Google+ was Google’s company-wide bet to beat Facebook, and on April 2, 2019 Google switched the consumer version off. Launched on June 28, 2011 as an invitation-only “social layer” meant to unify everything Google did, it offered Circles for sorting contacts, Hangouts for video, and a clean stream that early reviewers genuinely admired. It was technically capable and, for a moment, fashionable. What it never became was a place people wanted to be. Eight years later it died not in a blaze of competition lost to Facebook, but quietly, hastened by a security disclosure that gave Google a reason to do what its own engagement data already suggested it should.
The numbers tell the story Google spent years trying not to tell. The company liked to cite 540 million “monthly active users,” but that figure counted anyone signed in to a Google account who touched a socially-enhanced property — a +1 button, a YouTube comment, a Gmail profile photo — whether or not they ever visited Google+ itself. The narrower, more honest “in-stream” figure was roughly 300 million, and even that flattered reality. By the time Google announced the shutdown, its own audit found that 90 percent of Google+ user sessions lasted less than five seconds. The New York Times had already called it a ghost town in 2014. The accounts were real; the activity was not.
The proximate cause of death was Project Strobe, an internal review of third-party data access that surfaced a Google+ People API bug exposing the profile fields of up to 500,000 accounts to as many as 438 apps. Google had found and patched it in March 2018 but chose not to disclose it — a decision that, when the Wall Street Journal reported it in October 2018, looked less like prudence than like the Cambridge Analytica-era coverup it was internally compared to. Google announced the consumer sunset that day, October 8, 2018. A second bug disclosed that December, affecting roughly 52.5 million users, moved the date up by four months.
What users lost was modest, because so few were truly using it — but the manner of its life and death left a larger mark. For most people, Google+ was less a service they joined than one they were enrolled in, wired into Gmail, YouTube, and the real-name identity Google wanted to impose across its products. A stripped-down version survived for business customers as Google Currents until that, too, was wound down in 2023. The consumer social network that Google forced on a billion accounts could not, in the end, persuade them to stay five seconds.
Vine was the app that proved six seconds was enough, and on January 17, 2017 the company that owned it switched the network off. Launched on iOS on January 24, 2013, Vine let anyone shoot a looping six-second clip and share it to a public feed; the constraint was the genius. Out of that tiny window came a new comedic grammar — the jump cut, the running gag, the punchline that detonated and immediately repeated — and a generation of native stars who learned to perform in the length of a held breath. Twitter, which had bought the startup for a reported $30 million before it ever launched, owned all of it. Four years later Twitter discontinued the app, having never built a way for Vine or its creators to make money.
The scale was real and the company liked to quote it: by December 2015 Vine claimed more than 200 million active users, and at its 2013 peak it had been the most-downloaded free app in the iOS App Store. It minted a roster of talent that would go on to define the next decade of online video — Liza Koshy, David Dobrik, the Paul brothers, Shawn Mendes — all of whom got their start performing for a six-second loop. What Vine never built around them was a business. There was no creator fund, no revenue share, no advertising product to speak of. The views were enormous and the checks were nonexistent.
So the talent left, because talent goes where it gets paid. Through the first half of 2016, more than half of Vine’s biggest accounts — those with over 15,000 followers — stopped posting or deleted themselves outright, decamping for YouTube, Instagram, and Snapchat, which had figured out the monetization Vine had not. On October 27, 2016, hours after announcing it was cutting nine percent of its workforce, a cost-cutting Twitter said it would discontinue the Vine mobile app. The shutdown came that January; the app became a stripped-down “Vine Camera,” and the archive of every Vine ever posted lingered online until it, too, was taken down in April 2018.
What its users lost was a peculiar and irreplaceable corner of internet culture, and a public record of how an entire art form had been invented in real time. Vine’s death is the cleanest case in the catalog of a thriving network killed not by a rival or a scandal but by the simple, structural failure of its owner to ever answer the question of how it would pay for itself.
