Investigating social media’s spam-bot problem

36 Shares 36 Flares ×

Need some users for your firm’s social media account? There’s a bot—er, app—for that. For a few dollars, marketers can populate their Facebook, Twitter, Instagram and even LinkedIn accounts with thousands (millions? DONE.) of followers. Like some highs, it feels good for a while, but marketers are increasingly coming around to the conclusion that the subsequent crash—counted in reputation damage and lost ROI—isn’t worth it.

Twitter and the now Facebook-owned Instagram have over the last few years launched high-profile purges where followers and users were outed as bots and culled. Numerous celebrities and big businesses lost millions of followers. But since those followers were never real, no problem, right?

Maybe.

The explosive growth of social media has been a boon to what’s called “social selling,” according to Caroline Carter, director of tech and business PR at Vancouver-based Brix Media Company. “The upside is if you buy fake followers and somebody chances across your social profile and they’re considering going with your service, especially in the B2B area, maybe they look at that really quickly. And they say, ‘OK, this company has a lot of followers, they must have a lot of customers, they must be legitimate.’”

While she’s quick to point out the advantage is a fleeting one, it speaks to the relative immaturity of the social media industry which has perhaps yet to develop enforceable standards despite the efforts of organizations like the Direct Marketing Association.

Partly, this is because it’s a technology problem: bots that increasingly mimic human behavior, leaving programmers and developers in a never-ending arms race with spammers who use the same tools and know-how to create their malicious programs.

Social media networks by the numbers

By Twitter’s own count, just under 5 percent of its 300 million active users may be fake accounts. That number could rise to as much as 8.5 percent depending on how loosely you want to define potentially suspicious activity. Facebook reported in its 2014 10-K filings that fake accounts numbered somewhere between 67 million to 137 million out of about 1.3 billion active users.

Twitter referred further press inquiries to its online support pages. Facebook did not respond to a request for comment.

LinkedIn’s 10-K SEC filing reports 347 million users as of Q4 2014 but does not reveal how many of these may be fraudulent accounts. The company also says it does not have a “reliable system for identifying and counting duplicate or fraudulent accounts.” LinkedIn did not respond by press time to a request for comment.

A June 2015 study by Italian researchers claimed 8.2 percentof Instagram’s 300 million users were in fact bots. Their investigation revealed an extensive world of bot vendors who, for anywhere from a few dollars to thousands, can provide not just varying numbers of fake followers but followers that demonstrate different levels of human-like complexity such as following each other, and posting on a variety of subjects to simulate breadth of interests. They even keep human hours, sleeping and waking when you do.

Describing this shadow industry as a “competitive black market,” the study identified 270 online web sites offering bulk sales of followers, views and the ever popular “likes” and estimated annual sales to be somewhere between US$76.5 million to US$197.5 million. (Instagram declined to comment except to say it uses the same anti-spam and bot tools as its parent, Facebook.)

Inside the bot wars

Those potential riches make it “worthwhile for people with marginally better programming and developer chops to get into the business of making bots,” says Finn Brunton, author of Spam: A Shadow History of the Internet and assistant professor at NYU’s Steinhardt Department of Media, Culture, and Communication. “You can very quickly start to see bots that are harder and harder to distinguish in various ways from human actors.”

But it’s not a war marketers and other social media professionals are giving up on. Some, like Tania Yuki, are blunt. The founder and CEO of New York-based social media analytics company Shareablee says “there’s no business value upside” to using bot spam. “It’s impossible to get bots to buy things or to develop loyal word of mouth for your product or to refer customers in their personal networks.”

The allure of padding one’s social media numbers is being attacked, in ironic fashion, on two fronts. One, by client-facing marketers’ need to protect their industry from an early death and two, by customer-facing marketers’ getting over the honeymoon phase of social media hype and realizing real money has to generate real results. Explains Yuki, “The biggest threat with having overblown numbers was that ultimately you ran the risk long term of brands … and people in the industry saying things like, ‘Oh well, social media isn’t effective,’ but really it was holding itself to an unrealistic standard.

