Google Reader was the web’s dominant RSS reader, and on July 1, 2013 Google switched it off. Launched out of Google Labs on October 7, 2005, Reader did one thing exceptionally well: it collected the feeds of every blog, news site, and webcomic a person subscribed to and presented them in one fast, keyboard-driven inbox. For the news-obsessed it became indispensable — the home page of the open web — and over eight years it quietly accumulated tens of millions of users and a genuinely devoted core. Google never ran an ad against it, and that, in the end, was the problem.
The announcement came on March 13, 2013, buried in a corporate blog post cheerfully titled “A second spring of cleaning.” Reader, Google said, would be retired on July 1 because “usage has declined” and the company wanted to “focus on fewer products.” The reaction was immediate and disproportionate to the size of a free RSS tool: a Change.org petition gathered roughly 150,000 signatures within days, rival readers were swamped overnight, and a generation of power users concluded — loudly, and not for the last time — that Google could not be trusted to keep anything alive that it could not monetize.
What made the death sting was that it looked self-inflicted. In October 2011 Google had stripped Reader of its beloved built-in sharing and social features and rerouted them through Google+, the social network Google was then betting the company on. The de-featuring alienated the very users who made Reader special, depressed engagement, and supplied, eighteen months later, the “declining usage” that justified the shutdown. Reader was not killed because it was failing; it was made to fail, then cited as a failure.
Its users scattered to Feedly, NewsBlur, Inoreader, and The Old Reader, and RSS survived — more fragmented, less central, but alive. What did not survive was the assumption that a free, beloved Google product was a permanent fixture. Reader became the founding artifact of the “Google graveyard,” the case every subsequent shutdown is measured against, and a durable lesson in what it means to build your routine on something you don’t pay for.
AIM — AOL Instant Messenger — was how a generation of Americans first talked to their friends online, and on December 15, 2017 it was switched off after twenty years. Launched as a standalone app in May 1997, AIM turned the buddy list, the away message, and the soft door-creak of a friend logging on into the texture of everyday social life. For students especially it was the after-school commons: you came home, you signed on, and the little window told you who else was there. It was the place a whole cohort learned what it felt like to be reachable.
At its height it was dominant. AIM reached roughly 61 million users by 2000 and held something like 52 percent of the US instant-messaging market by the mid-2000s, the clear leader over Yahoo Messenger, MSN Messenger, and the ICQ network AOL also owned. Its cultural footprint outran even its user count: the away message — the little broadcast status you left when you stepped away — was a precursor to the tweet and the status update, a way of performing your mood to your buddy list before “posting” was a verb.
What killed AIM was not a better instant messenger; it was the disappearance of the need for a standalone one. Text messaging moved real-time chat into the phone in everyone’s pocket. Then Gchat folded messaging into email, and Facebook folded the buddy list into the social graph everyone already lived in. The behavior AIM pioneered didn’t die — it migrated into devices and platforms that did it without a separate program to open. AIM’s parent, AOL, declined alongside it, sold to Verizon in 2015 and folded into the Oath media division; by March 2017 AIM was down to single-digit millions of users, and keeping its aging OSCAR messaging protocol running for them no longer made sense.
This entry is wry about the corporate decline — a company once worth hundreds of billions reduced to a media-division line item — but gentle about the thing itself. AIM was where a lot of people had their first online conversations, their first crushes typed at midnight, their first sense of an internet that was social. It was outgrown rather than defeated, and that is its own kind of ending.
Google Wave was the most exciting product nobody could use, then the most baffling product nobody wanted, and on April 30, 2012 Google shut off its servers and gave the code to the Apache Software Foundation. Announced on May 28, 2009 at the Google I/O developer conference, Wave was pitched with a line that became famous: what would email look like if it were invented today? The answer was a single real-time medium that fused email, instant messaging, documents, and wikis into shared “waves” where every keystroke appeared live, conversations could be replayed from the beginning, and bots and extensions could plug in anywhere. The demo, led by Google Maps co-creators Lars and Jens Rasmussen, drew rapturous applause and instant hype.
