Have Brands Been Paying Leading Kenyan Influencers for Fake Engagement?
Earlier this year, we released data on the perceptions and attitudes of Kenyans towards influencers. Key points that stood in this context were the fact that;
- Trust in influencers to endorse brands and products is low
- Few digitally active Kenyans can attest to have bought products as a result of their interactions with influencers/celebrities online
Consumers expressed distrust in influencers due to a perception that they have sold their souls and credibility for advertising dollars. This arms race to make money off brands dabbling in influencer marketing has led many would-be influencers to use all available means to build their online profiles and following to appeal to brands. Unfortunately, as happens with most arms races, people will look to find unfair advantages to get ahead of the pack. In our case, that means influencer fraud.
Influencer fraud comes in many shapes but we would like to specifically focus on the use of online fake accounts to prop up an influencer’s popularity, either through the purchase of fake followers or fake engagement. We focused our study on Instagram for this study due to its popularity among Kenyan influencers. We carried out audits on all the 400 plus influencers we track on Instagram and what we found was astounding:
From our analysis, there is no single influencer who did not have bot activity happening on their account. This is a widespread problem in the Kenyan influencer market, cutting across all influencer categories from politicians, musicians, bloggers to news anchors.
Where Are Fake Accounts Found?
Fake accounts by some estimates, make up as many as 95 million of Instagram’s users and 48 million of Twitter’s users. These bots have become currency in the black market of online influence. A lot of them are designed to look and behave like real accounts by real people. The goal is that when they follow, like or comment on an influencer’s account, it comes off like real, measurable engagement. However, there are some key smoking guns that can quickly reveal a fake account. Firstly, you will often find that they have a disproportionate number of followers versus their number of posts. Growing an account organically requires a large number of quality content shared over time that attracts following.
Secondly, the nature of the posts is revealing. A common tactic among fake accounts involves identity theft where the fake creator will take pictures off a real account and use them to set up the fake account to make it look real. However, they will often only use 5–10 pictures that have a very similar aesthetic, and then infuse the account with 5,000 to 10,000 fake followers. Thirdly, a quick scan of their followers also reveals that these accounts are mainly followed by their fellow fake accounts.
As with many things, a simple Google search can reveal a lot about this shadowy industry. Fake followers and engagement are sold openly in online marketplaces with prices ranging from as low as USD 5 to USD 100 for higher quality accounts. Kenyans have joined this bandwagon with some individuals and companies asking for as little as Ksh.5,000 (50 USD) for 10,000 fake followers. However, most Kenyan vendors are resellers of accounts often created in Asia or Eastern Europe.
How do influencers use bots to fake their popularity?
- The Drip Method: social media fraud has been a long-running issue and as such most social media sites have put in place measures to flag suspicious activity like a sudden surge in followers for no particular reason or a surge in engagement on an account at regular intervals. To counter this, bot vendors now offer their services in a drip format, increasing engagement and following on a day by day basis until the full transaction is complete. The most surprising part is that they actually keep their word and deliver as advertised.
- The PR Disguise: As they say, any publicity is good publicity and it turns out that any publicity presents a great chance to inject fake followers to your account. For example, if an influencer is involved in a scandal and it goes viral, it would appear natural that they would be of sudden interest to many people thus justifying a sudden increase in following even if it comes from fake accounts. This can then be easily followed up with increased engagement as well.
- Gaming the System: Vendors have also come up with unscrupulous ways of gaming the Instagram algorithm. A common method involves using fake likes and comments to boost engagement and increase an influencer’s chances of appearing on Instagram’s Explore page, which is known to significantly increase exposure on the platform. This manufactured activity typically happens within a few minutes after the influencer has posted content on their page, thus signalling to Instagram that this person must be really “popular” and therefore the system pushes the content even further.
The Most Affected Categories
From our analysis, the Mainstream Music, Gospel Music, Radio, Socialites and Comedy categories appear to have the highest percentage of fake activity in Kenya.
Social media influence today largely equals real world influence. This is a global trend and Kenya is no exception. This makes it a lot more appealing and almost vital especially for categories like musicians, comedians or socialites whose relevance is judged by their most recent hit. This means there is daily mounting pressure on these individuals to display social media activity that matches their perceived social ranking.
Micro influencers generally seem to have a much lower amount of fake followers and fake engagement than more popular influencers and celebrities synonymous with mainstream media. It would make sense that they have lesser pressure at their level to prop up their influence. However, as they grow bigger they will most likely be inducted into influencer fraud as well.
What Does This Mean For Influencer Marketing?
Influencer marketing has exploded globally and Kenya has not been left out. While it still represents a small portion, brands are spending more of their marketing budgets each year on influencers. There is already wide skepticism across the marketing industry about the effectiveness of influencer marketing on consumer perception and sales conversion. From the results of our study, it seems that the majority of influencers are feeling this pressure and have unfortunately turned to unscrupulous methods to keep attracting advertising dollars. As a result, we now have brands spending money on an unregulated industry that could be rife with fraud.
On the other hand we have consumers, who continue to feel alienated and lose trust in influencers. With fraud fuelling influence, consumers get confused or duped into perceiving certain people or things as popular even though they actually are not. This in part explains why over 80% of consumers are not making purchases based on influencer marketing. If the influence is not real and an influencer is not connecting with real people, it adds up that they cannot effectively deliver positive brand perception or drive purchases.
As mentioned earlier, influencer marketing is still largely unregulated. As such, it falls on brands to properly audit the influencers they choose to work with. It is now imperative that brands do thorough background checks on following and engagement figures, audit past performance of an influencer with other brand campaigns and their industry reputation in general. Brands must also have clearly defined KPIs that can only be delivered by someone who has actual influence among their target audience.
We created a test account on Instagram and purchased both fake followers and fake likes for the account, taking note of the offerings that the various vendors made to help us boost our account. From there we created and algorithm that looked at the gained engagement and followers on the test account, taking into account things like number of posts, follower/following counts, the comment to like ratio, Like counts, username and whether the account is public or private. We then used the same algorithm on a data set of bots that we had purchased to which it identified 73% of the bots. We also trained the algorithm to a dataset of real accounts to which it correctly identified 80% of them. This was done across a random sample of 350 accounts per individual and repeated 3 times to come up with an average figure. Samples were also taken on the engagement each individual got and the algorithm was ran against the accounts using the same indicators.
How does this affect our Kingmaker Influencer Rankings?
Based on this revelation, we’ve had to make necessary updates in how we rank celebrities and influencers in our Kingmaker tool. We have begun discounting individuals with significant fake followers and engagement as per our findings on their accounts. In addition, we shall also begin to include talkability (the number of mentions a celebrity/influencers has garnered online in a specific period) into our rankings to give a more wholesome view of the popularity and influencer of an individual.