And this is where it gets suspicious. The event invitation contains a notice, supposedly from ‘Gmail Customer Care’, claiming that due to some difficulty Google will close the recipient Gmail account lest they send their Google username, password and date of birth. Although it is clearly suspicious, whenever you receive such notice from Google one may ignore the warning light.

As with many spam attacks we may notice some misspellings to fool the antispam software. Remember, never send anything to the Google “Customer Varifaction” department. We should at least assume that Google at least uses spellcheck…

False positive is the percentage of legit emails blocked by an antispam compared to the total number of blocked messages. In other words it is a measure that shows how effective a given antispam engine is in differentiating spam from normal messages.

False positive rate is inseparably connected with catch rate. However having a very high catch rate means nothing without low false positive rate, and vice versa. Every antispam can block 100% of spam simply by blocking every message that comes, but to do it without eliminating any legitimate message is what we want.

What is it worth for an AntiSpam engine?

In general, the smaller the false positive rate, the better AntiSpam you’ve got. It is hard to say what false positive rate is good enough, basically it should be somewhere around 0,04%, and everything below that value should work. You can try to estimate the value of this factor as I suggested with catch rate, but this may only give a very general impression.

What is a catch rate?

It is a very simple factor. Catch rate (or a percentage of spam blocked) is a number that describes the sheer efficiency of an antispam at identifying and blocking spam. In other words, a percentage of spam that has been stopped, with respect to the total number of spam messages that were sent to a given mailbox or mail server.

How to evaluate this factor and how much of spam blocked is enough?

Basically, the higher the catch rate the better the antispam is. If you are not interested in a freeware antispam you may as well filter out all the solutions that offer less than 90% catch rate.

If you already have an antispam solution and are looking for a better one, you should first find out your current anti-spam engine catch rate. Simply divide the number of spam blocked by the number of total spam (blocked and missed) and multiply by 100%. Are you still looking for a new antispam? :) 

How to estimate value of catch rate?

In other words, is one percentage point greater catch rate worth additional $100 or $200 a month? Perhaps it is worth much more? There is no easy answer to this. You could measure the cost of time wasted on processing spam (cumulative of all employees / mail users) then try to estimate how much time each pp of catch rate more would save. Compare the cost of better antispam with potential gain….

You can see why it is not truly measurable. But perhaps it’s worth to try, it’s your call.

Come back for more of my antispam buyer’s guide.

The sole purpose of this series of posts is to help you make informed decision. Show you which parameters and features are most important for your needs when evaluating antispam software.

Over the next few weeks I shall be writing about how to interpret all the parameters and how to value additional features, both the most basic ones as well as more complicated.

Come back for more of antispam buyer’s guide.