As influencer marketing grows in popularity, the industry that has sprung up around it has been trying to get to grips with how to measure campaigns more effectively. Advertisers no longer need to just throw money at a YouTuber or an Instagram star and then pray it has an impact — now there are plenty of metrics that brands can use track to both measure the impact of a campaign and inform pricing.
With the space evolving, and around 62 percent of marketers growing their influencer marketing budgets, VAN spoke to those working in the space to clarify what is and isn’t currently possible, and to gauge the direction the industry is heading.
Towards More Standardised Pricing
In its early days of influencer marketing, the rarity of influencer campaigns meant that prices tended to be negotiated on a case by case basis, as Sheetal Sahota, a senior manager at Rakuten Marketing, explained to VAN, “As part of our affiliate business, we’ve always worked to connect advertisers to content creators. In the beginning, for example, we might have just asked how much we’d have to pay for a certain level of exposure.” In some instances, this method is still used by some of the larger influencers who are in a strong enough position to be able to set their own prices.
Maria Cadbury, managing director at publisher Evolve Media’s influencer marketing department The Studio, gave an example of one in-demand influencer she worked with who wanted to cut down the number of campaigns she worked on to maintain her authenticity. This influencer was able to quadruple her rates, knowing that limiting the number campaigns she worked on would make any she did work on more valuable.
But as influencer marketing has grown in scale and brands are working for with much smaller ‘micro-influencers’, negotiating prices with every individual influencer has become impractical, and so new, more standardised pricing models have emerged.
The social platform a campaign is delivered on might partly dictate the pricing model. YouTube for example tells you how many views a video has received, meaning YouTube campaigns can be measured on a fairly standard cost-per-thousand impressions (CPM) basis. On Instagram however this isn’t possible, meaning campaigns tend to be traded on a cost-per-thousand followers basis instead.
Some in the industry think these pricing models need to evolve further. Dafydd Woodward, global product lead of content and influencer marketing at GroupM, says the cost-per-thousand followers metric is too “broad and generic”.
“At the moment we’re asking is there a more genuine metric that we can trade on, for example cost per engagement, whether that’s a like, a comment or a view,” said Woodward. This model again would be very platform dependent, as different platforms offer a variety of forms of engagement.
Rakuten Marketing’s Sahota said she “wouldn’t be surprised if that’s the way the industry goes”, as there are already significant moves towards cost-per-engagement across the industry. For example Rakuten Marketing, which in many cases uses affiliate links to lead users from an influencer’s content to a brand’s domain, can already reward influencers for being the first click on a customer’s purchase journey.
When it comes to key performance indicators, again the social media platforms themselves dictate in part what is possible.
Total views is perhaps the most basic metric measured, and pretty much all of the most popular platforms will show creators the reach of a given post. More sophisticated insight though, for example into how many users watched the entirety of a video post, will vary between platforms. Snapchat, for example, only showed creators their story view counts until recently, but now offers more in-depth analytics, including things like average time unique viewers spent watching and completion rates, as well as audience demographics. Instagram meanwhile currently gives analytics for regular photo and video posts, but not for Stories.
Engagement too, whilst still not commonly used as a currency, is commonly measured as an indicator of campaign performance. Likes, comments, retweets, and any other forms of engagement all point to a deeper level of interest in a piece of content of course.
Sahota says brands have to be careful with how they interpret engagement for a couple of reasons. Firstly, larger influencers expect to see lower levels of engagement (in relation to their follower count), so advertisers shouldn’t necessarily think campaigns with popular influencers that receive lower engagement rates have been less successful if they are looking at reach or awareness as their main KPI. It’s more likely that people are less inclined to comment on content when hundreds of comments have already been made.
Further to this, it’s important to understand the relative value of each type of engagement on each platform. Blog posts for example should expect to see far lower numbers of comments than Instagram posts, so brand should value 100 comments on an Instagram post and 100 comments on a blog post differently.
Some also dive deeper into comments and hashtags related to piece of content to analyse their sentiment. Sahota says Rakuten Marketing uses AI to analyse the overall sentiment of comments, helping brands understand whether the post has had an overall positive, negative, or neutral effect on the brand’s image.
“That’s a great indicator of whether the audience is liking and commenting because they like they creator, or because they like the content,” she said.
If the campaign itself involves driving users to a specific destination, for example an online store or an app download, clicks through these links can be accurately measured. This is a more platform-independent metric, as advertisers can measure the amount of traffic coming from content posted across a range of social media sites.
Harry Hugo, co-founder and chief campaigns officer at The Goat Agency said that his company can track how many clicks a piece of content on any platform gets, how these clicks are converted to downloads, and how those downloads are converted into sales. Rakuten Marketing meanwhile uses affiliate links to track directly when audiences move from a social media post the the brand’s website, although this still isn’t completely platform-independent. Instagram for example doesn’t allow affiliate links within posts, making affiliate links more awkward to use as they have to be posted in the influencer’s bio, or in the ‘swipe-up’ of an Instagram story instead.
The Inevitable Fraud
Of course several of these metrics are liable to being tampered with by wannabe influencers who pay for followers, bot-driven views, likes and comments in order to dupe brands into working with them. The issue was raised at this year’s Cannes Lions festival when Unilever CMO Keith Weed pledged not to work with influencers who buy followers.
Weed’s concern reflects the worries still held by many advertisers, and reignited the debate around how prone to fraud influencer campaigns are. Some are calling for the social media platforms themselves to do more to help identify anyone artificially inflating their follower or engagement counts.
GroupM’s Woodward says some platforms are fairly active in their efforts. “Instagram every now and again will go through a purge where they identify fake accounts and remove them,” he said. “But I don’t think the social media companies see it as their responsibility to ensure that accounts on their platforms are their genuine, or at least it’s not their top priority.”
Often this is just due to technical challenges of creating a solution that addresses all the concerns and needs of different brands and agencies, said Woodward, and the data advertisers get from these platforms is already by and large sufficient to weed out fraudulent influencers.
Sahota said that one obvious clue is if an influencer’s audience is largely from a different country and has a different primary language to the influencer themselves. Other data on the ratio of engagements to comments, the level of their audience’s activity on social media, and level of variation in audience variation across platforms provide further indicators.
That’s not to say though that platforms couldn’t be doing more to help. Woodward said that Instagram opening up its API to advertisers for example would make it much easier to spot when followers have been paid for.
However those that VAN spoke to generally agreed that if advertisers are serious about avoiding working with fraudulent influencers, they should already be able to do so with the data available today.