Friendster was the first social network to reach a mass audience, and on June 14, 2015 the company that owned it shut the site down. Founded in 2002 by the Canadian programmer Jonathan Abrams and live to the public in 2003, Friendster introduced millions of people to the idea that you could build a profile, connect to your real friends, and traverse the chain of acquaintances that linked you to strangers. It grew explosively — three million users in its first few months — and was, for roughly a year, the most exciting thing on the consumer internet. It was also the network that taught the entire industry, by negative example, almost every lesson about how to run one.
Its peak figure was large and is best stated as a claim: by 2008 Friendster reported more than 115 million registered users, with its real strength concentrated in Asia, particularly the Philippines and the rest of Southeast Asia. But registration is not the same as life, and Friendster’s life had drained away years earlier in its home market. As early as April 2004 it had been overtaken in page views by MySpace, and then comprehensively buried by Facebook. The site was famously, chronically slow — pages that took agonizing seconds to load while a small team struggled to scale a service growing faster than its architecture could bear — and that technical creakiness, combined with missteps in design and leadership, sent American users fleeing to faster, friendlier rivals.
The afterlife was a slow translation. In December 2009 Friendster was acquired by the Malaysian internet and payments company MOL Global for a reported $26.4 million; in 2010 Facebook bought Friendster’s portfolio of social-networking patents for a reported $40 million, extracting the one durable asset. In June 2011, conceding the social-network race entirely, MOL relaunched Friendster as a social-gaming site aimed at its strong Southeast Asian base. That pivot, too, ran out of road. On June 14, 2015, Friendster suspended the site and all its services, citing a challenging industry and a lack of user engagement; the company itself was formally wound down by 2018.
What its users lost was the first social graph many of them ever built — early-2000s profiles, testimonials, and connections from the dawn of the form. What the industry kept was Friendster’s blueprint and its cautionary tale, studied ever since as the canonical example of a pioneer that proved a market exists, then lost it to the competitors its own existence summoned.
Orkut was Google’s first social network, and on September 30, 2014 Google switched it off. It launched on January 22, 2004 — a week before Mark Zuckerberg registered TheFacebook — as a “20 percent” side project by a Google engineer named Orkut Büyükkökten, who gave it his own first name. For a brief moment in 2004 it was the social network Silicon Valley was talking about. Then it did something none of Google’s later social efforts ever managed: it became genuinely, overwhelmingly beloved, just not where its makers lived. Orkut was a phenomenon in Brazil and India and an afterthought everywhere else, and that geographic accident is the whole of its story.
The numbers, where they can be pinned down, point south and east. By 2008 Orkut was among the most-visited sites in both Brazil and India, commanding a reported 90 percent-plus of the Brazilian social market at its height; estimates of its peak put it around 300 million registered users worldwide, the large majority of them Brazilian and Indian. When Google finally published the traffic breakdown at shutdown, Brazil accounted for roughly 55 percent of users and India about 18 percent, with the United States — Google’s home, the world’s largest social-media market — a distant also-ran. Orkut was a hit record that never charted in the country that pressed it.
That mismatch sealed its fate. Facebook overtook Orkut in India around 2010 and in Brazil around 2012, and Google had by then placed its entire social bet on Google+. On June 30, 2014, a Google engineering director in Brazil announced the shutdown in a blog post, explaining that “Facebook, YouTube, Blogger and Google+ have taken off” and that “the growth of these communities has outpaced Orkut’s growth.” It was a polite way of saying that the one social network Google built that people actually loved was not the one Google had decided to keep.
What its users lost was a genuine community and a decade of “scraps,” testimonials, and communities — Orkut’s distinctive features, more affectionate and freewheeling than Facebook’s. Google let people export their data via Takeout and preserved the public communities as a permanent read-only archive. But a network that had been the digital town square for two of the world’s largest countries went dark while a near-empty Google+ marched on. Five years later Google+ would die too, having never been wanted; Orkut died having been adored, just not by the right people.
Path was the social network built on the radical premise that a social network should be small, and on October 18, 2018 the last of it switched off. It launched in November 2010, founded by former Facebook executive Dave Morin alongside Napster co-founder Shawn Fanning and designer Dustin Mierau, with a single defining constraint: you could have only 50 friends. The cap was a thesis — that real intimacy lives among your closest people, not a sprawling list of acquaintances — and Path wrapped it in some of the most admired interface design of its era. It was the thinking person’s anti-Facebook, and for a few years it was the app other apps were measured against for craft.
