Inside Facebook’s Decision to Blow Up the Like Button
bloomberg.com
bloomberg.com
Facebook researchers started the project by compiling the most frequent responses people had to posts: “haha,” “LOL,” and “omg so funny” all went in the laughter category, for instance. Emojis with eyes that transformed into hearts, GIF animations with hearts beating out of chests, and “luv u” went in the love category. Then they boiled those categories into six common responses, which Facebook calls Reactions: angry, sad, wow, haha, yay, and love.
Facebook Reactions won’t get rid of like—it will be an extension. Within the company, there was some debate on how to add the options without making every post look crowded with things to click. The simpler Facebook is to use, the more people will use it. Zuckerberg had a solution: Just display the usual thumbs-up button under each post, but if someone on her smartphone presses down on it a little longer, the other options will reveal themselves. Cox’s team went with that and added animation to clarify their meaning, making the yellow emojis bounce and change expression. The angry one turns red, looking downward in rage, for example. Once people click their responses, the posts in News Feed show a tally of how many wows, hahas, and loves each generated.
edit: Yay was rejected as a choice.
Personally I think a 'helpful' or 'useful' reaction is badly needed also, but apparently it's not a common enough reaction to show up on Facebook's radar.
FB 6 basic emotions - angry, sad, wow, haha, yay, and love
Humintell's 7 basic emotions - anger, sadness, surprise (aka wow), joy (aka haha/yay?), fear, disgust, contempt. those last 3 aren't buttons on FB for obvious reasons :)
Obviously, given the near six billion dollars they made, they know what they're doing. But....
This doesn't fix anything currently making Facebook suck. It makes it noisier.
My solution would be a tagging system. People can add tags (that don't even need to be shown by default), and other friends' feeds can be filtered by the tags.
For instance: This is a baby pic? Tag it as baby. In my feed, I will never need to see another baby again.
Other people don't like politics or football. They can get tagged as such, and ignored by those who don't like those topics.
This would absolutely positively fix Facebook for me. They are smart enough to make it work. No more babies PLEASE!!!
I imagine it's on their radar considering all their interest in AI.
It'd be a messy solution for Facebook but an awesome and nonintrusive solution for annoyed users once it's working.
#baby #myLittleGirl #littleHands #myWorld #parenting #healthyLiving #mom #momKnowsBest #sanFrancisco #sanFran #USA #earth #apple #grannySmith #farmFresh #fuckGluten
Looking at it I was wondering what are the chances that it would lead to posting of an unintended reaction ("Haha" instead of "Love"). Once a reaction is tapped on, it is gone to the wild forever. There won't be any undo.
With 'Like' button there is no chance of accident as there is nothing else. And considering the reaction feature is being built for mobile - small screen, distracted attention, unsteady hands, the chances of 'oops' scenario are quite high.
Talk about unintuitive UI. I don't understand this recent fad of pressing harder/longer on a button to discover a secret hidden feature that's unavailable otherwise.
Also I liked yay the best, too bad they axed it.
On the other hand these were needed let's say because some items are not appropriate to Like. Like say a piece of news for some tragedy.
But I do think this will backfire as I think it will lower engagement. Right now when people cannot Like, they Share. And for other reactions they place comments. Many in the form of emoticons but still.
For more and more users touch is going to be their primary experience and eventually you are going to need some method to provide more options to the "power users". Longer-click seems the most obvious and common of the patterns.
Most people won't even realise that a web-app can have a right click context menu (even those that use right click in their desktop programs all the time).
As for desktop, if they're intelligent at all, they'll use a different method there. Perhaps all the new variations will be displayed side by side. Perhaps instead of only trigfering on a "long-click", it'll expand on a normal click.
Desktop applications are a different story.
Overall I agree with the parent comment. Right click and long-press are hard for non-power users to find/discover.
No kidding, as a 28 year old I'm embarrassed to admit how long it took me to "get" snapchat's UI.
I believe the UI has been tweaked for users that are more impulsive. I know it sounds strange, but you open the app the first time and you are immediately asked to make a picture of something and then send it to someone. (even though you haven't seen your available contacts yet)
- A user that has friends on snapchat. For them, just sign up and snap a picture. You're off to the races. IMO, there's no reason to pick who I want to send it to until I take the picture.
