SpoonRocket

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

~ the obituary ~

spoonrocket was founded in 2013 by anson tsui and steven hsiao, two UC berkeley engineering and social engineering majors who had run a late-night campus food project called munchy munchy hippoes a few years earlier. they pitched the new idea to y combinator in may 2013 and launched at the end of june. the proposition was a single number: $6. one vegetarian option, one meat option, both cooked in a commissary kitchen, both delivered hot in fifteen minutes or less by drivers circling the berkeley flats with insulated bags in the back seat.

the model worked when the routes were tight. drivers ran loops with a heated trunk full of identical meals. the app placed the next customer on the route, and a meal came off the rack. it was less a restaurant and more a moving warehouse. through 2014 spoonrocket raised about $13.5 million between a seed round and a 2014 series a led by foundation capital with sherpa and base ventures, and pushed into san francisco. the price held at six dollars. food costs and driver pay did not.

on 14 march 2016 berkeleyside broke the news that spoonrocket had stopped running and was looking for a buyer. the next day co-founder steven hsiao confirmed the shutdown to techcrunch and pointed customers at sprig, the other bay area meal app on the way to its own funeral. on-demand was on fire that spring. homejoy was gone the previous july. sprig and maple would follow in 2017. spoonrocket lit up first.

i was watching this from ankara on a ttnet line, working on edge computer vision for unmanned aerial systems, eating öğretmenevi lunch downstairs from my desk. the idea of a hot meal driven to you in fifteen minutes did not translate to a turkish city where every block already had a lokanta and a simit cart. spoonrocket lived in the same north-california geometry as caltrain and dolores park, and that geometry did not export.

~ the case file ~

~ what it actually was ~

the trick was the route, not the kitchen. spoonrocket's drivers carried meals that had already been cooked and packaged in a central kitchen, so the app's job was to keep the trunks turning over before the heat died. uber eats had not yet launched as a separate product, and doordash was still focused on restaurant pickup. spoonrocket's bet was that pre-batched meals on a moving truck would beat the on-demand kitchen-pickup model on speed and on unit economics.

the bet only paid off in places where the customer density was high enough to keep a route hot. berkeley and a slice of san francisco qualified. anything further out broke the math.

~ how it ended ~

by early 2016 the on-demand fundraising window had closed. spoonrocket had reached a positive contribution margin per meal but could not cover overhead, and the next round was not coming. brazilian food delivery company ifood bought the technology and brand stub on 16 march, the day after the shutdown, and relaunched a different spoonrocket in brazil later that year using restaurants instead of a commissary kitchen. the us product never came back.

~ what carried over ~

spoonrocket sits in the canon of the 2016-2017 on-demand collapse next to [[sprig]], [[maple]], [[homejoy]], [[munchery]], and the rest of the post-uber x-for-y wave. it was the first of the meal-batch startups to discover that pre-cooked food in a moving trunk only beats a real restaurant if the route stays full, and routes get expensive when growth slows. the six-dollar lunch was a thing for two and a half years and then it was a paragraph in a techcrunch shutdown post.

@}-->--

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

« previous  ·  random  ·  next »