EVOPEDIA

← all systems · see it in The System

Social media

Online services on which users make and share content (text, photos, video) and connect their profiles to each other, carried by the internet through web and mobile apps: the platforms (Facebook, YouTube, WhatsApp, Instagram, TikTok, WeChat, X and many more), the posts and interactions users generate, the social graph of their connections, and the ranking systems that choose what each user sees. Grown from PLATO's notes and chat (1960), bulletin boards and Usenet (1970s-80s), and profile-and-friend-list sites (SixDegrees, 1997) to platforms used by billions; most are paid for by selling advertising slots targeted with users' data. Communication research reads it as a many-to-many (dialogic) medium beside one-to-many mass media; computer science as feeds ranked by machine-learned relevance models; sociology as networked publics whose studied effects on news, politics, mental health and social ties are disputed.

Properties published scienceour proposal

Sourced fields, and our own readings.

aliasesSocial media platforms, sys:social-media-platform, Social media platform
born atPLATO 1960; Community Memory by 1973; CBBS 16 February 1978; Usenet 1980; SixDegrees 1997 (contested which counts)
born precisioncontested / decades
boundaryInside: the platforms (their services, software, ranking and recommender systems, rules, moderation and business), users' profiles, the content users post, the connections between profiles (followers, friends, groups) and the interaction data logged on them. Outside: the people themselves and their offline lives and ties, the internet and the devices that carry the services, advertisers, news organisations, governments and courts. The line is porous: users make the content but the terms of service set who owns it, and the platforms 'control access to the content'.
nested levelA part of the mass media in the communication field's reading (its dialogic, many-to-many form; contested, C-1140); built from user-generated content, profiles, the social graph, ranking systems and platform companies; carried by the internet's services floor and by people's own output (their posts, taps and talk).
scaletechnological

Measures (24) published science

Social media users worldwide4.76 billion (about 59% of the global population) people
Social media users worldwide (Statista, 2022)3.96 billion people
Share of online time on social networks22 percent (2015)
Share of active Facebook users in the giant component99.91 percent
Average distance between Facebook users4.74 (3.74 intermediaries) hops
Stable relationships per Twitter user100-200 relationships
Algorithmic amplification of parties' tweetsover 200% in some cases (3 times the chronological audience); Canada Liberals 43% vs Conservatives 167% percent over chronological reach
Size of Twitter's chronological control group1 percent of all global users (since 2016)
Diffusion increase per moral-emotional word20 percent per word
True and false news cascades studied (Twitter 2006-2017)~126,000 stories, ~3 million people, more than 4.5 million tweets count
Additional votes from one Facebook message (2010)340,000 (60,000 direct, 280,000 through friends) votes
Direct effect of the social message on voting0.39 percentage points
Per-close-friend effect on voting0.224 percentage points per close friend
Variance in adolescent well-being explained by digital technology useat most 0.4 percent
Lag from social media use to life satisfaction in adolescenceone year years
Lifetime of a story24 hours
Experience-sampling rate (Kross)five a day for two weeks text messages
Median daily social media use (sleep study)61 minutes per day
Median age of starting social media (US teens 16-18)14 years
Users who regularly get news from social media (Pew 2021)roughly 70 percent of users
Weight of an emoji reaction against a like (Facebook 2017)5 times
DSA maximum fine6 percent of global sales
Articles in the digital media and democracy review496 (24 causal) articles
Posts analysed for reward learningover one million posts from over 4000 individuals

Open contentions (6) open contention

Where sources or readings disagree. Both stay on the map.

Cross-scale patterns seen here (21) our proposal

Shapes that recur when the same kind of problem is solved at different scales, and how each shows in Social media. Each links to every other system that shows it.

Valence: what it trades, and with whom (19) published science

Each exchange with a neighbour, read from both sides: who gains, who pays, and what carries it.

Partners and parts published science

Built from / builds (11)

Sits inside / holds (3)

expressed by (12)

floor filled by (24)

Kinds (17)

wired (1)

Notes published science

Layers

The ways it takes in its world, where they meet, and how it acts back, each from L1 (where it meets the world) up.

social media as communication research and computer science describe it: people's lives and posts to the ranked feed, the public conversation and its effects on the world (reader's view)

  1. L1: People and their lives, events and devices: users reach the services through web and mobile apps carried by the internet; 4.76 billion people
  2. L2: User-generated content and profiles: text posts, comments, photos, videos, and the data generated by interactions; the profile a user builds
  3. L3: The social graph: friends, followers, groups and lists connecting profiles (721 million users and 69 billion friendship links on Facebook in 2011, 99.91% in one component)
  4. L4: The feed: ranking and recommender systems select and order what each user sees from content features, social connectivity and user activity (reverse-chronological feeds before; machine-learned relevance models from 2016 on Twitter)
  5. L5: The platform company: rules, moderation, terms of service and the business model of selling targeted advertising slots; under law (Section 230, the EU's DSA and DMA, age limits)
  6. L6: Networked publics and collective attention: virality and cascades, trending, homophilic clusters and echo chambers, emotional and moral contagion
  7. L7: Effects back on the world: news consumed and shared, political self-expression, votes and protest, moods and well-being, purchases; beneficial and detrimental by context

the boardroom (shared floors)

