Effect of Social Scams on Engagement Initiatives
By Edwin Margulies
In addition to outright spam, there are many social posts that can clutter your engagement efforts and bog-down agent productivity. Here we will discuss how to identify scam-based posts and how to avoid them.
What are the top Social Media Scams? Let me count a few ways people have figured out how to "game" social networks. These in turn become a headache for legitimate social engagement for customer care practitioners.
First, there are robotic or scheduled posts that are essentially clones of one another. These seek to push a certain agenda and create fake trends without having a basis for "real people" behind them. For example, if you see: "I wish ABC company would just send me a new phone and give me unlimited data…" you might think that's a legitimate tweet. But if you see the same exact wording from dozens upon dozens of other authors in a short time frame, it's Robo-Cloning.
Robo-Cloning is perpetrated by individuals who hide behind fake twitter accounts and use those accounts to do "Silent ReTweets." A Silent ReTweet is a post that is just a repeat of what someone else said without the attribution of "RT @." The idea these clone scammers have is that if they tweet out the same message over and over again from fake accounts, some legitimate users may see them - think they are spontaneous and sincere messages from real people - and then legitimize the fake posts by ReTweeting them.
Another Scam are the folks you can actually pay to retweet or follow you for $27 or some other small amount. Here, unscrupulous business people can promote a product or service by paying shills to retweet or follow them - thus adding un-deserved credibility to their tweets. This scam has the same motivation as the "Clone Wars" scam but it's less transparent.
Influence Gamers try to fake-out influence algorithms by doing behaviors that pump up their influence score. One of these behaviors is to goose-up the score by animating their number of followers or likes. This is achieved by paying third parties to follow them or by aggressively following other people in the hopes they will be followed back. There are some people on the social web that have huge influence scores, but very little in the way of meaningful posts.
What's a Social Care Supervisor to Do?
As you can imagine, it can be frustrating to read posts that seem sincere on the surface but which end up being social scam posts. This wastes time for social care agents and keep them from their duty of reaching out and helping real customers with legitimate concerns. There are a few ways modern social engagement for customer care platforms can do battle with these scammers:
- Trending Cluster Identification
- Author Suppression
- Auto-Dispositioning based on Content
Trending Cluster Identification
Rent-A-ReTweet and Clone Tweets can easily be "caught" by systems that have the ability to track trending clusters. Trending clusters (sometimes called conversation topics) are "word clouds" that form based on Natural Language Processing (NLP) algorithms. These NLPs can be set to create a word cloud when the same phrase is repeated more than a predetermined number of times. Let's say a particular phrase like "they owe me a new phone" is repeated 100 times. That's no coincidence if there is not a ReTweet in the content. That's more than likely a sign of the clone scam.
Any social engagement system that automatically tags posts with a word cloud name or ID can be used to give agents a heads up that posts meeting a certain profile are fake. This is determined by clicking randomly on the source author and viewing a few native Twitter pages. If the link does not load up because the account has been closed, that is a sign that the author was fake. Or if the author has no original content, that is a sign that the author may be fake. Doing a spot check on posts caught by the word cloud is a great way to identify and then ignore these scammers.
Some modern social engagement for customer care platforms incorporate rules engines in addition to the NLP engines. These allow you to run rules for blacklisting content or even authors. For example, if you are getting a lot of scam-based content from a specific author, you can "black hole" that author using the rules-based engine to suppress tweets originating from that handle. This method is especially helpful if a particular author (real or not) is clogging your engagement platform with more than ten bogus tweets a day.
Auto-Dispositioning based on Content
By combining dynamic rules with dispositions, you can be pretty creative in ignoring and marking as "scam" any posts that meet a certain criteria. For example, if you have determined that a word cloud represents Robo-Clone or Silent ReTweet content, you can automatically close or delete those posts. It is a best practice; however, to bin suspected posts caught in this net so q quarterback can quickly scan through the list in case there are any legitimate posts mixed in with the bad ones.
Whether it's Silent ReTweets, Rent-A-ReTweets, or Influence Gamers, you don't want social network scammers getting in the way of your customer service operation. These people rob you and your colleagues of precious time that should instead be spent with customers. You can protect yourself from this threat by adopting the use of modern social engagement tools and practices. Take the time to inspect trending topics, build solid rules, and suppress fake authors and you will be in much better shape than just ignoring the problem.