Five days ago, we tried a new approach to spam.
It didn't work as intended.
In short, we created a one click link to tag posts with the community tag #spam in the global feed. New users also filtered #spam by default.
The intention was to leverage existing behavior to more easily send spam posts into the abyss. The effect was that we codified the community spam tag as the principle means to avoid spammy posts, and made moderation less individual. The spam tag became more powerful, and its application more controversial.
Today we are trying a different approach that we feel is more aligned with the nature of Hubski and its characteristic individual moderation.
In the global feed, you can now find a 'filter' link on posts. Clicking this link will filter the user, removing their posts from your feeds.
We have long used filter as one signal to bring spammers to our attention, and will continue to do so. Thus, filtering a spammy account does help remove spammers from Hubski. However, unlike the #spam community tag, it does not have an immediate effect upon everyone else.
You can still community tag posts with #spam. However, like with other community tags, you will have to open the post first to do so.
As always feedback is welcome.
I'd like to add that tags provide a wealth of information, and should be leveraged to better enable discovery and curation. We have a number of ideas in this regard, many of which have been generated by you all. We are going to be experimenting with improving the utility of tags in the not-too-distant future. Not having a relational database has been a hindrance we are close to shedding.
I think the filter approach is a lot more nuanced and is going to give you much better results. Question: based on what you see on the back end, is it more effective to filter the user or the domain? Because in the end, it's the domain they're driving (albeit they're mostly shotgunning a gajillion domains these days).
There seems to be an equal share of users spamming multiple domains, and multiple users spamming the same domain. It's a bit easier to target users as some content platforms like medium, imgur, etc. require exceptions. We can use the filtering of users to highlight spammy domains, however. I just added the user's name to the filter link to make it clear it is the user that you are targeting.Question: based on what you see on the back end, is it more effective to filter the user or the domain?
Sorry, but I'm not seeing the tag. Am I too new to get the link or am I just missing it? I only see hide, save and stick next to posts. I see links next to comments but there are no comments in spam. edit: It's probably right in front of me, but I'm not seeing it.