StackMob

» see stackmob.com as it lived, on the wayback machine

~ the obituary ~

StackMob sold mobile developers a backend they never had to build. You wrote the iPhone or Android side, and the accounts, the stored objects, the push notifications and the API in front of them all came from somebody else's servers. In 2011 that was a real saving, because the alternative was standing up a database behind an application server and then wrestling a certificate chain for push, none of which is the app.

The company started in San Francisco in 2010, raised $7.5 million, and spent four years in a category that briefly looked like the future of app development. Parse was the better known competitor. Both were selling the same argument to the same audience, and both ended up being bought by a much larger company for reasons that had nothing to do with the product continuing.

PayPal announced the acquisition on 17 December 2013. The framing at the time was that PayPal wanted a developer platform and StackMob had built one. Eight weeks later, on 12 February 2014, PayPal said it was closing the backend service and retiring the StackMob name so the combined team could work on payments instead. Data export closed on 31 March. The servers went off on 11 May.

~ the case file ~

~ what a backend service sold you ~

The value was in what you did not have to think about. A push notification looks like one line of code and is actually a certificate, a token store, a per-platform gateway and a retry policy. User accounts look like a form and are actually password hashing plus a reset flow that has to survive people typing their email address wrong. StackMob had all of it behind an SDK.

The cost was that your entire data layer sat inside a startup's business model. That is fine while the startup is independent and looking for growth. It stops being fine the moment somebody buys the startup for its engineers, because the engineers are the asset and your app is an operating expense attached to them.

~ the acqui-hire shape ~

Fourteen engineers were reported as working on the service when it closed. PayPal did not want the service. It wanted a mobile team that had already shipped SDKs to strangers and survived the support load, and $7.5 million of prior investment had paid for the training. Nobody involved described it that way in public, and the shutdown post used the standard phrasing about focus.

Two years later Facebook did the same thing to Parse on a longer clock: announced in January 2016, off in January 2017, twelve months of notice instead of three. Parse also open-sourced its server on the way out, which StackMob's developers were never offered.

~ what happened to the category ~

The idea survived its companies. Firebase went to Google in 2014 and is still the default answer for a small team that needs auth and a database without a server. Supabase and Appwrite arrived later with the open source version of the same promise, which exists mainly because of what happened to Parse and StackMob.

I have never shipped a phone app that needed one of these. The work I do runs on hardware that has to keep going when the network does not, so the whole premise, that somebody else holds your state and you call them, is the wrong shape for it. That does not make the category wrong. It makes it a bet on connectivity that a lot of products were correct to take and a few were not.

@}-->--

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

« previous  ·  random  ·  next »