What are the typical best practices for handling reporting/flagging inappropriate content on a social/UGC site?

Question

I'm designing a site that is driven by user-generated content and will not be heavily moderated. So users need to be able to flag/report inappropriate content and users. I've read that if the asset is reported by 5 distinct IP addresses or users, it should be taken down automatically. I've also seen examples where the site administrators "look into" a reported asset before removing it. But what happens to the user who's being reported? Are they emailed? Notified the next time they sign in? Are those assets still visible to them, but no one else? Is their account frozen if they're reported too many times? Any and all best practice recommendations are greatly appreciated.

Answer 1

Slightly off topic - The nicest feature I've seen on this kind of site (if it allows online 'discussion') is an 'Ignore' Button so that users can choose to ignore the postings of a particular user (their posts then show afterwards as 'Ignored User').

This adds a "pre-reporting" stage so that users who post content which could end up producing more heated arguments is diffused at an earlier stage, as argument is prevented from occuring.

This has the knock on effect that there is less content which gets reported and therefore needs dealing with.

Answer 2

unless the user did a really really bad spam i will take that user down,

But if they did something bad (like a bad comment), then i won't take down user even they did that > 5 times cos it's internet & we need freedom. However, I have a solution. That is you can have many level of reputation.

-Level 1: very good

-Level 2: good

-Level 3: bad

-Level 4: really bad

If a comment is in Level 3 or 4, & if someone click on it, you make a popup saying

"that this comment is not polite, do you still want to view?" Yes No
"are you > 18 years old" Yes No
... more question depending on the "BAD" level

If the user clicks all the Yes, then show the comment. If user click No for 1 question, then won't show the comment.

The more "BAD"s then you put more questions.

By doing this you save a lot of time to managing the bad comment since all bad comments was managed automatically.

1 more, how we know a bad comment? if a comment was voted down many time then it is bad, if more votes down, then more "BAD" level applied

Answer 3

In my opinion, the simpler and more discrete way to manage these kinds of alert is to include it in the notation.

^{download bmml source – Wireframes created with Balsamiq Mockups}