  1. L4: The feed: ranking and recommender systems choose and order what each user sees from content features, the graph and engagement; a personalised relevance model or a reverse-chronological list
  2. L5: The platform company: rules and moderation, terms of service, the advertising business, compliance with law, copying rivals
  3. L6: Networked publics: virality, cascades, trends, homophilic clusters, emotional and moral contagion; the competition of memes for limited attention
  4. L7: Effects back on the world: news consumption, participation, votes and protest, trust and polarisation, moods and well-being, purchases

way in: Advertisers' money and targeting requests

  1. L1: Firms promoting products and seeking customers
  2. L2: Advertising slots bought
  3. L3: Matched to data about each user so ads are individually relevant
  4. L4: Placed in feeds, where users can share and comment on the ad

way in: Automated accounts: bots, social bots and cyborgs

  1. L1: Automated programs operating on the internet
  2. L2: Programmed to like, comment and follow like people; some pose as real people
  3. L3: Mixed into the interaction signals so that bot and human interactions are hard to tell apart; they spread true and false news at the same rate

way in: Law and regulation: legislatures, regulators, courts

  1. L1: Legislatures act (the US Communications Decency Act, 1996; the EU's DSA and DMA, 2022; Australia's under-16 law, 2024)
  2. L2: The rule: Section 230 exempts platforms from liability for third-party content
  3. L3: Duties: very large platforms must redesign systems for privacy and minors, end advertising on personal data, and assess recommender risks
  4. L4: Compliance machinery: content moderation and age verification, at large cost

way in: News: outlets' stories, politicians' statements and links shared by users

  1. L1: Events and the news about them; fake news often first, before conventional media publish
  2. L2: Politicians, political organisations and news outlets post to large audiences
  3. L3: Published, consumed and shared along the graph; about 70% of users regularly get news this way

way in: Rival platforms and their features

  1. L1: Other platforms adapting functionality beyond their scope
  2. L2: A rival's feature taken up (Snapchat's stories by Instagram, then YouTube)

way in: The platform reading itself: randomised experiments and control groups

  1. L1: A randomly chosen 1% of all Twitter users kept on the old chronological feed since 2016
  2. L2: Feeds changed for a test group (e.g. emotional content reduced in the News Feed)
  3. L3: Differences between treatment and control read back and used to improve the ranking

way in: What people do with posts: likes, shares, comments, reactions, time spent

  1. L1: Users' responses to posts; the like as a signal of approval
  2. L2: Logged as interaction data; sharing tools (the like option, re-share)
  3. L3: Read as engagement signals that the ranking tracks; reactions can be weighted (emoji reactions five times a like on Facebook in 2017)

way in: What people post: text, photos, video, live streams

  1. L1: People's lives, memories, friendships and communities, and the events around them
  2. L2: Captured and uploaded through web and mobile apps (cameras, camera phones, keyboards)
  3. L3: Attached to the poster's profile and sent along their connections

way in: Who connects to whom: friend requests, follows, groups

  1. L1: People's ties, most of the strongest also face to face
  2. L2: Profiles and friends lists under actual names (SixDegrees, 1997)
  3. L3: The social graph, held by the platform and capped in use by people's own limits (100-200 stable relationships on Twitter)

way out: Likes, comments and counts returned to the poster

  1. L1: Shown to the poster as likes: read in the brain's reward regions
  2. L6: Endorsement accrues on a post

way out: Messages that move offline behaviour (get-out-the-vote, calls to protest)

  1. L1: Votes cast at polling places (about 340,000 more in 2010)
  2. L3: Passed on through close friends more than through the message itself
  3. L7: A platform message to millions, with friends' faces

way out: Moderation: deletion, reduced reach, bans

  1. L1: Content and accounts removed from the shared space
  2. L4: Problematic content deleted or its reach reduced; bots blocked with limited success
  3. L5: Rules for what is allowed and what is promoted

way out: Targeted advertisements

  1. L1: Ads seen, shared and acted on; peer communication raising purchase intentions
  2. L3: Targeted with data about each user
  3. L5: Ad slots sold

way out: The ranked feed shown to each user

  1. L1: Light and sound from the screen to the user's eyes and ears
  2. L2: Rendered in the app, curated to keep the user scrolling
  3. L4: The ranked selection for this user (in some cases an audience three times the chronological one)

way out: Users' data passed on: to advertisers, third parties, governments

  1. L1: Made available to third parties, including for targeted advertising; used by law enforcement
  2. L5: Data collected beyond the profile, often without knowing consent

