Last year, advertisers spent an estimated $1B on Instagram influencer marketing. That figure is expected to double by 2019.
There’s just one problem with that: It’s relatively easy to game the system by creating a fake account (AKA, a bot), buying fake followers and convincing brands to shell out thousands of dollars for sponsored posts on a completely made up account.
Now, a cottage industry of verification software companies is popping up, enlisting algorithms to give advertisers a better bang for their buck.
Down on the farm
With $500 and a dream, you can buy thousands of followers from an Instagram bot farm, and thousands of “likes” and comments from a click farm.
Once an account has 10k followers, it generally qualifies for influencer-marketing platforms (companies that hook up influencers with brands) — and from there, it’s easy money, so long as your fake account goes undetected.
A 2015 study found that 8% of Instagram accounts are fraudulent — and on average, 16.4% of the top 20 Instagram accounts’ followers are fake.
To catch a bot
Instagram has made some halfhearted attempts to crack down on fake accounts in recent years, motivating advertisers to turn to third-party account “verifiers” to ensure that accounts are real.
One such company, Dovetale, uses “more than 50 metrics” — including the rate of follows and likes, country of origin, the wording of bios, and excessive emoji use in comments — to determine an account’s legitimacy.
When an account’s post activity is graphed out, there are also clear markations: A legit post that gets popular organically usually sees a swell in traffic followed by a gradual decline; a fake post’s traffic shoots off the charts and falls to nothing within a few hours.
Numbers aren’t everything
Part of this issue stems from the fact that most advertisers’ commitments (and the rates they pay) are based on follower count, rather than engagement or other more qualitative markers.
Certain advertisers who invest heavily in influencer marketing care more about reach than substance. And unless the target demographic is “millennial bots,” that’s a problem.