- Another type of user would be someone who does not have any friends using SnapChat. For those people - you don't want to just say, "Sorry, you don't have any people to talk to - but feel free to play around with the camera." That probably won't encourage them to do much more with the app. Instead, you get them to create something first. When it's time to send their creation, they don't have anyone to send it to, but they are way more likely to say, "Hey <other person in the room>, download SnapChat so I can send you this hilarious thing I just made." Both of those people are way more likely to download and use the app - but the chance of them doing so goes way down if they never invest any time and creativity into the app.
Just my two cents, but I think it makes perfect sense.
Snapchat's confusing interface complements the way their users interact. Users tend to learn about features from social interactions with their friends (face to face, or digital), which builds a sense of ingroup. Additionally, the confusing interface tends to filter out older users, who lack the social guidance of the younger users. This keeps the userbase heavily skewed young, which further reinforces its branding as a fun app for the cool kids. After all, nothing is less cool than your mom doing the things you like.
I'm not sure that this is what the company wants (they've had screenshots of a redesigned UI leak this week), but it makes logical sense to me.
Just... why???
But over the time more and more high profile apps removed these features simply because so many users never knew that they could long press on something to reveal additional functionality.
Also the Android old menu button was removed for the same reasons. It was a hidden menu that could show different options depending on which screen within the app you were.
I could imagine that it could be different for users with Apples force touch or 3D touch, but we'll have to wait and see.
My estimate is that the average person knows about maybe a third or a half of the features on their smartphone, because the rest are hidden away behind never-explained gestures or long presses on things that don't look like buttons.
In terms of conveying functionality the popular smartphone and webpage UIs today are atrocious. Just today I was talking with someone who was really frustrated that they can't see webpages anymore because web browsers all got rid of the magnification feature. He was amazed to discover that it has been there all along but hidden away in a menu that doesn't even appear unless you hit alt. He thought all of that was moved over to the "hamburger" menu, and they just dropped it. He's been squinting for months (years?) because someone doesn't like the look of dropdown menus.
Gnome (ongoing dumbification since gnome), Canonical (Unity, although HUD was a nice idea), Microsoft (ribbon, although it has some decent ideas) etc etc.
There is room for huge improvements in UI: more search-friendlyness, a way to explain scary buttons in mobile UIs since they don't naturally seem to support hover etc etc
Dumbifying things, both by assuming your users are useless and cannot learn new features, as well as just plain weird ideas like breaking alt-tab without checking (Unity) or hiding save button in Office behind what looks like a logo (MS ribbon early edition) should hopefully be less important.
But I generally agree with your sentiment. You can't rely on users to poke around and find features on their own. I recently talked to a teenager who despite having used Windows all his life said he only "just" learned about Alt-Tab.
I'd prefer a halo ring around the button that appears when you touch or mousedown that presents additional options related to the button. If you touch, move your pointer, then release over a secondary function, it activates, otherwise a simple touch and release activates the main function.
Similar functionality is available for gamepad devices, where a choice wheel appears, and a directional input can select any of the 1 to 8 available options.
I also think that perhaps there needs to be someone willing to take a few arrows in the chest before new UI concepts become so common as to be generally accepted as 'intuitive'.
On the contrary, this is non-intrusive and discoverable. The user isn't presented with a dense interface and too many choices, but if he does want to do more advanced interactions he will find a way to learn about it.
There are numerous mentions of failed product roll-outs(News Feed as a failure at first, Paper, etc.) But the key quote is here, when they talk about the meeting: "Cox agrees. 'My intuition, which we could prove wrong, is people just want more stuff.'
I'm sure that Facebook uses data to back up some of its assertions; after all, they did mine comments to come up with the most common reactions. But how many people's entire internet experience (i.e. consumption of news through the News Feed) is impacted just by some senior guy in a room going, "This sounds right to me," and everyone scrambling to work on what he proposes?
Data has its place, but you can only user-test things after they're already built, and you can't build everything. At some point, somebody has to make a gut decision on what's worth sending to user testing.
Like senior exec of IT buying into enterprise solutions that sound good on paper and make everyone's life living hell. "Business people will be able to make programming changes without developers using visio diagram!"
Like that'll ever happen. Senior execs of IT are so stupid, amiright?
Everyone knows only programmers should be doing financial calculations with a computer, and that UIs are for losers. /s
On the other hand, ion they use this too heavily, then the newsfeed becomes a serious echo chamber, where it is not only correlated to your friend group, but also to your "optimal" emotional reaction. Might be great for Facebook to sell more ads, but not necessarily great for the end user.