The thesis ran straight into the economics. A network deliberately capped at 50 friends — later raised to 150, then higher — is a network that has engineered away its own virality, because the whole engine of social-media growth is the unbounded friend-of-a-friend cascade that Path had explicitly forbidden. Path raised something on the order of $70 million from blue-chip investors and reportedly turned down a Google acquisition offer north of $100 million in 2011; reports later put its registered base around 50 million, with its single largest market not the United States but Indonesia. Impressive numbers for a boutique app, and nowhere near the scale its valuation and its venture backers required.
What did the most damage to its reputation was self-inflicted. In early 2012 a developer discovered that Path’s iPhone app was silently uploading users’ entire address books — names, phone numbers, emails — to its servers without asking. For a product whose entire pitch was privacy and intimacy, it was the worst possible scandal, and it cost Path an $800,000 FTC settlement in February 2013, not for the upload itself but for collecting personal information from children under 13 in violation of COPPA. The contradiction lingered: the network that promised to protect your inner circle had quietly harvested it.
Path never found its second act. It sold to South Korea’s Kakao in May 2015, largely for its foothold in Indonesia, and limped on as a regional product before the new owners announced its closure on September 17, 2018, with final shutdown on October 18. Its users lost a genuinely lovely, carefully made place — one of the few social apps that ever felt designed rather than optimized. Path’s epitaph is a paradox the industry still hasn’t resolved: it set out to prove that smaller could be better, and proved instead that smaller could not be a business.
FriendFeed was the real-time social aggregator that the technology industry’s early adopters loved and almost nobody else used, and on April 9, 2015 Facebook — which had owned it since 2009 — finally switched it off. Launched in 2007 by four former Google engineers, FriendFeed pulled a person’s activity from dozens of services — blog posts, Twitter updates, photos, bookmarks, any RSS feed — into one live, continuously updating stream, and then layered fast conversation on top of it. Comments and likes appeared instantly; a popular thread could surface and ignite a discussion in real time. For a particular kind of user — bloggers, developers, journalists, the people who lived on the leading edge of the social web — it was the best conversation tool of its moment.
It was also, commercially, a service ahead of its audience. FriendFeed’s user base was small, intensely engaged, and disproportionately influential, but it never crossed into the mainstream, and it had no obvious path to the scale or revenue that would make it a standalone business. So in August 2009 Facebook acquired it. The reported terms — roughly $47.5 million, split as about $15 million cash and $32.5 million in Facebook stock — were never officially confirmed; Facebook’s announcement said only that the financial terms were not disclosed.
What Facebook bought, in the end, was less the product than the people who made it. The acquisition was widely understood as an acqui-hire: FriendFeed’s four founders — Bret Taylor, Paul Buchheit, Jim Norris, and Sanjeev Singh — and their team joined Facebook in senior engineering and product roles, and over the following years helped build large parts of Facebook’s infrastructure. Bret Taylor became Facebook’s chief technology officer; Paul Buchheit, who had created Gmail and coined “Don’t be evil,” went on to Y Combinator. The team’s fingerprints ended up on the real-time News Feed, the Like button, and tools that outlived FriendFeed by more than a decade.
FriendFeed the service, meanwhile, was kept on life support — left online but undeveloped for nearly six years, its small community slowly thinning, until Facebook announced on March 9, 2015 that usage had “declined steadily” and it would shut the service down. On April 9, 2015, FriendFeed went dark. It had been acquired not to grow, but to be absorbed — and it had done its real work the day its founders signed on.
Google Buzz was Google’s first real attempt to build a social network, and it lasted barely a year and a half before the company switched it off to chase the next one. Launched on February 9, 2010, Buzz was not an app you signed up for so much as a feature that appeared, unannounced, inside Gmail. It let users post status updates, share links and photos, and follow one another in a stream stitched directly into the inbox hundreds of millions of people already opened every day. On paper this was Google’s killer advantage: it did not have to recruit a single user, because it had simply enrolled them all.
