Stay marketing and tech-savvy. Get the latest in martech - subscribe to MarTech Today.
Fohr Card launches a service to identify bots that follow Instagram influencers
On average, the influencer platform found that 7.8 percent of those followers are fake -- and it is lowering brands’ fees accordingly.
Influencers’ level of influence is largely dependent on how many followers they have.
But how many of those followers are actually, you know, human?
Influencer platform Fohr Card is today releasing a new tool that it says can accurately determine which followers of Instagram influencers are bots — and then clients’ fees to the platform are reduced proportionately.
The New York City-based company says it is the only one to charge brands for the activity of verified human followers on Instagram, and not for bots. On average, Fohr Card found that 7.8 percent of the followers of its Instagram influencers are fake.
Fohr Card also manages influencers on YouTube, blogs, Twitter, Tumblr and Facebook, but this anti-bot-follower tool is initially targeted at Instagram followers.
Each Fohr Card influencer grants permission to scan their account’s followers via an API. A Follower Health Score is determined by automated scanning of followers’ usernames, bios, activities on other social platforms, how recent their posts are, and the ratio of their own followers compared to the number of Instagram users they follow.
Nord said some clues are giveaways, such as a skimpy bio that says, in essence, “I follow and you follow back.” Or a user name might be something like Instagram374.
To figure out some of the subtler bot strategies, Fohr Card bought bot followers from every service it could find. And they’re cheap. Nord says you can buy 50,000 followers from bot farms in places like Romania or the Philippines for $200 total.
The resulting Follower Health Score for each follower — generated after analyzing about 20 million Instagram accounts for followers — ranges from -8 to 8. Nord says his company is “100 percent certain” that a follower is a bot when it scores at the very bottom, or that it is a human when it scores at the top.
Each Instagram influencer is also given a score from 0 to 100, reflecting Fohr Card’s assessment of the quality of that influencer’s followers. Here’s a screen shot from the bot-sniffing tool:
Nord told me that “7.8 percent is not a large amount,” adding that he’s “happy it’s not super widespread.”
But his company has pointed out that the resulting damage is potentially large, given estimates that influencer marketing on Instagram is currently worth about a billion dollars annually, possibly doubling next year. That would mean about $78 million of Instagram influencer budgets are being wasted on bots.
Nord said that “nobody has built a product like this.” Previously, his company focused on engagement metrics and would note fake followers when the engagement rates indicated something fishy.
He pointed to only one company doing something in the same ballpark. FollowerCheck says it takes a random sample of 150 Instagram followers and assesses if they are real by checking profile completeness and followers-to-followed ratio.
Fohr Card argues that it would not be possible to do such random checks unless FollowerCheck had API access, which Fohr Card says it doesn’t. Nord told me his engineers assume FollowerCheck is looking at the last 100 followers for any influencer.
By contrast, Nord said, his company does a more comprehensive analysis of every follower for its Instagram influencers.
How much are influencers to blame for fake followers?
Nord said many influencers acquire fake followers without any effort on their part, because bots want to establish credibility by following a diverse group. But, he acknowledged, some influencers probably subscribe to bot farms to pump up their numbers.
Why aren’t other influencer platforms undertaking similar efforts?
Nord said there’s a “look the other way” attitude in the influencer industry, since the bot influencer problem — if Fohr Card’s finding of 7.8 percent on Instagram is any indication — has not yet reached crisis levels. He also compared it to “athletes who are doping, [where] the NFL isn’t interested in catching them” because the illegal practice helps their game performance.
While Fohr Card may extend its anti-bot-follower tool to other platforms, Nord said the company’s next goal is to “weed out purchased engagement,” where bots generate likes and comments.