Using Facebook Reactions to rank content is akin to counting beans. We don't need it.
It's just a game we play with robots... and ourselves.
For years, social media marketers have been trying to reverse engineer the content ranking algorithm that powers Facebook’s News Feed. At its highest, most broadest level, the majority of the industry believes organic ranking is based on four causata:
The total inventory of organic posts available to display for any given user
Signals that tell Facebook’s learning engine what the post is about and how it is already being interacted with
Predictions on how any given user, based on the user’s aggregated behavioral profile, will (or will not) react to each post
A score assigned to the post based on the factors above, which sit somewhere in the order of tens of thousands of data points
Understanding how ranking works is an exercise in understanding how to improve what content is published, and how content is delivered over time—an exercise that is coveted by social media managers and marketers alike to produce the “best performing” content over time.
We surmised that a Like is not as meaningful as sharing a post or writing a comment because both require more time and effort to execute. When Reactions came onto the scene, some of that changed.
Leaving a Reaction is a bit of a challenge from a UX perspective: a user has to press and hold the Like button, navigate their thumb upward, swipe right and release on the desired emoji. The outcome of the added tap motions suggests using a Reaction over a Like means a stronger interaction signal, and the algorithm is programmed to weigh Love, Haha, Wow, Sad, Angry, and newly-added Care over a Like.1
Sammi Krug, current Director of Product Management at Facebook, in a February 2016 press release, wrote2:
Over time we hope to learn how the different Reactions should be weighted differently by News Feed to do a better job of showing everyone the stories they most want to see.
So when the news broke this month that Reactions are being weighed five-times over Likes3, there was no shock and awe. All this was known to us back in the mid-2010s. What has been boggling us since then was the intentional choice by Facebook to equally treat signals such as Angry and Love. Love and Angry—even though they represent two dichotymous emotional states—are weighed in such a way that they both signal to the algorithm, “Yes, I’d like more of this, please.”
About a year after Facebook’s wider rollout of Reactions, they analyzed the data and found that Love was the most used Reaction with a 41% share, while Angry sat at just 5%.4 If people use Love so much more than Angry, shouldn’t they be seeing more content that is being loved than hated?
The short answer is no.
The reason has to with social comparison. Facebook earlier this year commissioned a study to better understand the effects of social comparison on its platform. They discovered that seeing proportionally more posts that show higher Reaction counts led to negative emotional responses such as inadequacy, sadness, and jealousy—particularly among teens.5 It didn’t quite matter if the Reaction was representative of love or hate.
The question is: does Facebook really need to assign value to Reactions at all? Assigning value to Reactions—including Likes—commodifies them: they have been (and continue to be) a source of digital authority, validity, and currency. Facebook is already factoring hundreds of thousands of data points to rank content in the News Feed. Is their value worth the pain points of Reactions as a social media commodity? What exactly is the value of an emoji in a world where AI can semantically interpret the nature of any given post? Or is it just counting beans?
In my opinion, we could do without them.