“We need to optimise the grey area of the commercial newsletter.”

Matthias Leisi is a chief strategist at SEPPmail and one of the key players responsible for its rapid breakthrough into the Champions League of email security providers. The spam filter was named the world’s best by the renowned Virus Bulletin, which the CTO said he was proud of, while also being aware that there was “room for improvement”. One of several good reasons, therefore, for the SEPPmail blog to ask Matthias Leisi a few questions about this topic.

Matthias, congratulations to you and your team: the great performance of the SEPPmail.cloud solution’s filter is plain to see. What are you particularly proud of?

Our accuracy, because that’s what it’s about. And actually, it’s pretty simple. This is an area in which we have performed well – and that could be the secret to our success. After all, it shows that we are on the case all the time. For us – and for our customers.

In the press release, you wrote that there is “room for improvement”. Where exactly do you consider this to be?

The grey area of the commercial newsletter.

What do you mean by that?

Cases in which it is difficult to distinguish between an email which is spam and one which isn’t, even for someone who is usually in the know.

What are you doing about this?

We have now been implementing user feedback data sets for a very long time, and we analyze this data very carefully. After all, genuine feedback can improve our processes more than automatic procedures.

What are the technical hurdles to achieving these goals?

The hurdles are more content-based than technical. The hurdles are collecting and organising data both better and more specifically. To see how the users behave, and to draw the conclusions accordingly.

How rapidly do you need to develop the filters to ensure that they stay at the “state-of-the-art”? After all, it is reported time and again that criminals have very quick development cycles.

There are two factors which we urgently need to improve. The manual, important work that we no longer follow up on because the information changes second-by-second. And the weighting of the automatic procedures. In the case of an extended time frame in particular, the findings of a wave of spam can destroy more than they save over an extended time frame.

What advice would you give companies which already have a different filter in use today? What should they pay particular attention to?

They should pay attention to how much malware and spam they are receiving. And then analyse exactly what the data contains. Our filters are more targeted, and particularly aimed at SMEs with much more accurate details in their spam history. All in all, however, they should certainly do as much as testing as possible.