Then Google did something that would define the saga: it made Wave artificially scarce. Rather than a public launch, it opened a limited preview to 100,000 users in September 2009, each able to dole out a few invitations, turning access into a status symbol and the wider audience into spectators. By the time Wave opened to the general public in May 2010, the buzz had curdled into a single, devastating question that no one — not even Google — could answer cleanly: what is this actually for?
The reckoning was swift. On August 4, 2010, barely three months after the public launch and a little over a year after the standing-ovation debut, Google announced it would stop developing Wave as a standalone product, citing a lack of user adoption. The service lingered as a courtesy: existing waves went read-only on January 31, 2012, and the servers were switched off on April 30, 2012. Google had already, in December 2010, donated the code to Apache, where it became Apache Wave (the “Wave in a Box” server) — a project that never left incubator status and was formally retired on January 15, 2018.
Few users lost data they grieved, because few had truly committed; Wave’s tragedy was not a bereaved community but squandered brilliance. It remains the textbook case of a technically dazzling product undone by hype it could not satisfy and a purpose it could never articulate.
Picasa was the fast, much-loved desktop application for finding, organizing, and lightly editing the photos already sitting on your hard drive, and on March 15, 2016 Google stopped supporting it. First released by a small company called Lifescape on October 15, 2002, Picasa was acquired by Google in July 2004 and from that point given away free. It did something that, in retrospect, feels almost quaint: it scanned your computer, gathered every image into one clean library, and let you crop, retouch, tag faces, and make albums without uploading anything to anyone. For more than a decade it was the default answer to “how do I deal with all these photos,” and it earned a loyalty that outlived its own updates.
The end was announced on February 12, 2016, in a blog post titled “Moving on from Picasa” by Anil Sabharwal, the head of Google Photos. The desktop application would no longer be supported as of March 15, 2016; the companion Picasa Web Albums service would begin shutting down on May 1, 2016. The stated reason was consolidation: Google wanted to “focus entirely on a single photo service in Google Photos” rather than “divide our efforts across two different products.” Picasa was not failing. It was simply on the wrong side of a strategy.
That strategy was the cloud. Google Photos, launched in May 2015, was mobile-first, automatic, and stored everything on Google’s servers — the opposite of Picasa’s local, you-own-the-files model. Killing Picasa was a bet that people no longer wanted to manage a library on a machine they controlled; they wanted it managed for them, somewhere else. For most users that bet was right. For the considerable minority who valued Picasa precisely because it kept their photos on their own disk and out of anyone’s cloud, the discontinuation was a quiet eviction from a workflow that had no real successor.
The desktop app, notably, did not detonate. Google did not push a kill switch; the program kept running on machines that already had it, unsupported and slowly aging. But “still works for now” is not the same as “alive,” and Picasa joined the long ledger of beloved Google products retired in the name of focus — this time not because users left, but because the company decided where they ought to keep their pictures.
Windows Phone was Microsoft’s genuinely good third mobile operating system, and it died not for lack of quality but for lack of apps. Launched as Windows Phone 7 on October 21, 2010, it broke decisively from the iPhone-and-grid template with “Live Tiles” — dynamic, color-blocked panels that surfaced information on the home screen instead of static icons. Reviewers praised the design’s clarity and originality, and on the well-built Nokia Lumia hardware it had a real argument as the most distinctive phone you could buy. By 2017 active development had effectively ended; Microsoft confirmed the platform’s last release would reach end of support on December 10, 2019.
The cause of death has a name developers used at the time: the “app gap.” A modern phone is only as useful as its software, and Windows Phone never attracted the breadth of apps that iOS and Android took for granted. The pattern was a vicious circle — too few users to justify the cost of building and maintaining an app, and too few apps to attract more users. Headline services arrived late, in crippled form, or never at all, and every missing or outdated app was another reason for a buyer to choose an iPhone or an Android instead. The OS could be elegant; the ecosystem was empty, and the ecosystem is what people actually live in.
