Facebook documents show its News Feed algorithm treated emoji reactions as 5x more valuable than Likes starting in 2017, amplifying controversial content
Five years ago, Facebook gave its users five new ways to react to a post in their news feed beyond the iconic “like” thumbs-up: “love,” “haha,” “wow,” “sad” and “angry.”
Washington Post
Context & Ripple Effects
When Facebook rolled out Reactions to everyone in February 2016, it told users and publishers the new buttons carried the same weight as a Like in News Feed ranking "for now" — and the design process behind narrowing the set to six was framed purely as an empathy exercise (Wired's account of the icon choices). The Washington Post's document reporting rewrites that framing: since 2017 the algorithm quietly counted each emoji reaction as five Likes, with "angry" among them.
The significance is that the empathy feature became an amplification mechanism — posts engineered to provoke strong feelings earned five times the distribution per interaction, and the company's public description of equal weighting no longer matched the system's actual behavior.
First-order effects
Publishers and page operators who optimized for Likes saw their reach calculations invalidated overnight: a post drawing "haha" or "angry" responses out-distributed an equally Liked one by roughly 5-to-1, rewarding outrage-bait and divisive content in the Feed.
Second-order effects
Content strategies across the pages ecosystem shift toward provoking any strong reaction — including negative ones — which feeds exactly the inflammatory and disinformation dynamics documented elsewhere on the platform rather than the positive-affinity signals Reactions were sold on.
Third-order effects
The gap between stated ranking behavior and internal weighting strengthens the case for algorithmic-transparency requirements: if engagement signals as innocuous as emoji can be secretly multiplied to reshape what billions see, regulators and researchers will push for auditable ranking inputs.
The trend: Platform engagement features designed as user-expression tools are being revealed as ranking levers, pushing social media toward external scrutiny of how algorithms weigh attention.
To me this is not a story of Facebook intentionally fanning anger for profit. It's a story of how arbitrary initial decisions, set by humans for business reasons, become reified as the status quo, even as evidence mounts that they're fueling harms. https://t.co/CdsciNOpQV https:/…
Facebook secretly weighted reaction emojis, including “angry,” as 5x the value of “likes”—over the integrity team's warnings. We wrote about the obscure, often arbitrary, human decisions that shape Facebook's algorithm and how we all interact online: https://www.washingtonpost.co…
My subjective experience of Facebook's algorithm for the last few years was that it was consistently showing me the Trumpiest content it could find. https://www.washingtonpost.com/ ...
FB's stock is still close to historic highs amid a dump of internal documents, whistleblower testimony and a barrage of global media coverage. Investors are looking at all this, including regulation, and not finding anything that would significantly harm the business https://twit…
Internal experiment with removing algorithmic ranking and showing a mostly chronological newsfeed produced lots of problems. It did apparently make FB more revenue, though. https://bigtechnology.substack.com/ ...
The takeaway should be that humans are perfectly capable of curating noisy chronological content feeds, even with the limited controls they have today. Engagement-maximizing (algorithmic) or noisy (chrono) is a false choice...the actual choice is AI-curated vs. user-curated.
We don't yet have access to the original study, but this take reads like it was written by an FB lobbyist. Removing AI manipulation from News Feed leads to a 50% increase in intentional user curation of feed stories (hides.) That's A Good Thing! https://twitter.com/...
Documents reveal internal fury and dissent over site's policies https://www.nbcnews.com/... // This way more balanced view shows some things: • lots of docs, but clearly carefully chosen • still no focus on role of Media as source for sharing • big companies are complex
Points to @Kantrowitz for: “Though I joined this consortium on Monday, I'm in favor of its dissolution and the speedy, responsible release of all Haugen's documents to the public” https://bigtechnology.substack.com/ ...
The consistent theme in FB docs is that anything related to hate or right wing extremism was given a “First Amendment” allowance and anything remotely progressive was filtered through a “politically neutral” veto. Every single time. #FacebookFiles https://twitter.com/...
Why did Zuckerberg alter Facebook's algorithms in 2018, making toxic content viral? One explanation: Facebook had been hemorrhaging and losing teens and young adults for years in its richest markets. The situation was dire. @BarbaraOrtutay https://apnews.com/...
Without a News Feed algorithm, engagement on Facebook drops significantly, people hide 50% more posts, content from Facebook Groups rises to the top, and — surprisingly — Facebook makes even more money from users scrolling through the News Feed. https://bigtechnology.substack.com…
New: A Facebook experiment in 2018 effectively deleted the News Feed ranking algorithm. It didn't end well. https://bigtechnology.substack.com/ ... via @BigTechnology
Regulators would love to abolish Facebook's newsfeed. Florida & Texas have already passed laws making it illegal, & Congress is evaluating proposals to eliminate it too. But would FB's users benefit from that? https://bigtechnology.substack.com/ ... https://twitter.com/...