In a recent episode of The Digital Marketing Podcast, ethical entrepreneur Anders Ankarlid told Target Internet about his investigation into the inflation of influencers’ engagement metrics. Ankarlid learnt that the $744mn of spend on Instagram influencer marketing could be being wasted on bots and fake engagement.
This article brings together some of Ankarlid’s key findings and presents them in the context of a new set of best practice guidelines on how to work with influencers on Instagram. We hope this will help brands fine-tune their influencer relationships and get a fair return on their investment.
When brands assess which influencers to work with, two of the key factors usually considered are the number of followers the influencer has, and the average number of likes and comments their posts attract.
Anders Ankarlid, founder of the eco-friendly everyday goods brand A Good Company, made waves back in July with the publication of his study into the extent of inflation by Instagram influencers, The $744 M Influencer Marketing Scam.
A Good Company had been ploughing money into an influencer marketing programme involving about 4,000 micro-influencers, but the results were disappointing.
“The expected effects of influencer marketing were missing – partly brand reputation, but mostly sales,” Ankarlid told The Digital Marketing Podcast.
“We had a potential reach of maybe 50 million people, which is like the UK audience on Facebook, but sales were not where we expected.”
Ankarlid decided to investigate the reasons behind his influencers’ under-performance.
A Good Company asked its micro-influencer partners to respond to a survey with about 12 questions. Many of the questions were easy, but some dug into the influencers’ approach to gaining followers and engagement. The influencers, most of them from the UK and elsewhere in Europe, were assured their responses would remain anonymous. About 400 responded – 10% of the micro-influencers working with the company at that time.
“When I got the initial results from the study, I was in complete shock,” says Ankarlid.
“When we asked them “Have you ever used any type of service to boost engagement with your Instagram account?”, 36% said we have not. That means a little above 60% have done something, and several said they have done many things to increase their engagement rate, follower rate and so on.”
Influencers who want to artificially inflate their engagement figures have plenty of options to do just that, ranging from mutual arrangements with fellow influencers, to the use of tools that exploit fake profiles – often referred to as ‘bots’ – to artificially boost engagement.
Best practice learning: Instagram engagement figures seem to be inflated by many influencers. Accordingly, marketers should lower their expectations of influencers, and increase their scrutiny of new influencer partners.
Shocked at the results of his qualitative study, Ankarlid decided to commission farther-reaching research into the practice of engagement inflation.
“We contacted an external company – an analytic tool called HypeAuditor, based in San Francisco,” says Ankarlid.
“They looked into 1.8mn accounts worldwide using an AI, and they presented us with a massive Excel file distributed by country, follower number, bots, engagement pods and so on. We were dumbstruck with their findings.”
According to Ankarlid, the insights from the report included:
“The one number that has struck me the most is when you compare the gross users in millions,” says Ankarlid.
“In the US they have 110mn active users per month, but when you take away the bots, you end up with 60mn. So in total, 45% could be fraudulent, meaning the market of all those people could be 45% less. That’s still a significant number. You have maybe 600mn people using Instagram actively each month – but it’s not a billion people using it.
“A lot of marketers are shifting to investing a lot on Instagram, and if 45% is fake, then maybe the market is not that important. It’s a number to take into consideration because you can’t really say if you go in full throttle on Instagram in the UK that you will reach 25mn active users per month since there are not 25mn active users per month.”
More detail on A Good Company’s report into Instagram engagement is available via Anders Ankarlid’s blog on Medium.
Best practice learning: marketers should take social media usage figures with a pinch of salt since a large share could be made up of fake engagement. This presents a problem for social media platforms that wish to reduce fake engagement: by removing bots, they could end up making their usage figures appear to shrink.
A Good Company’s report indicates micro-influencers are likely to have particularly inflated engagement figures.
One method used by such influencers is participation in so-called ‘engagement pods’ – groups of micro-influencers who agree to engage with each other’s profiles and posts for mutual benefit. Anders Ankarlid explains:
“An engagement pod is a ring of engagement. Let’s say a group of people has one Instagram account each. They go in a circle of trust and say whatever you put out online, I will watch it, like it and comment on it.
“It’s different from hiring a bot – it’s authentic people – but it’s not authentic engagement. This is the biggest thing our influencers were doing. 40% in our first study said they were using engagement pods from time-to-time.”
The use of engagement pods offers an explanation as to why micro-influencers often have impressive engagement stats.
“Compared to Kim Kardashian who has an engagement rate of something like 0.5% or 1%, micro-influencers often have an engagement rate of around 10%, meaning that for every post you put online, 10% will engage with it,” says Ankarlid.
“But, if 9% of these engagements come from friends and family, then you really have 1% engagement, and if you post to an audience of maybe 2,000 people, that’s maybe not a whole lot of people that actually engage with you.”
