There has been a new wave of spam these last few days, so I thought I’d again talk a bit about the tactics spammers use on Twitter.
Why is that many Twitter users see lots of spam come in over 3-7 days or longer and then mostly disappear?
This is because spammers try to send out a LOT of spam all at once over a few days, rather than spreading their efforts out over longer periods. That way, once Twitter starts to catch them, they’ve already sent out a lot of spam.
Their basic principle is to try to spam as many people as possible as quickly as possible before Twitter catches on and can effectively auto-suspend accounts. The idea is that more spam you send out quickly, the more people in total you will reach. But spammers have been getting better and better at pushing the limits on Twitter’s system of catching them, and sometimes now new spam networks can push for a couple of weeks before being well shut down.
As Twitter begins to figure out what kind of system and tweets are being used to send the spam, and starts to suspend the spamming accounts faster and faster, it become harder to get the new kind of spam through. After a bit, the spam network will stop sending out that particular kind of spam, and being preparing for their next big push, using what they have learned to try to create methods of propagating spam that are harder to detect.
Stopping spammers is not as easy as you might think
One reason it’s tough is that there are many networks, some with several thousand spam accounts, and many of these accounts are set up to act like real people. These “fake users trying to appear real” only send out small amounts of spam so they can’t easily be caught and get suspended. With the push of a button, the network operator can start to move spam out through these thousands of real-seeming accounts, each sending perhaps one or two a week.
There are many other approaches spammers take as well, depending on particular goals. A spammer might burn (allow to be suspended) a large group of accounts if they need to send a lot of spam out in a hurry, perhaps to try make a certain link be counted highly by sites that track the popularity of links across social networks. Trying to create a trending topic would be another example of trying to get something counted and promoted.
What makes it particularly difficult is that whatever Twitter does to stop spam, spammer take note of, and do less of it. It’s a cat-and-mouse game where spammers can keep trying new things once old methods stop working as well.
Twitter is upgrading their spam blocking
Twitter has said they are preparing to move to a preventative system that stops spam before it appears, but that they are still working to keep Twitter running well, and don’t have enough engineers in total to do everything. So yes, eventually, Twitter’s goal is to predict spam before it happens.