Stars, Likes, and Angry Reacts
Remember several years ago, when people used to use stars to rate things? It still exists in certain places, like Yelp and Medicaid I guess. It used to exist on YouTube until about 2010, when they bent the knee as well. It's worth understanding why we made the switch – not only because of the way that likes pervade our world, but also because it seems to be a primitive regression at face value. Going from five options to just two feels like a step back; after all, doesn't a rating of 3.2 tell me more than 150,000 likes?
Foremost, there's the obvious consideration that most people only bother to do something if they feel strongly. It's a reason why things like voluntary surveys and polls are a fundamentally flawed metric – the middle ground is underrepresented. When's the last time you took time out of your day to rate a restaurant three stars? Or to call in to your local news station about their exceptionally average coverage of a story? So in a sense, it could be argued that likes and dislikes cover the extremes, which encapsulates most of the results anyway. Absent a rubric or some kind of formality by which to channel those intermediary sentiments, there's an overwhelmingly arbitrary nature to the way that 2, 3, and 4 stars are awarded. It's infuriating when I have to give feedback for something like customer service, and I'm given a choice from 1 (dissatisfied) to 9 (extremely satisfied). Does a 7 mean that the associate solved all my problems, but farted in the middle of the phone call? Does a 3 imply that they were maddeningly unhelpful but wore a charming smile? Literally as I was writing this piece, I got a call from UC Berkeley's alumni outreach, during which I was asked to rate my time at Cal on a scale from 1-10; I considered it for about seven seconds before I realized that the silence was getting a bit awkward. The rules weren't really defined, and no star-related site bothers to do so – am I rating a news video one star because the news presented pisses me off, or because the video was poorly made?
A subtle psychological consideration for this inclination toward the extreme is the power of an individual vote. When you rate or review something, you are passing judgment on it as you see fit, usually without an external influence. Now if you were to vote a 2, 3, or 4, it doesn't hold the same weight as a vote of 1 or 5 would. Statistically, those extreme values have a higher effect on the overall rating, and as a judgment-passer, you want to make your metaphorical voice heard. To that end, it's unproductive to give a middling vote.
On the flip side, there's a considerable amount of nuance removed when making this kind of a switch. It's like removing the overtones from a note played by a musical instrument, or chewing something to taste it but spitting it into a bucket instead of swallowing – you get the general idea, but you're not capitalizing on it as much as you can. Something isn't always 'good' or 'bad'. Netflix is a clear example here: they inexplicably converted to likes/dislikes some months ago. Before, a three-star rating was equivalent to "I guess you can let me know if there's a new season and I'll see if I have time," but now I'm forced to downvote it into the abyss.
And there are niches that successfully employ some manner of informed rating. Feedback for eBay sellers, for example, has four different categories (item as described, seller communication, shipping speed, and shipping cost) of five-star ratings, compounded by an 'overall' category with three options (positive, neutral, and negative). Fortunately eBay doesn't make an attempt to condense these parameters into a single statistic, instead showing each category separately. The Google Play store breaks its overall 5-star rating down into subcategories for certain apps, like having Gameplay/Controls/Graphics ratings for games. But both of these examples are for product-driven websites, where every interaction (a purchase) comes with a respectable amount of forethought. In a more social or interactive framework, there's a greater drive to fire off interactions at will, and mentally parsing through that amount of data is counterproductive to the aim of getting through as much content as you can.
Quick tangent: At least for videos, there's some measure of advanced statistics available. Things like playback patterns, drop-off times, and skips are far more telling of a video – considering how they lack subjectivity. But for everything else, you don't have that luxury. Seeing that an Instagram photo has 100k likes is only marginally impressive; seeing the like count juxtaposed with things like the follower count or view count would add some perspective. But of course, this is all adding context into a system explicitly designed to kick context to the curb – imagine if you could rate an Instagram photo with 4 stars.
