The rise of ‘micro-influencers’, social media personalities whose followers number in the thousands, rather than the millions, has given brands a new way to advertise to audiences through endorsements or brand-sponsored content. While they may not reach huge audiences, they are significantly cheaper than the big-name influencers (it emerged last week that Kim Kardashian-West can charge up to $500,000 for a single Instagram post), and can often give brands easy access to specific audiences.
But it throws up new challenges too. If a sports brand wants to run an influencer campaign with just a few big name stars, the process is relatively straightforward. The brand will likely already know athletes who fit their values, and given they’re only working with a handful of people, they can maintain a lot of control over the content produced, and track performance.
Working with micro-influencers though requires a lot more heavy lifting – from identifying and vetting influencers, to overseeing content creation, to making sure they’re all properly fulfilling the brief.
As a result we’ve seen a number of influencer tech solutions emerge which try to simplify the process through automation, enabling micro-influencers to be used at scale. But is automation fundamentally incompatible with influencer marketing, which prides itself on authenticity?
The Dangers of Commoditisation
Influencer marketplaces, the likes of Influencer, TapInfluencer, and Kasitoki, all claim to help take care of this heavy lifting. They allow brands to put out briefs to a network of social media influencers, with the platforms helping select which influencers might be a good fit.
Once the brand has chosen its influencers, depending on the level of automation involved, the platform may then take care of assessing the content put out by the campaigns, making sure influencers are honoring their obligations, and analysing the campaign’s performance.
These platforms often feature dashboards which look very similar to those brands and agencies are used to from their display and video campaigns, and the appeal of cutting out the manual labour of sifting through thousands of potential influencer partners is obvious.
But Maria Cadbury, founder and CEO of Spring, says that in the rush by some brands and agencies to commoditise and scale influencer marketing, the quality of their campaigns has taken a hit.
“You can end up with micro-influencers who are just looking for a quick buck, and they don’t actually understand how to edit their own work, or understand what a creative strategy is,” said Cadbury. Even when the influencers selected are more experienced, their creative can end up missing the brief without sufficient human oversight.
A good example of the danger here cropped up last year with a post put out on Instagram by fitness influencer Andrew Papadopoulos in partnership with Nescafe.
At a glance, this post might appear to have ticked all the boxes. The product appears prominently in the post, the couple featured are both smiling, and there’s nothing offensive in the background. In the caption, the brand name is included, as well as a link to Nescafe’s Instagram account, the branded hashtag, and a reference to the fact that it’s a sponsored post.
But the post was slated on Instagram and Twitter, for the obvious reason that many found Papadopoulos using a sponsored post to announce his engagement to be somewhat tasteless. Cadbury says this is the danger of automation – these sorts of things won’t be picked up by a computer, but can be easily avoided with sufficient human input.
“Influencer marketing needs somebody to be checking the output looks authentic,” she said. “Authenticity is a word some people think is overused, but it is very important. The perfect campaign is where you as a brand are using influencers who are really passionate about the area, whether it’s the brand itself or the area the brand is working in.”
“Leaving a machine to determine whether the creative story and content looks right is where influencer marketing got a really bad rep last year,” said Cadbury.
Adam Williams, CEO of Instagram influencer marketing platform Takumi, agrees about the need for human oversight. “Originally, Takumi was very much about trying to automate the whole process, or at least a big chunk of the process,” he said. “But what we found over time is clients become more demanding and better educated, they wanted to be much more specific and bespoke with what they did. So we had to build much more flexibility into the service, and a lot of that comes from human input.”
Williams said that even when it comes to just sending the brief out to influencers, often a human touch is needed – a phone call might be required to properly convey what the brand is looking for.
A Human Behind the Wheel
Neither Cadbury nor Williams suggest that automation has no role in influencer marketing, and indeed in some areas it is nearly essential. Williams said for example that for influencer campaigns using hundreds of influencers, it would be almost impossible to manage communication without an element of automation.
Automation can be extremely beneficial for the areas of influencer marketing which require a lot of manual work, with no real need for creative oversight. Technology can, for example, track whether influencers working on a campaign are posting using the right phrases, and at the right frequency. When it comes to collecting data on influencers themselves or the performance of their campaigns, tech can take care a lot of the heavy lifting here too.
And in other areas, even when automation does need to be overseen by human beings, it can still be a very useful tool. “What technology can do is flag things up, and then people manage those flags,” said Cadbury.
Williams said that when it comes to identifying influencers to work with a brand, technology can simplify the process significantly. “The automation does the initial number crunching – is this person relevant, and should we even be having a conversation with them,” he said. Then, a human team can step in to look choose which of those influencers flagged up by the tech will be most appropriate.
For spotting fraud, a key concern for advertisers investing in influencer marketing, automation can be very helpful too. But while an algorithm can spot unnatural looking spikes in an influencer’s follower count or engagement, humans still need to be involved. “If you speak to the influencer you might find out they were on TV or had a press release put out or something, which gave them that jump,” said Williams.
For brand safety too, tech can be helpful for filtering through an influencer’s old social media posts to make sure there’s nothing problematic buried in their Instagram feed.
As influencer tech becomes more sophisticated, there may be less human oversight needed in the future, but Williams believes that there will always be a trade off which comes with cutting out human oversight.
“If you want to automate stuff to make it simpler and cheaper, you end up losing the personal element, and you lose an element of quality,” Williams said. “Often it’s a human instinct to recognise a quality image, as opposed to a computer’s attempt to recognise what might work.”