App.net

» see app.net as it lived, on the wayback machine

~ the last words ~

app.net launched on august 8, 2012 from dalton caldwell's startup mixed media labs. caldwell had been publicly arguing that twitter was about to break its api in ways that would kill third-party clients, which turned out to be roughly correct a year later. his pitch was an alternative microblogging network funded by user subscriptions instead of ads, with a real api that developers could build on top of. the bare minimum to start was $50 a year for a personal account, $100 for a developer tier.

it crowdfunded fast. the goal was $500,000. the campaign hit it on august 13, 2012, ended at around $750,000 with over 11,000 backers. tech press covered it as a serious experiment. messages were 256 characters instead of 140. the early user base was disproportionately developers and podcasters, the kind of twitter user who already paid for software. for about six months it felt like a working second-tier social network for the people who built things.

the model never scaled past that early base. mixed media labs introduced a free tier on february 25, 2013, hoping to grow the user count past 100,000. it climbed to around 100k registered users by may 2013 and stalled there. paid renewals were the real number that mattered. on may 6, 2014, caldwell posted that subscription renewals had been so weak there was no longer cash to keep development staff on. the team was let go. the service kept running on autopay infrastructure for almost three more years.

the final shutdown announcement came on march 13, 2017. the deadline was march 15, then got extended to march 17 because the data export script kept failing under load. at 06:45 utc on march 17, 2017, app.net stopped responding. the post archive is gone except for what users had personally exported. the api is closed. the domain still resolves to a static notice.

i never used app.net. by 2012 i was on twitter where the network already lived in turkey and across the global engineering crowd, with no friction to leave. paying $50 a year for a private second twitter only made sense if your actual circle came along. mine did not. app.net was a small US developer-tier social network with almost no presence in turkey, even among engineers who would have been the natural audience. i read the announcements when they happened and watched it from the outside.

~ the record ~

~ what the pitch was ~

a microblog network as paid infrastructure. dalton caldwell argued openly that twitter could not stay free and developer-friendly forever, since the only revenue model that would scale was advertising plus the closed app ecosystem that ad funding would require. his alternative was a network where users paid the bill themselves, and where the api was a first-class commitment instead of a tolerated side door. the math required either a small loyal paying base or a much larger free base willing to upgrade. neither one materialized at the size needed to fund a team.

~ what came of the api ~

a small ecosystem of clients and bots. third-party readers like felix and riposte shipped on ios. a podcast network called 5by5 used app.net as its show comments backbone for a while. there were hashtag-tracking bots and file-storage extensions, plus a couple of working federated chat experiments. when the funding model gave out, the surrounding ecosystem went with it. some of those developers later moved their work to mastodon when the fediverse took shape a few years later.

~ why it lost ~

twitter did not break the way caldwell expected on the timeline he expected. the api restrictions came in 2013, then again in 2018, but each time twitter remained free at the front door and paid clients stayed niche. the average user had no reason to pay $50 a year for a parallel network where their friends were not. by the time twitter actually started feeling broken on the open-internet axis, around 2022 to 2023, the alternative that picked up the runoff was mastodon, not a paid twitter clone. app.net was right about the diagnosis and wrong about the cure.

~ how it ended ~

on time. dalton caldwell had been transparent about the financial situation since the may 2014 layoffs. when the maintenance budget finally ran out, the company posted a march 13, 2017 notice giving users a few days to export. the export tool itself was overloaded and pushed the cutoff to march 17. the shutdown was unusually clean for a social network. there was no surprise and no quiet decay over months, just a posted date and a service that went offline at the hour stated.

@}-->--

              )
             (_)
            .-'-.
            |   |
            |   |
            |   |
            |   |
          __|   |__   .-.
       .-'  |   |  `-:   :
      :     `---'     :-'
       `-._       _.-'
           '""""""

« previous  ·  random  ·  next »