New types of spam appearing in our community – happening to anyone else? any suggestions?

Our community occasionally receives similar spam, except the responses don’t seem to be copied from anywhere. The initial post or reply will sound relevant, but then they edit it to add the spam links. Sometimes the edit gets caught, sometimes it doesn’t.

The poster is almost always a yahoo.com email address. I’ve also noticed some of these spam accounts rotate through a small set of profile pictures, so sometimes I can spot them and ban the account before they post anything.

Dealing with these accounts can be pretty frustrating, but it doesn’t seem much can be done other than having inSided block free email domains. but there are legitimate situations where our users would register with a personal gmail or yahoo address.

Maybe registration approvals for certain domains would be an appropriate way to start monitoring this a little more closely. Unfortunately, the feature doesn’t currently work this way.

Thanks for your thoughtful response, @Drew C.! I just reviewed the email addresses associated with these posts in our community; they’re almost exclusively gmail accounts. :( And the majority of our community members are using free email domains. 

Hi all,

sorry to hear that both of you have to deal with this kind of spam, I get that this is really annoying to you and your users, especially if they initially leave an impression of a genuine request for help.

The problem here is that it is not possible to check each post with the entire rest of the internet - there are so many posts being created each second that something like this is not really scalable. We do check of course if a similar post is existing on the community already, this would trigger the spam protection. And then there are also other means like the detection of certain unusual characters/words/urls which could trigger the system, depending on what’s considered likely to be spam content. Next to that, we of course check each new registration against a database of known spam email accounts, and as mentioned we also do blacklist certain questionable email hosts (and specific ones on request).

I think the bottom line here is that we are efficient in handling automatically created spam content (classic spam if you like), however tackling individual, personally created spam like the one described above is extremely tricky, as the spammer invests quite some time to stay undetected. As our systems become more advanced, so do spammers with their strategy. What’s problematic in this specific case is that there is not much standing out besides the fact that this content already exists on reddit. Make sure I will forward this to the product team so they can investigate if there is anything we can do about this.

Could you give us an indication of the frequency of these kind of incidents? It will help us to determine the impact compared to other types of spam.

@Drew C. You mentioned that you see spam connected to edits - I am curious if you have limited the edit time for new users to 60 min, and how you think about disabling editing for new users completely?

@Julian, there currently aren’t any options for the edit time for new users/ranks. Only for roles. I’d like to see the ability to disable links altogether until an account is at least a few days old and/or has a few posts. IMO, that would help in cases like these.

Thanks for your response, @Julian. That all makes sense. I’m seeing the same thing as @Drew C. – the spam only appears in the OP’s reply, when they’ve copied their original post verbatim but included two to four spam links within the text in random places. It seems that these sorts of posts should always be caught by the spam filter, but they are not. When I get a chance, I’ll look into exactly how often this is occurring (both the spam posts and their being caught by the spam filter) and will report back. Thank you!

Thanks for your feedback! 

Yes, I know, I think I did not word it correctly... I was just wondering if you give users unlimited editing time via a user role at start (probably not) or if you use the default (60 minutes). Of course giving new users unlimited editing time would increase the risk of spam.

I see what you mean, I am a little bit worried that new users would be confused by this (if not pointed out specifically). Regular users who just joined and who would like to use links to refer to some external information might be thrown off by this… but maybe I am just too worried. In an ideal situation it would be better to detect spam and block it without limiting functionality for regular users.

Thanks for doing that! It surely is challenging for the system to detect which hyperlinks are spam and which are not, if I get more info on how this part works (and if it can be improved) I will let you know!

It looks like we had about 10 users doing this kind of spamming, over a period of about 1 month – most of the spam occurred in mid-late December and early January. Most of these spammers posted 2–4 times (that includes both topics & replies). I initiated this post when I felt this sort of thing was increasing in frequency; now we’ve had a relatively quiet spell so I’m hopeful it will last! Thanks again!