That same shortcut was its fatal flaw. To give every new Buzz user an instant social graph, Google auto-generated each person’s public follower and following lists from the Gmail contacts they emailed and chatted with most. The list of the people you talked to most was, by default, displayed on your public Google profile for the world to read. For most users this was merely unsettling. For some — a journalist’s confidential sources, a therapist’s patients, a woman whose abusive ex-husband could now see her new partner and workplace — it was dangerous. The backlash was immediate, loud, and entirely justified.
Within days Google was apologizing and rewiring the product’s privacy defaults, but the damage was structural. A class-action lawsuit produced an $8.5 million privacy-education settlement, and the Federal Trade Commission charged Google with deceptive practices, securing a landmark consent order: the first time the FTC required a company to build a comprehensive privacy program and submit to independent audits every two years for twenty years. Buzz never recovered its reputation, and Google’s attention had already moved on. On October 14, 2011, in a corporate-blog post titled “A fall sweep,” Google announced Buzz would be retired so it could focus on its new social bet, Google+. The service was fully discontinued on December 15, 2011.
What users lost was small — Buzz never became the place anyone genuinely lived — but what it taught was large. Buzz was the moment Silicon Valley’s “launch first, ask permission never” instinct collided with the simple fact that an email account is a map of a person’s whole private life, and that you cannot make it public as a convenience.
App.net was the ad-free, subscription-funded “Twitter alternative” that asked a genuinely good question — what if the users were the customers instead of the product? — and then could not find enough of them to keep the lights on. Founded by Dalton Caldwell and Bryan Berg under Mixed Media Labs, it launched on August 8, 2012 on a single, principled premise: members would pay a fee, so the service would answer to them rather than to advertisers, and the perverse incentives that warped ad-funded social networks would simply never apply. It was a real-time microblogging platform with a developer-friendly API at its core, and for a stretch it was the most idealistic thing on the social web.
The founding act was a crowdfunding campaign that became a referendum on whether people would pay for a better social network. Caldwell set a $500,000 goal and a hard deadline, and the campaign cleared it, ultimately raising roughly $803,000 from over 11,000 backers. Paid plans ran $50 a year or $5 a month, and a developer tier targeted the builders Caldwell believed would make the platform indispensable. In May 2013 App.net reported it had passed 100,000 users. The early excitement was real, and so was the affection of the small, technical community that gathered there.
The arithmetic, however, never closed. The thesis depended on a flywheel — paying members would attract third-party developers, whose apps would attract more paying members — and the flywheel never reached escape velocity. Renewals lagged, growth stalled, and on May 6, 2014 the founders conceded that subscription income could no longer fund full-time staff; App.net went into maintenance mode, kept running on autopilot from the revenue it still had. On January 12, 2017 Caldwell and Berg announced the service would close, and on March 14, 2017 it shut down for good, with the platform code open-sourced and user data deleted thereafter.
What App.net lost was not scale — it never had much — but a genuinely better idea, run by people who believed it. Its failure did not disprove the thesis that users should be the customer; it proved how hard that thesis is to fund when the incumbent is free and already has everyone’s friends.
Periscope was the live-streaming app that arrived as a sensation, defined a moment, and then was quietly digested by the company that owned it. Built by Kayvon Beykpour and Joe Bernstein, it was acquired by Twitter in January 2015 — before it had even launched publicly — for a reported sum somewhere between $50 million and $100 million. Twitter unveiled the app on March 13, 2015, and shipped it on iOS on March 26, with timing that was no accident: a rival app called Meerkat had just become the breakout hit of the South by Southwest festival, and Periscope was Twitter’s pre-loaded answer. The app let anyone broadcast live video from their phone to a worldwide audience that could tap hearts and comment in real time, and for a year or two it felt like the future of the platform.
The early numbers were genuinely strong. By August 2015 Periscope reported it had surpassed 10 million accounts, with viewers watching the equivalent of 40 years of live video every day, and in December 2015 Apple named it iPhone App of the Year. For a brief window, Periscope was the place where breaking news, spontaneous events, and ordinary life went live, and it gave Twitter a credible claim to own a format it had never built itself.