Sources (26) published science

Social media (Wikipedia, via Wikimedia Enterprise snapshot 2026-11-01, version 1382672990)
Lindstrom B, Bellander M, Schultner DT, Chang A, Tobler PN, Amodio DM (2021) A computational reward learning account of social media engagement. Nature Communications 12:1311. doi:10.1038/s41467-020-19607-x (PMC7910435, CC BY) (behind its links)
Network effect (Wikipedia, via Wikimedia Enterprise) (behind its links)
Brady WJ, Wills JA, Jost JT, Tucker JA, Van Bavel JJ (2017) Emotion shapes the diffusion of moralized content in social networks. PNAS 114:7313-7318, doi:10.1073/pnas.1618923114 (behind its links)
Orben A, Przybylski AK, Blakemore SJ, Kievit RA (2022) Windows of developmental sensitivity to social media. Nature Communications 13:1649. doi:10.1038/s41467-022-29296-3 (PMC8960761, CC BY) (behind its links)
Huszar F, Ktena SI, O'Brien C, Belli L, Schlaikjer A, Hardt M. 2022. Algorithmic amplification of politics on Twitter. PNAS 119(1):e2025334119. doi:10.1073/pnas.2025334119 (PMC8740571, CC BY-NC-ND) (behind its links)
Kramer, Guillory & Hancock 2014, Experimental evidence of massive-scale emotional contagion through social networks, PNAS 111:8788 (doi:10.1073/pnas.1320040111); editorial expression of concern in the same record (behind its links)
Ugander J, Karrer B, Backstrom L, Marlow C (2011) The anatomy of the Facebook social graph. arXiv:1111.4503 (abstract) (behind its links)
Backstrom, Boldi, Rosa, Ugander & Vigna 2012, Four degrees of separation, arXiv:1111.4570 (WebSci 2012) (behind its links)
Goncalves B, Perra N, Vespignani A (2011) Modeling users' activity on Twitter networks: validation of Dunbar's number. PLoS ONE 6(8):e22656. doi:10.1371/journal.pone.0022656 (PMC3149601, CC BY) (behind its links)
Vosoughi S, Roy D, Aral S. 2018. The spread of true and false news online. Science 359(6380):1146-1151. doi:10.1126/science.aap9559 (PMID 29590045) (behind its links)
Bond RM et al. (2012) A 61-million-person experiment in social influence and political mobilization. Nature 489:295-298, doi:10.1038/nature11421 (behind its links)
Orben A, Przybylski AK (2019) The association between adolescent well-being and digital technology use. Nature Human Behaviour 3:173-182. doi:10.1038/s41562-018-0506-1 (abstract via Europe PMC) (behind its links)
Kross E, Verduyn P, Demiralp E, et al. (2013) Facebook use predicts declines in subjective well-being in young adults. PLoS ONE 8(8):e69841. doi:10.1371/journal.pone.0069841 (PMC3743827, CC BY) (behind its links)
Lorenz-Spreen P., Oswald L., Lewandowsky S., Hertwig R. (2023) A systematic review of worldwide causal and correlational evidence on digital media and democracy. Nature Human Behaviour 7:74-101. doi:10.1038/s41562-022-01460-1 (behind its links)
Wikipedia: Social networking service (behind its links)
Sherman LE, Payton AA, Hernandez LM, Greenfield PM, Dapretto M (2016) The power of the like in adolescence: effects of peer influence on neural and behavioral responses to social media. Psychological Science 27(7):1027-1035. doi:10.1177/0956797616645673 (PMC5387999; abstract only) (behind its links)
Weng L, Flammini A, Vespignani A, Menczer F (2012) Competition among memes in a world with limited attention. Scientific Reports 2:335. doi:10.1038/srep00335 (PMC3315179) (behind its links)
Bail CA et al. (2018) Exposure to opposing views on social media can increase political polarization. PNAS 115:9216-9221, doi:10.1073/pnas.1804840115 (behind its links)
Journalism (Wikipedia, via Wikimedia Enterprise snapshot 2026-11-01) (behind its links)
Dickinson GM (2025). Section 230: A Juridical History. Stanford Technology Law Review 28:1. arXiv:2503.05814. (behind its links)
Newman et al. 2025, Reuters Institute Digital News Report 2025, executive summary (48 markets) (behind its links)
Orben A, Przybylski AK, Blakemore SJ, Kievit RA 2022. Windows of developmental sensitivity to social media. Nat Commun 13:1649. doi:10.1038/s41467-022-29296-3 (behind its links)
Sarin A, Murthy P, Jain S. Addiction and technology: Plus ça change plus c'est la même. Indian Journal of Psychiatry (2017), doi:10.4103/psychiatry.IndianJPsychiatry_371_16 (behind its links)
Cinelli M, De Francisci Morales G, Galeazzi A, Quattrociocchi W, Starnini M. 2021. The echo chamber effect on social media. PNAS 118(9):e2023301118. doi:10.1073/pnas.2023301118 (PMC7936330, CC BY) (behind its links)
News (Wikipedia, via Wikimedia Enterprise snapshot 2026-11-01) (behind its links)