The aim of using two programs; POPFile and Daisy POPFile, together was to increase the proportion of 'spam' recognised as such to a maximum.

Our POPFile is now showing a success rate of over 97 %, which is probably a bit lower than the actual as when we get a new customer, we create a new 'bucket' and then reclassify the message.

This of course counts as an error. But it is not!

We have also learned a lot more about where our 'spam' originates, how it is sent, and how to identify messages that are of a lower level 'spam' that can be stopped.

This is a very much a similar conclusion that you will get, when you use Daisy or another data analysis program on your problem.