It seems that every year, the spammers change up their tactics to see how they can get past our content filters or automated spam system.
This year, the trend has been to copy a legitimate post from another user and repost it but then the spammer will:
- Embed spam links where legit links were once used
- Hide spam links / terms in the HTML
- Hide spam links / terms by changing the color to match the background
These tactics are nothing new, but the frequency at which is occurring is new for us. Unfortunately, these types of posts are often missed by the automated spam system as they appear to be legitimate. We are unable to "mark the post as spam" as then the system could start to learn incorrectly on legitimate words that we don't want "blacklisted". That results in a manual process to remove the spam.
We also see that other users do not recognize when these posts are spam and they interact with the spammer or even click the links in the posts. This could result in malicious or inappropriate content being presented to the user.
Ideally, the spam system would take other variables into account such as account creation date, if the content is a duplicate of existing content, etc. But since we don't have that...
Any tips for other approaches to combat these particular spam tactics?
Content filters have always been our best bet, but with the spam URLs typically changing, it isn't a catch-all.
Recently we found spammers using tags to insert certain keywords or sentences. I’m not sure if I understand the use case. They didn’t appear to be links just some type of advertisement. Has anybody else seen that?
It just kind of feels like it’s useless even for Spammers
@Wendy_S not knowing what the words were, my guess is trying to get some Google juice out of it? I haven't seem them use tags for that....yet. I've seen them do it in their profile fields.
@Wendy_S- we receive that type of spam too. I assume it's some type of SEO spam? Oddly, these users were able to rank up when doing this, and getting to bypass the spam system due to the roles they earned with the rank. We reported this bug to support back in Nov. 2017.
I spotted this technique here in Atlas... replacing the username link with the spam link!
We have been under attack by scammers for a while now. The post fake support phone numbers in the email replies pretending to be legit support.
you know how that story may end for many customers.
they use. Symbols or small images to make up the phonenumber so that we have to constantly catch them and add to the filters. Of course scammers are most active during our low mod hours such as evening and weekends.
we have removed the content preview from the email notifications and set up an alert on all posts to warn users but I think we could do more to scare these scammers away for good but we are running out of ideas.
if you have any tips, gladly accept them