“When you think you’ve engaged a million people or when you think you have 10 million fans but really you have half that you’re judging your media, your campaigns and your performance off of the big number. And suddenly your conversion rates feel disappointing.”

Brix Media’s Caroline Carter gives a specific example of how damaging this can get: “If you are buying retweets and followers, you run the risk of bots—which will retweet and repost things at random—highlighting the fact you are tampering with your numbers. There is  nothing worse than having a bot retweet a time-sensitive tweet over and over again three weeks later when it’s no longer relevant and when you’ve tagged somebody high profile or of importance in the body of the text.”

And there’s a payoff to doing it right. Yuki says the Twitter and Instagram purges were necessary steps in the evolution of social media into a trusted, transparent platform. “It really helped many marketers that we speak and work with feel so much more confident about the numbers they were getting.”

Fighting the spam

Data provided by Shareablee suggests the initial shock of the Instagram’s December 2014 cull immediately gave way to qualitative improvement. For a universe of 100,000 American brands, Shareablee’s tracking indicated an average decline of 6 percent among Instagram followers between December 18-19. But at the same time there was a 10 percent increase in engagement (e.g. likes and comments) on content posted by brands while the volume of that content remained flat.

That said, it’s not a problem that’s going away any time soon. According to threat detection company BrandProtect, “threat incidents,” which can be anything from phishing attacks to fraudulent accounts, across social media platforms (the company’s definition includes blog sites, YouTube, Craigslist and more) rose 125 percent from 2013 to 2014. CMO Greg Mancusi-Ungaro says Twitter and Facebook account for about one-third of total incidents and cautions that those platforms represent “only the tip of the social risk iceberg.”

Rob Waller is the founder and lead developer at UK-based StatusPeople, which is an app that identifies “fake users” on Twitter and offers related functionality such as bot blocking and analytics. He says bot customers run the gamut from small to big businesses and that, unfortunately, there are still some among them who would rather not know what’s really going on in their community. “A lot of people love our service because it helps them improve the service they deliver for their company, brands or users,” he says. “We have of course had others who have got quite upset with us for exposing them, essentially as frauds. I think though generally most marketers now understand the importance of quality over quantity.”

Because of the scale of social media communities, getting that quality is another challenge in itself. For large organizations such as  Facebook or Twitter, they can and do deploy small armies of engineers to fight bots and spam. But what’s a small business with far less resources to do?

Carter suggests a three-pronged approach combining honest self-appraisal, technology and some old-fashioned elbow grease. She advises managers to assess the business in terms of who your real target audience is and then choose platforms based on that—as opposed to thinking you have to be on all platforms. Use affordable tech solutions like Sprout Social or Hootsuite to manage your social media efforts and spend time first hand on your channels. It means “dedicating a certain amount of hours per week or per day to really owning those channels.”

For all the increasing sophistication of computer technology, the solution to the spam problem may ultimately lie with human intervention. Author Finn Brunton thinks human moderation is going to make a comeback in the near future precisely because bots are becoming too clever and nuanced for computer intelligence to discern real from unreal.

This raises the possibility of false positives—acting on which can be just as damaging as letting bots do their work. His advice to businesses is perhaps the most “social” of all: “A really smart move at this stage for a company that’s moving into [social media] would be to identify your heaviest, most beloved users and put them on the payroll. And turn them into people who can actually adjudicate these things, and who can do things like make sensible and thoughtful calls about how advertising is going to work.”

Photo via Flickr user The Hamster Factor

36 Shares Twitter 0 Facebook 0 Google+ 0 LinkedIn 36 Email -- 36 Flares ×
The following two tabs change content below.
Samson Okalow

Samson Okalow

Samson Okalow is a Toronto-based business writer. He has held editorial positions at a number of publications , including Canadian Business, Strategy and Adweek. He has contributed to numerous other publications including Playback, the Globe and Mail and Report on Business.