The bet that was supposed to break the cycle made the failure catastrophic instead. In 2014 Microsoft bought Nokia’s devices and services business for roughly 7.2 billion dollars, aiming to control the hardware and force scale. In July 2015 it wrote down approximately 7.6 billion dollars on that acquisition — more than the purchase price — and cut about 7,800 jobs, overwhelmingly in the phone division. The numbers told the story plainly: Microsoft had spent billions trying to buy its way into a market whose users were already locked into two other ecosystems, and the market refused to follow.
Windows Phone’s discontinuation is the textbook case of the app-gap death and of network effects that money cannot purchase. It was not killed by a rival feature or a scandal; it was starved. A platform with no users gets no apps, a platform with no apps gets no users, and Microsoft, despite all its resources, never found the lever to break that loop before the loop broke the product.
Parse was a mobile backend-as-a-service — a hosted set of tools that let an app developer skip building and running servers entirely — and on January 28, 2017 Facebook switched it off. Founded in 2011 by Tikhon Bernstam, Ilya Sukhar, James Yu, and Kevin Lacker, Parse sold a deceptively simple promise: write your iOS or Android app, point it at Parse for data storage, user logins, and push notifications, and never think about a database, a server, or a sysadmin again. For a generation of mobile startups and solo developers, that promise was irresistible, and Parse grew fast.
In April 2013 Facebook acquired the company in a deal widely reported at roughly $85 million, its first serious move into paid developer tools. Under Facebook’s banner Parse kept growing; by 2014 it was reported to power around 500,000 mobile apps, and at shutdown TechCrunch noted that at one point some 600,000 apps relied on the platform. For nearly three years it looked like a rare example of a developer service that an acquisition had strengthened rather than smothered.
Then, on January 28, 2016, Facebook announced that Parse would be wound down over exactly one year and shut for good on January 28, 2017. There was no scandal, no security breach, no collapse in usage cited — only a corporate decision to focus elsewhere. Facebook gave developers a year, a database-migration guide, and, unusually, the platform’s own source code, open-sourced the same day as Parse Server so that anyone could stand up their own copy. It was about as graceful as a shutdown gets, and it still stranded hundreds of thousands of apps whose owners had to scramble to migrate or go dark.
That is the lasting weight of the Parse story. Tens of thousands of developers had taken Facebook at its word and built their products on infrastructure they did not control. When the owner’s priorities changed, the polite one-year window did nothing to change the basic fact: every one of those teams now had to rebuild the foundation of a shipping product, on a deadline they did not set, for a reason that had nothing to do with them.
Grooveshark was a free music-streaming service where users uploaded the songs, and on April 30, 2015 it shut down overnight as the price of a legal settlement — admitting infringement, surrendering everything it owned, to escape damages that could have run past $700 million. Launched in March 2006 by three University of Florida undergraduates, Andrés Barreto, Josh Greenberg, and Sam Tarantino, and run through Escape Media Group, Grooveshark let anyone upload an audio file and stream anything in the library for nothing, supported by ads. It was, for years, one of the easiest ways on the internet to play almost any song instantly.
That ease was its appeal and its original sin. Grooveshark’s catalog — at its height the company claimed over 15 million songs, more than a billion streams a month, and around 20 million users — was assembled largely from files its users uploaded, very few of them licensed. The service leaned on the legal shelter that protects platforms from what their users post. The labels argued, and ultimately proved, that Grooveshark was not a passive host but an active participant: its own employees had been instructed to upload copyrighted recordings as a condition of employment.
The litigation was long and, once that fact emerged, lopsided. Universal Music Group sued in 2010; a nine-label coalition including Sony, Warner, and Arista followed in 2011. In September 2014 a federal judge in New York granted summary judgment, finding Escape liable for direct and secondary infringement over thousands of recordings. With statutory damages of up to $150,000 per work and 4,907 works at issue, the company faced a theoretical $736 million in liability — an extinction-level number for a startup.