Micro-influencers using engagement pods seem relatively blameless, compared to others who use services that deliver artificial engagement from bots.
“A lot of those influencers get affected by one another, and it’s easy to imagine thinking okay, maybe I should also do this,” says Ankarlid.
“You just go into this pod, then in two seconds your engagement rates are up, and suddenly brands reach out to you and say okay, it looks like you have a super-decent audience, maybe we should do a collaboration together.”
However, A Good Company’s research indicates micro-influencer engagement pods could have the greatest inflationary effect of all, relative to the engagement figures used to win contracts with advertisers.
Best practice learning: content engagement is perhaps not the best way to establish how influential a micro-influencer is, since many of this group seem to participate in engagement pods to inflate their stats.
Instagram has been trialling a new update that could profoundly alter the influencer marketing industry. Users in Australia are no longer able to see the like count on other users’ content, neither on their feed, nor by clicking a ‘Liked by’ link beneath an image.
According to a spokesperson for Instagram’s owner Facebook, removing the like count is intended to “remove the pressure of how many likes a post will receive, so you can focus on sharing the things you love.”
We can’t help but wonder if the trial might also have something to do with Instagram’s ongoing crackdown on “inauthentic activity”. An artificial like would clearly be easier to pass off as genuine than an artificial comment. By removing the easier option of artificial likes, Instagram have given the app’s users – including marketers who are thinking about working with influencers – a better chance of differentiating between true and false engagement, at a glance.
Anders Ankarlid strongly supports Instagram’s like-free trial.
“That’s a super good thing to do. The whole glossiness around influencers with comments and likes has to stop,” he says.
“Leave likes, leave all of this hunting and chasing for comments and engagement rates, and really dig deeper and look for brands that really can connect with people.”
Best practice learning: An end to like counts could make it harder for influencers to impress brands with inflated engagement figures.
Since conducting its study into engagement-inflating tactics on Instagram, A Good Company has stopped giving away free product samples to influencers.
“We used to hand out loads of freebies,” says Ankarlid.
“The higher the engagement rate the influencer had, the more interested we were as marketers to give them a freebie, so they could do a review, do a post about it, do a story or something. We’d see they had an engagement rate of 10% and think awesome, we’ll reach maybe 200 people who surely are interested in our brand.”
In light of the findings from A Good Company’s study, many micro-influencers’ real engagement rate could be closer to 1% than the 10% sometimes stated. This may have meant that the free gifts given out by the brand engaged as few as 20 legitimately interested people each – not enough to guarantee the required return-on-investment.
Anders Ankarlid was shocked to find this out – but A Good Company didn’t stop distributing products to influencers altogether. The company now ships products to influencers at heavily discounted prices; not for free.
“We decided to drop off this a little bit, then work with influencers who really are engaged with the brand, who were willing to pay for the product and service,” says Ankarlid.
“They can have a huge discount, but they must be willing to pay for it.”
Best practice learning: offering discounts, not free gifts, could be an effective way to identify influencers who are truly compatible with your brand.
Discovering the scale of engagement-inflation among influencers could easily have spelt the end for A Good Company’s micro-influencer collaborations.
However, Ankarlid and his colleagues concluded that this approach could still achieve great results – if they could identify influencers who could deliver success more reliably.
“We have maybe ten influencers a day contacting us and applying to our programme. They brag about the numbers; we don’t care about the numbers,” says Ankarlid.
“What we care about is that they are authentic to our company, that they care – and the only way to prove that is that they are willing to pay something for the product itself.
“Secondly, we can pay for when we are asking them to do a certain campaign, to promote something. It’s much better to pay them for the work required than based on the number of likes they are achieving. That stops people using a bot and earning maybe £200 in 2 minutes.”
Best practice learning: to get reliable value from influencers, get them to pay for product samples; and pay them based on the work they do for you, not their engagement stats.
In October 2011, The Guardian reported on the tactics newspaper and magazine publishers were using to inflate their circulation figures. In one high profile example, Rupert Murdoch’s Wall Street Journal Europe was found to have inflated its circulation for the period May 2009 to April 2011 by 41%, through a deal with a Dutch company that agreed to purchase thousands of copies of WSJE for as little as €0.01 each.
What this tells us is that the inflation of engagement metrics is nothing new. Some advertising partners – whether they be a newspaper publisher, an influencer, or whatever comes next – will always use whatever methods they can to increase engagement and make themselves more attractive to advertisers.
As marketers, we must be ready to analyse potential partners carefully and keep up-to-date on the tactics used by advertising partners to boost their stats. Influencer collaborations can deliver great results – but we need to be mindful that the value of an influencer cannot reliably be calculated based on the numbers next to their posts.
iOS14 has changed the game for advertisers using Facebook – and also for the social media giant itself. In this article, we’ll talk about the big change to Apple's operating system that advertisers need to know about.