You would think that after seeing how well received the Slack Reactji are, Facebook would have expanded its available options.
I can understand not venturing into custom emoji, but at least let there be a :pile-of-poo: reaction! After all, it's an election year!
Facebook is absolutely correct in keeping a very tight rein on this stuff, and in making the options uniformly "cute". If emotions start running too high, people will start abandoning the entire platform, I think.
The facebook system allows a few critical things:
- a simple tally of reactions. With an open-ended set of emoji, it's a bit silly to tally emoji responses. With 6 options, the tallies are interesting.
- advertising. These six actions are a known quantity and can be baked into their advertising and display algorithms at every level. Using emoji, you'd have to build some kind of emoji-emotion-algorithm which classifies the use of all emoji in real time to a small set of factors for use in the algorithm
But yeah nothing is stopping you from filling up facebook messages and comments with :poo: emoji and stickers and gifs.
Slack's implementation is worth using to question whether of your speculated problems would actually be significant. It updates the tallies in real-time and requires only a single click to increment the count. Although people could use any emoji, in practice there are only a small percentage which are relevant and people tend to add to previous ones rather than add a random new one.
> - advertising. These six actions are a known quantity and can be baked into their advertising and display algorithms at every level. Using emoji, you'd have to build some kind of emoji-emotion-algorithm which classifies the use of all emoji in real time to a small set of factors for use in the algorithm
If an advertiser was trying to do this, they'd have to do all of that work either way – the only difference is the size of the lookup table. I think they'd be more worried about trying to gauge the target of the reaction – e.g. if I post something about politics and you use the angry reaction, it's much harder for a machine to tell whether you disagreed with me or were directing some righteous indignation at the subject of the story and that's the most valuable bit of info to know for deciding which candidate or party's ad to show.
Facebook's five choices are clear and universal human emotions that leave very little room for subjectivity.
Emoji, on the other hand, are used as a famously arcane new language where each symbol has many potential uses, as a pictogram in conjunction with others, or as multiple uses per emoji, as viral "inside" jokes meaning completely different things than depicted (tree, fire, or pineapple are often used to mean a drug, but how do you know when a tree symbol means a drug and when it means a pine tree?) "X_X face" is often used to mean sexytimes, but the X_X emoji is also used to mean exasperation.
Sad, happy, angry, these are simple to work with.
But how do you classify Back-Button? What does a user think of a post when they rate it "Alien-Face". When they use "Skull/Crossbone" are they saying danger/poison, or death, or Halloween, costumes, a movie?
I'm not sure you could ever get to an adequate level of programmatically understanding the intent behind emoji, but I think that it's not too difficult with 5 universal emotions.
Besides, if Facebook only designed things for advertisers every post would have a mandatory form collecting demographic info. They know they have to make using the service enjoyable and this would be an easy way to boost interaction which doesn't make advertising fundamentally harder.
With a smaller set of emoji, you can analyze what text often is coupled with what emoji to help infer the meaning of text automatically. Great if you want to serve ads based on private messages, less of a concern if you just want to deliver a useful product.
Good to have actual reactions anyways. "Someone announces her divorce on the site, and friends grit their teeth and 'like' it." Such cases are everybody's life events and there were no reactions to share unless you comment.
D|A appears to have axed this system though, for whatever reason. I hope it works out better for Facebook.
It has everything to do with Facebook tracking users on 3rd party web pages. This way they know where you have been, for how long and how often. IMO this data is much more relevant to knowing users interests that enables FB to sell targeted ads.
Same goes for Google Analytics: that's the reason it is free - because you are the product.
Mind blown. What a great idea to ban meetings on Wednesdays!
Replace the Like with Reactions. There are 6 reactions you can choose from: angry, sad, wow, haha, yay, and love.
That sort of jives with this anecdote as well:
"When the company redesigned the News Feed in 2013, it looked great on the iMacs in Facebook’s headquarters but made the product harder to use everywhere else."
Most computers aren't macs, let alone iMacs. How could you possibly miss out on testing for even one other thing? It's weirdly close-minded given what facebook is and how it presents itself.
"Cox says Reactions’ biggest test so far was during the November terrorist attacks in Paris. Users in the test countries had options other than like, and they used them. “It just felt different to use Facebook that day,” he says."
Does this disturb anyone else?
Knowing that there was a national tragedy happening, Facebook takes advantage of all that great data coming from their french users.
it's disgusting. and the fact that I'm in a minority on HN that feels this way is pretty disgusting.