Then the absorption began. Live video was Periscope’s whole reason to exist, but it was also exactly the feature Twitter most wanted inside its own app. Beginning in December 2016, Twitter wired Periscope’s broadcasting directly into the main Twitter client, so users could go live without ever opening the standalone app. From that point the standalone Periscope was a redundancy waiting to be retired. On December 15, 2020, in a Medium post titled “Farewell, Periscope,” Twitter announced the app would shut down, describing it as stuck in an unsustainable maintenance-mode state with declining usage and rising upkeep costs. The standalone app was discontinued on March 31, 2021. Past broadcasts shared to Twitter survived as replays; the app itself did not.
Periscope’s was not a death by failure but a death by success — its core function worked so well that its parent took it in-house and let the original husk go dark. What its broadcasters lost was not the ability to go live, which moved into Twitter, but the dedicated home and community that had grown up around a separate app.
Friends Reunited was Britain’s pioneering social network — a school-reunion site that taught a country how to find its old classmates online years before anyone said the word “Facebook” — and on 26 February 2016 its owner quietly switched it off after sixteen years. Conceived in 1999 by Julie and Steve Pankhurst of Barnet with their friend Jason Porter, and launched in June 2000, it began as a homemade directory of UK schools that let people register, find their year group, and reconnect with friends they had lost touch with for decades. It was a genuine national phenomenon: 3,000 members by the end of 2000, 2.5 million a year later, and a fixture of British internet life when most of the country was still on dial-up.
The numbers that matter most bracket its decline. In December 2005, with over 15 million members, Friends Reunited was bought by the British broadcaster ITV for £120 million up front plus performance payments of up to £55 million — a deal widely reported as worth £175 million. ITV believed it had purchased the future of social interaction; what it had actually purchased was the past, because Facebook had launched the year before and was about to make the entire “find your old schoolmates” premise feel quaint. By August 2009 ITV gave up, selling Friends Reunited to the Dundee publisher DC Thomson’s Brightsolid subsidiary for £25 million — a loss of roughly £150 million on the headline figure, and one of the most expensive misjudgements of the British dot-com era.
What followed was a long managed decline. DC Thomson valued the site at just £5.2 million by late 2011, relaunched it in March 2012 around “memories” rather than reunions, and eventually handed it back to co-founder Steve Pankhurst. His conclusion, posted in January 2016, was blunt: the site was no longer used for what it was built for, most members had registered a decade earlier with contact details long out of date, and the business could not cover its costs. Britain’s first social network did not die in a crisis. It died of being outgrown, having already cost its most ambitious owner a fortune.
SixDegrees.com was the first thing that looked like a social network, and it died before the world was ready for one. Launched in 1997 by the New York entrepreneur Andrew Weinreich and his company MacroView, it let users create a profile, list their friends, family, and acquaintances, and then traverse those connections outward — sending messages and seeing how they were linked to anyone else on the site, first degree to second to third. The name came from the “six degrees of separation” idea, and the concept was, in retrospect, exactly right: the profile, the friend list, the friends-of-friends graph. Every social network since has been a remix of what SixDegrees did first.
The numbers, for 1997, were genuinely large. At its peak the site had around 3.5 million fully registered members and close to 100 employees, and in December 1999 it was sold to YouthStream Media Networks for a reported $125 million — a headline price that, arriving near the top of the dot-com mania, made Weinreich a textbook example of selling at exactly the right moment. The patent he and his co-inventors filed on the underlying method, US Patent 6,175,831, would later be regarded as the foundational social-networking patent and was eventually bought up by companies including LinkedIn.
What SixDegrees could not buy was time. The technology and habits that make social networks work — broadband always-on connections, digital cameras and photos worth sharing, a critical mass of friends already online — did not yet exist in the late 1990s, so users built their networks and then found little to do in them. When the dot-com bubble burst, YouthStream’s finances collapsed under it, and the site that had invented the category was wound down around 2000–2001, catalogued as shut by 2001. It was not outcompeted; nothing yet existed to compete with it. It was, in its founder’s own diagnosis, simply too early — the right idea waiting roughly five years for the rest of the internet to arrive.