To make ratings into a real indicator of success, there has to be a recognition that ratings and rankings are distinct. And there's space for a Facebook-esque 'like' button here. You explicitly like it if you like it, and you implicitly don't like it if you don't hit the like button. Each post is then 'ranked' by the number of likes divided by the number of views, which not only evens out the gap between different creators' audiences, but also gives a way to see more overall content. A old-world 'who has the most likes' system isn't likely to give way to new content; a photo that has 4 million likes probably won't be dethroned easily. But the more views that photo gets without any additional likes would drop its ranking, meaning that people who might have seen that photo instead see something else. You can even use the 5-star approach here, but the 5-star output has to consist of more than just the 5-star input.
There are several reasons that this system doesn't include a 'dislike' button, the most important of which is the connotation. Disliking something is at its core a negative action, whereas granting even one star has some association with generosity. The former is a -1 out of 0, and the latter is a +1 out of 5.
Regardless of the relative ineffectiveness of the 'like' rating system, Facebook held steadfast in their choice to not add a 'dislike' button, until they partially caved with the inclusion of an 'Angry' react over a year ago (a notable departure from their clutter-free ethos that learned from the lessons of MySpace). The motivation behind this is likewise interesting: at its core, the like button is supposed to range from a gentle expression of support to a filter for quality content. Reactions moved the goalposts of user feedback from quality to actual curation, while still maintaining its generally positive vibe. This differed from Buzzfeed's longstanding reactions (including <3, </3, LOL, Cute, Win, Fail, OMG, WTF, most of which pretty much mean the same thing, or at least have substantial overlaps) in that the Facebook reactions actually have a possibility of relating to the content. If you LOL at something or think it's a Win, chances are that you probably want to <3 it. But being Sad about something and Wow'ing at it feel more genuine, because you would include those in your list of core emotions – even the most woke among us wouldn't confess to having 'yaas' as a baseline for feeling.
As with any feature, the unintended consequences create problems far beyond the ones that the feature was originally meant to solve. The Facebook 'Like' was meant to purge the system of people commenting 'I like this', it also dumbed down the interaction. There's some redundancy to having a series of comments all professing props for a post, but something humanizing about it as well. A like became a substitute for all that riffraff, and more importantly became an endorsement. Liking something was implicitly expressing a desire for more people to see that thing, a greater responsibility than a regular supportive Facebook user anticipates or asks for. Reactions' unintended consequence was that of confirmation bias. It's too easy to see whether I'm 'supposed' to enjoy something or be mad at it, considering how the top three reactions are all there with numbers for each. Mob psychology means that I don't want to risk any of my social capital by doing something unconventional, such as love'ing a post that has majority angry reacts, or liking a picture that has no reactions at all, or even interacting with something that's more than a few days old.
Reactions collectively pose a curious question when it comes to changing your mind. Does initially being 'angry' at something and then later changing it to 'wow' count doubly? Or is it simply the equivalent of wow'ing it in the first place? They also offer a somewhat shallower approach to making up your mind in the first place – let's say I posted a link to an article. My commentary on the article (the text I wrote for my post) and the subsequent comments are prominently featured, but the substance (the actual article) requires me to temporarily exit Facebook, which isn't really in my best interests. Before, likes were a measure of intrigue in the subject – something with a lot of likes would indicate that this was something worth making that external trek for. Now with reaction data coupled with minor comments/commentary and a headline, I'm less inclined to even read the link or watch the video or whatever.
Interestingly, the angry react plays a unique role as the only overtly negative reaction (the 'sad' react can be interpreted as a supportive expression of grief or disappointment more so than a bawling-your-eyes-out sentiment that the little animation seems to imply). It contradicts the 'more is always better' phenomenon that has been drilled into us since birth. In the heavily-studied topics of brain hacking and like addiction, one common thread is that our reward centers flash every time we get something. An angry react threatens to upset that norm, considering how they contribute to the overall reaction count, but are largely undesirable. This dichotomy is passable from a marketing standpoint (say if you're a company trying to get any kind of interaction on your content) but confusing on an individual level. In certain situations, this could even dis-incentivize me to post something at all, in fear of the angry wrath. In that sense, the reactions have run counter to their initial mission of curation; instead of curation by the consumer, that curation has shifted to the creator. This is because the notification bubble – the first place our eyes go when we open Facebook – has ceased to be a positive sure thing.
Opening Facebook used to be like playing the slots, now it's more like Russian Roulette with a nerf gun.