E-Mail Spring Cleaning

 
 
By Cade Metz  |  Posted 2003-05-02 Email Print this article Print
 
 
 
 
 
 
 

Ella cleans up your in-box and filters out spam—simply and effectively.

Welcome to the era of the antispam tool. With the average e-mail user receiving 2,200 unsolicited and unwanted messages a year, according to Jupiter Research, the market for software designed to block the flood of junk is getting bigger by the month. The latest entry is Ella, a $29.95 download that not only filters spam from your in-box but also helps organize correspondence.

In our February "Slam the Spam" story, we reviewed 26 tools capable of blocking spam—20 designed for personal use and 6 for corporate use. But these only scratched the surface of the antispam market. A recent First Look story, "More Ways to Slam the Spam" detailed even more antispam tools for personal use. As soon as we sent that story to the printers, Ella arrived.

Attempting to set its app apart from the other products on the market, Open Field Software bills Ella not as a spam blocker but as a smart in-box assistant. The software doesnt just separate spam from legitimate mail; it sorts your messages into three folders of your choosing, one of which you can designate for spam. Once you have defined, for example, one folder for spam, one for newsletters, and one for personal correspondence, you give the program examples of the sorts of messages that should go to each.

Ellas best use is as a spam filter, of course. Junk e-mail is the first thing most of us want filtered from our in-boxes. However you use the utility, you should be pleased with the results. Its not perfect—no e-mail filtering tool is—but even after the initial set-up, a majority of messages go to the right place. And as you continue training, Ella becomes increasingly accurate. Moreover, the software is wonderfully easy to use, integrating seamlessly with Microsoft Outlook 2000 or Outlook XP.

After installation, Ella launches a training wizard the first time you open Outlook. The wizard asks you to name the three folders you wish to sort mail into. One of the three is your Outlook in-box (Ella created the other two during installation). The default names are Inbox, Later, and Spam. Later is for unsolicited e-mail that is somewhere between legitimate correspondence and spam and neednt be dealt with until youve gone through your in-box. The wizard then asks you to scan Outlook and identify examples of messages that you would like sorted into each of the three folders. This takes no more than a few minutes. Initially, you have to give only 10 to 15 examples for each category.

Ella examines more than 100 attributes of each example—items like the senders address, the subject line, the message ID, and various aspects of the body text—identifying common characteristics. The program then sorts your incoming mail guided by these characteristics.

Even after a few minutes of training, Ella is reasonably accurate. We trained it to leave personal mail in our in-box, move computer-company PR newsletters and mass mailings into a second folder, and shuttle spam to a third. We filtered 1,000 messages—546 we considered personal correspondence, 137 we classified as PR, and 317 we considered spam.

The app correctly filtered 86 percent of the personal correspondence, 40 percent of the newsletters and PR mailings, and 96 percent of the spam. Thats quite impressive considering we gave the utility only 30 messages for training. As you might expect, the initial false-positive rate is high. Thirteen percent of personal correspondence wound up in the spam folder and 1 percent in the newsletters folder. To improve the results, you can provide additional training using the three buttons Ella adds to the Outlook toolbar. The buttons correspond to your three mail categories. If a message ends up in the wrong folder, you select the item and click on the button indicating the correct folder. Ella, which moves the item to the right folder, alters its filtering rules accordingly. After we trained the software with another 200 messages, the false-positive rate dropped to less than 2 percent.

Ella uses your examples and continues to learn—its as simple as that. You can rerun the training wizard to reorganize your folders and retrain the filters, but the software offers no other dialog boxes, buttons, or windows. We would like to see a whitelist so messages from certain senders always make the in-box, and we would appreciate a blacklist for the opposite purpose, but Open Field says it wants to keep the app as easy to use as possible. Ella certainly achieves that goal.

 
 
 
 
 
 
 
 
 
 
 

Submit a Comment

Loading Comments...
 
Manage your Newsletters: Login   Register My Newsletters























 
 
 
 
 
 
 
 
 
 
 
Rocket Fuel