> Christian Quest wrote: > > Are statistics that important ? > > > Fact: some spam will get past any antispam solution. The trick is that > you want to be using a solution that minimizes that amount > > So when a senior manager complains about a particular piece of spam that > reached her mailbox, and says "your system is useless - why aren't you > using 'vendor Y'?", you need to be able to refute that with... > Stats! :-) > > ...and if you find out your system is only catching 85% of spam, then it > is indeed time to be looking for a new solution (assuming you have > configured it/etc)
Just, unplug the senior manager and plug a junior one in. :) Well, right. Actually stats are matter (at least at some degree), but using things like graylisting doesn't mean you loose any statistic about non-retriers. There are greylisting products which do rely on an external sql db to store triplets, and you can easily count expired, open triplets (ie: the ones for which there was not a second delivery attempt) as accounting for spam+viruses. You just need a bit more effort in putting numbers together, but it isn't that difficult, anyway. Also, having to talk to a manager, you probably don't even need to show the whole design of the thing: you total receiveds are the SA checked messages + the expired, open triplets. You total blocked count is the SA+AV's blocked ones + the expired, open triples. Here it is: you'll get your numbers. Also note that, provided you don't have a too narrow time window in your GL product between first and second contact, you actually get really a low figure on GL's FPs (at, this seems to me). Regards, giampaolo > > -- > Cheers > > Jason Haar > Information Security Manager, Trimble Navigation Ltd. > Phone: +64 3 9635 377 Fax: +64 3 9635 417 > PGP Fingerprint: 7A2E 0407 C9A6 CAF6 2B9F 8422 C063 5EBB FE1D 66D1 >
