words found in most spam messages. If these words come up a certain number of times in a message, the system refuses to accept it and it is bounced back to the sender. The problem: certain words contained in most financial spam messages (mortgages, lending, credit card names, etc.) are also contained in legitimate financial e-mails.

On paper, the easiest solution to this problem would appear to be the removal of certain words from the filter dictionary. In practice, this answer is impractical. The e-mail filter was set up company-wide; and many employees on non-bank publications probably appreciate the fact they are no longer bombarded with junk financial e-mail, just the way I'm content with no longer receiving unwanted information on Viagra.

In retrospect I suppose another solution would have been to use a number of smaller e-mail filters-tuned to the needs of each publishing group-as opposed to one all encompassing system. This would fly in the face of current bank technology logic however, which almost always stresses a universal instead of silo approach.

But the one thing this whole experience has taught me is the value of a new product audit. A number of consultants now offer this service, designed to alert institutions to all the possible impacts of employing a new technology. It may add cost to a new system, but anything is better than alienating trusted customers.

Take it from personal experience.