Active Moderation

Spam is one thing, but moderation is another beast entirely. As a message board moderator, or somebody operating a comment-enabled blog, you want to prevent flame wars and perform some sort of moderation on inappropriate posts. Depending on your subject matter, you might have any of a number of different problems - it could be copyrighted images, links to bad sites, run of the mill spam, or perhaps even the promotion of a competitive entity.

With low volume, it's easy to keep up to date with the things people are posting and manually moderate once a day or on an as needed basis. Higher volumes generally mean more moderators.

Corey Doctorow had a story published on the escape pod podcast recently and part of that story got the hamster in my head running around. Why not active moderation? You implement a back end process that identifies postings that potentially need moderation, and that system notifies the moderator when something potentially inappropriate comes across that needs human review.

Best of Both Worlds

One of the problems with automated moderation controls, especially heuristic ones, is that they end up weeding out the good with the bad. The problem with doing it the manual way is that the moderator can get over-run with work so much so that moderation ends up taking a back seat to other tasks.

Workflow can be designed to present moderators with a subset of information to review as it comes through the system - reducing the amount of critical review that is necessary, and decreasing the time period between when a problem presents itself and when it is resolved. Once implemented it can be augmented and tuned to the point where it sees a very high success rate at identifying those messages that need to be moderated.

Heck, you could even push out moderation requests to a service like Amazon's mechanical turk and get rid of professional moderators all together.

Talk About Active Moderation