Both of these feeds have been essentially useless the past two days due to a stream of bots posting endless, lengthy spam posts. I tagged several dozen as spam yesterday but that's a rather time consuming method of cleaning up global and has no effect on chatter. If this is to be a community or self-moderated site, how can we improve the tools available without giving trolls too much ability to interfere with beneficial content.

insomniasexx:

This is a response to like..everything in this thread.

I do not like 24 hour rule. I know I first created my account on Reddit, Hubski, and others because I finally found a discussion I could add value to.

Forums that have a manual review process....well....even the ones I have signed up for I never used. By the time I get approved, I'm doing something else.

The barrier to entry for using Hubski is already surprisingly high. New users consistently describe Hubski as "clique-y" and not the most friendly place for newcomers. By preventing users from commenting for 24 hours, we will see a significant drop off after users create an account.

I don't have stats on this so imagine:

10% of people who visit Hubski, ever consider making an account.

Of the 10% who consider making an account, only 10% of those actually make the account and comment or post.

Of those 10% who actually comment or post, only 10% come back again the next day and make another post or comment.

So with these imaginary numbers, .1% of new people who visit Hubski will ever comment or post twice. By implementing a 24 hour rule, we may see that .1% drop to .01% or even .001%. It's not that new users won't create the account. It is that the new users who do create an account will never post or comment.

If you think I'm being ridiculous with those numbers, I'm not. Conventional wisdom says that a good conversion rate is somewhere around 2% to 5%. In fact, having a conversion rate above 10% is basically unheard of, and places you with the unicorns and other fantasy animals the internet world uses to describe amazingness.

So....how do we prevent spam without causing a massive drop off in legitimate users? You do something that is very easy for a normal human being to do but impossible for a bot to do and tedious for a spammer. Originally, they figured Captchas were the way to go but unfortunately.....they don't really stop anything beyond basic bots. Bots are reading Captchas today.

Or... you let spammers be spammers and let everyone else call them spammers. Right now, in order for a new user, even a spammer, to be flagged as spam, multiple people must do the following things:

1. Go to global.

2. See spam.

3. Click the user's name.

4. Scroll down, find the little check boxes, and click filter or block.

5. Repeat.

I propose we do it differently. Put that filter button on the global fucking feed. If the user is under 24 hours old and they get 2 filters, they're automatically globally filtered. They want to not be globally filtered? Then send us a message and we'll look into your shit.

This does a bunch of things:

1. No additional barrier to entry for new users.

2. No need for the Hubski team to be moderating and deleting spam -- we only have to deal with false positives.

Additionally, putting a limit on the number of posts/comments a new user can make in a set amount of time will decrease the amount of spam. But, admittedly, it will make it harder to tell if it's spam, self-promotion, or just really bad content and it may drive spammers to simply make new accounts. (This is actually quite funny: most spam accounts limit themselves to 2-3 posts and then they're gone. We have no mechanism for this but they do it anyways)


posted 3123 days ago