3 Improvements to Spam Management in 4.6

[[R] NuclearGeneral]

Spam is as much a part of life on the internet as emoji overuse, serial GIF abuse and regretful tweeting. But I'm not here to talk about how I conduct myself online; I wanted to talk about three spam improvements coming to Invision Community 4.6. As you may be aware, Invision Community has its own Spam Defense functionality, which uses a mixture of crowdsourced data, publicly available data and our own special sauce to help reduce the number of spam accounts that get through the registration system. Invision Community also has several other tools to mitigate spam post-registration. These tools have served us well, but as spammers evolve, so must our systems. Here's what's coming to our next release. Spam Defense Scoring I can't divulge too much on our Spam Defense system lest we give spammers targeting Invision Community information that can assist them. Still, we have made several changes to our Spam Defense system. These include rebalancing the score thresholds, checking against known TOR networks and proxies and using other data in the public domain to inform our scoring decisions. Spam Defense Blocking The current implementation of our Spam Defense only allows options to either prevent registration entirely or put the registration in an approval queue. However, the days when Spam Bots stood out from normal registrations are long gone, and it's hard to know if an account in an approval queue is legitimate or not. In 4.6, we've added a new Spam Defense option that you can choose to allow the registration but put the new members into the posting approval queue, meaning their posts will need moderator approval before being published. This reduces the decision burden and makes it easier to take a chance on a low score from the Spam Defense system and review their posts before they are made public. Word Filters We have added a new option to the Word Filters to allow content containing specific words or phrases to be held for moderator approval where the author has less than a set threshold of posts. For example, you may notice an increase in spam targeting "CBD Oil" and add it to the word filter list to hold the content for moderator approval. This works great and captures a good number of spam posts; however, your regular members get frustrated when they want to talk about CDB Oil in their posts. This new option allows you to set a trust level for allowing these words to be used without capturing them for approval. We hope these three changes to our spam controls will reduce the level of spam you get in your community! I'd love to know what's the weirdest spam (that is safe for work!) you've seen in your community.

View the full article