Is Influencer Marketing Dead If Instagram Hides Likes?

BY ADAM GROSSMAN

Instagram announced that it will hide likes for “some users in the United States” beginning this week. This expands on previous tests the company has conducted internationally where Instagram makes likes data only available to account owners. In a recent post responding to these tests, Senior Vice President, Digital Strategy at Octagon Meredith Kinsman asks “Will hiding ‘likes’ lead to the end of influencer marketing?” While Block Six Analytics (B6A) agrees with Kinsman that the answer is no, we do not fully agree on why this is the case.

The reason Kinsman (along with other digital strategists) are asking this question in the first place is that determining influencer value from an engagement and reach perspective typically requires examining likes and comments. Removing likes not only makes this much more difficult but also reduces overall engagement. Kinsman points out that “marketers know hiding the number of likes reduces the chances of a user liking a post” and this then makes influencer content less valuable.

In addition, many companies (including B6A) have found that likes have a strong correlation to impressions. Companies create formulas to determine reach by scaling (or weighting) the number of likes since impressions are not publicly visible in posts. Eliminating likes as a public metric means that digital strategists would need a different approach to determine impressions.

Kinsman primary suggestion for addressing these issues is centered around authentication. Influencers can share their metrics (including likes) with “with [approved] third-party platforms to allow brands and marketers to view them as well.” The challenge with relying on authentication as the primary source for engagement and reach metrics is that it does not reflect the full impact influencer content. More specifically, the benefit of partnering  influencers is to see the impact they have on a larger population beyond their owned accounts (i.e. earned media impact).

The solution comes from looking at the publicly available data that will still be available even if likes are hidden in addition to influencer account authentication. This starts with looking at comments on owned posts. More specifically, the issue with likes (whether they are public or private) has been that it is a binary calculation meaning someone either liked a post or does not like a post.

Sentiment does not have to be a binary calculation. For example, our Social Sentiment Analysis Platform (SAP) uses natural language processing (NLP) to determine posts on a +100% (most positive) to -100% (most negative scale). For Instagram, SAP can evaluate individual comments on an influencer post to determine the content’s exact positive or negative score (i.e. examine public content beyond an owned post). Our research has shown that the specific way SAP calculates lifts in sentiment has a strong and statistically significantly correlations with lifts in revenue. Therefore, SAP enables our clients to determine the revenue impact of brand perception along with the number of overall comments.    

In addition to comments, the text and visual content of social media posts will still be publicly available. This is critical to determining influencer value because it enables marketers to examine content outside of owned accounts. Traditionally, digital strategists have used specific hashtags and keywords to find earned media posts and content. However, our work with clients has shown that company logos are often present in posts that do not use a specific keyword or hashtag. In addition, these posts frequently generate more value for brands than ones with a featured keyword and hashtag.

Our Media Analysis Platform (MAP) provides our clients with the capability to complete earned image analysis in Instagram (in addition to Facebook and Twitter). We work with our clients to determine the accounts that likely will generate the most earned media value in addition to an influencer’s owned accounts. We then use MAP to evaluate the video and image content in earned posts for logo activations. This enables us to better understand the full reach a post generates by looking at the number of posts with logo activations.    

While Kinsman analysis focuses primarily on engagement and reach metrics, she does not directly touch on the third component critical in determining an influencer’s value. Fit should be a critical consideration for brands determining if a brand should work with an influencer. In this context, fit typically means how does an influencer’s audience align with a company’s targeted demographics.  

Account authentication can provide demographic and psychographic information to marketers for owned accounts where influencers provide access. However, what if brands want to consider a large number of influencers at the same time? That would mean that a marketer would need to contact every single influencer to obtain account authentication. If the influencer does not provide authentication than how does a marketer determine fit?

Once again publicly available information can provide a solution. B6A’s Audience Inference Platform (AIP) uses NLP and our proprietary algorithms to determine the demographic and psychographic profiles of virtually any account with 900+ followers. Essentially, what people say in public posts and who they are connected to has a strong relationship with who they are (demographics) and what they find to be important (psychographics). Our clients can then use AIP, SAP, and MAP to determine the comprehensive influencer value from publicly available data.  

To clarify, B6A would prefer that Instagram not remove likes from posts for reasons that Kinsman articulates. However, it is clear that the social media companies are investigating how to “depressurize” the user experience and eliminating likes is one potential solution. Influencers will still have significant value even with these changes to likes and publicly available data will be critical to determining this value at scale.