So Grooveshark settled to survive being erased rather than be bankrupted by a verdict. On April 30, 2015 it ceased operations immediately, posted a public apology, wiped its catalog, and handed its website, apps, and intellectual property to the record companies. It is the cleanest illustration in this catalog of a simple rule: a service whose product is other people’s copyrighted work, taken without permission, is not a business with a legal problem — it is a lawsuit that happens to stream music.
Rdio was a beautifully made music-streaming service that almost everyone who used it admired, and on December 22, 2015 it was discontinued because admiration does not pay royalties. Launched on August 3, 2010 by the Skype founders Niklas Zennström and Janus Friis, Rdio married a large licensed catalog to an interface widely judged the best in the category — clean, fast, with social features that let listeners follow each other and see what friends were playing. It reviewed brilliantly. Entertainment Weekly called it the best app and online interface in streaming. For five years it was the connoisseur’s choice.
It was also, the whole time, losing the only race that mattered. Spotify launched in Europe a year before Rdio and reached the United States in mid-2011, and from there it simply outspent, out-marketed, and out-grew its more elegant rival on every axis that compounds in a network business. Rdio raised serious money — reported at well over $100 million, against the more than $200 million its backer Janus Friis poured in across its life — and still could not buy the scale that Spotify’s marketing, free tier, and partnerships were assembling. By the end, Rdio reportedly had fewer than 200,000 paying subscribers.
The economics were unforgiving. By 2015 Rdio was reportedly spending around $3.5 to $4 million a month, mostly on payroll, against roughly $1.5 million in monthly revenue — losing some $2 million a month with no path to closing the gap. When it filed for Chapter 11 bankruptcy on November 16, 2015, court documents listed about $188.5 million in secured debt. The company that built the best product in streaming was, by every financial measure, a failure.
The end was a managed transfer rather than a collapse into nothing. Pandora agreed to buy Rdio’s technology and intellectual property for $75 million in cash, hiring roughly a hundred of its employees to accelerate Pandora’s own subscription ambitions; the deal closed on December 23, 2015, and the Rdio service went dark the day before. Rdio is the catalog’s clearest case of a hard truth founders hate to hear: in a market governed by network effects and the deepest war chest, the best product does not always win.
Mailbox was the iOS email app that arrived in February 2013 as the most wanted download in the App Store, and on February 26, 2016 its owner Dropbox switched it off without ceremony. Built by a tiny San Francisco startup called Orchestra, it reimagined the mobile inbox around two gestures — a short swipe to archive, a long swipe to snooze a message until later — and around the idea, briefly seductive, that email could be tamed into a to-do list you actually cleared. For a few weeks in early 2013 it was the most talked-about app in technology.
What made Mailbox a phenomenon before anyone had used it was the waitlist. Rather than open the doors, Orchestra made every new user take a numbered ticket and watch a live queue counter tick down the hundreds of thousands of people ahead of them. The artificial scarcity worked exactly as designed: it manufactured demand, generated weeks of press, and turned a free email client into an event. Then, barely a month after launch, Dropbox acquired the company for a figure reported at roughly $100 million — and the app’s real purpose quietly changed.
Because the acquisition was never really about email. Dropbox, then racing to justify a multibillion-dollar valuation, wanted Orchestra’s design talent and a beachhead on mobile, and Mailbox was the vehicle. The app got an Android version and a Mac beta in 2014, and then very little else. By late 2015 Dropbox had decided that its future was workplace collaboration, that email was not core, and — in its own words — that it could not “fundamentally fix” the inbox. On December 7, 2015 it announced that Mailbox and its photo app Carousel would both close.
When the servers went dark on February 26, 2016, the lesson was already written: Mailbox was an acqui-hire that was never built to last, a beloved product acquired chiefly for the people who made it. The gestures it popularized — swipe to archive, snooze for later — outlived the app by years, absorbed into Gmail, Outlook, and Apple Mail. The app itself became a tidy parable about what happens when a product millions queued for becomes, on the acquirer’s balance sheet, a means to an end.