In the early summer of 2018, you could buy an Apple Watch with built-in GPS, wireless payments, and speakers that buzz water out after a swim. Meanwhile, Katharine Berry was hustling to keep five-year-old watches with black-and-white screens alive. Berry had worked for Pebble, maker of the first notable smartwatch. The company was acquired by Fitbit and shut down a year and a half earlier. Now Fitbit was turning off the servers that fed Pebble’s apps, weather, and other useful data.
But Berry and a cadre of crafty enthusiasts, the Rebble Alliance, had prepared for this moment. They had archived Pebble’s web and development assets, opened up the devices’ firmware a bit, and worked with former Pebble and Fitbit developers inside a Discord channel. Berry, between jobs, sprinted for two weeks to code a replacement cloud infrastructure. She guessed that, if they could pull it off, maybe a thousand people, at most, would try it out.
One morning seven months later, Berry realized that Rebble had 100,000 accounts. Today, more than 212,000 accounts have been created — more than 10% of the two million Pebbles ever sold — and nearly 9,000 have subscribed. Press coverage certainly helped. But really it is the Rebblers’ enthusiasm that keeps their watches running, precisely because Pebbles are not modern, in all the best ways.
Rebble is an inspiring repair story, and the way Pebble enabled this second life is a path that every gadget manufacturer should strive to emulate. Pebble created an open (and open-source) environment for developers and enthusiasts. As a direct result, Rebble is saving thousands of gadgets from the bin and building a real community around dogged longevity. Keeping Pebbles running, in the face of much fancier options, knitted the community together.
Arduino Roots
Prior to founding Pebble, Eric Migicovsky was studying abroad at the University of Delft, Netherlands in 2008. He fit in with his sturdy, reliable cruiser bike (known to locals as an opafiets, or “Grandpa bike”), but he knew he would crash if he kept checking his text messages while riding. In his dorm room,
he patched together an Arduino controller, a few buttons, a battery, and the screen from a disassembled Nokia 3310. His first idea was a bike computer, but “someone was like, you should probably just make it a watch,” he later said.
Migicovsky’s refined model, called inPulse and built for BlackBerry phones, got him into startup bootcamp Y Combinator in 2011. He impressed founder Paul Graham, who said Migicovsky was the most likely pick to be “the next Steve Jobs.”
Creating the next watch required a lot more money, and venture capitalists are notoriously cautious about hardware. So in the spring of 2012, Migicovsky turned to Kickstarter. Yet again, his timing was keen. Kickstarter projects were becoming not just a way to fund your cousin’s art project, but a viable option for pre-order funding. Months earlier, a video game and iPhone dock had broken the $1 million mark on the same day.
The launch of Migicovsky’s next watch, called Pebble, left Kickstarter’s records in the dust, raising $1 million in 28 hours and finishing with a new milestone for Kickstarter at the time — over $10 million.
First Mover Disadvantage
The first Pebble, shipped after delays in January 2013, was a plastic watch with a black-and-white e-paper screen. Between that and the 2014 Steel release, Pebble had proven that people appreciated an alternative to constant phone-opening. But then came the inevitable: Apple debuted its Watch. Could Pebble perfect its device and broaden its appeal before Apple ate
the entire field?
The Kickstarter for the next Pebble, the Time, launched exactly two months before the Apple Watch would ship. The campaign still holds the record for the largest-ever Kickstarter at over $20 million. The Time, and its fancier Time Steel version, added a first-of-its-kind 64-color e-paper screen, voice dictation, and a novel, whimsically animated OS.
At the same time Pebble was breaking records and touting its indie appeal, the company had stopped being profitable. One company source told Business Insider that Apple had “sucked up all the oxygen.” Pebble missed its sales goals in late 2015, and its Black Friday sales in 2015 were down from the year before. What followed were layoffs, a failed acquisition by Intel, and trouble finding more capital. Pebble’s final Kickstarter in May 2016 was, in effect, a bridge loan from its fans.
The next wave of Pebbles focused on fitness, something Apple was already pivoting toward with its second Watch, but Pebble didn’t have Apple’s money. After months of last-ditch fundraising attempts, Fitbit acquired Pebble’s software assets and engineer hiring rights in December 2016 for a scant $23 million. While Fitbit would not officially support Pebble’s customers, Migicovsky worked out a deal that would refund Kickstarter pre-orders, and, he hoped, keep the more than two million Pebbles sold, and their apps, working for as long as possible.
In hindsight, he nailed it. Pebble has been the most successful hardware company failure in history. Compare this to Revolv, whose acquisition by Nest led to an abrupt shutoff of smart homes around the world, or personal cloud device Lima, or, really, any Android device more than a couple years old.
The Panic Store
News of the shutdown shook the Pebble community. The official message was that Pebbles “will work normally for now,” but “Functionality or service quality may be reduced down the road.” Without web services from Pebble, watches would lose their app store, language packs, voice dictation, weather, and even the icons on their notifications. Nobody knew anything for sure, other than that Pebble, the company, no longer existed. “We were all just freaking out,” says IShotJr, a longtime Pebble developer, hardware hacker, and community organizer. “We’re all just sitting on the Discord, panicking.”
From the rubble formed Rebble, a team of motivated fans, developers, and ex-employees, rushing to reproduce years of development in a matter of days. Frantic to document critical APIs and development tools before the servers shut off, they grabbed everything they could. The first replacement app store appeared quickly, aptly code-named Panic App Store. Within a few days, they had firmware, core and third-party apps, all the dev tools, and more. And they preserved it all on a wiki, right down to the pinouts.
Meanwhile, Fitbit kept the servers running longer than expected. But the axe would fall in June 2018, and a replacement was needed. Using her inside knowledge of Pebble’s server setup, and painstakingly working through a man-in-the-middle proxy, Berry, the ex-Pebbler, created replacement web services for nearly everything Pebble had provided. With just 16 days until the shutdown, Rebble opened up account sign-ups. When Fitbit finally killed Pebble’s servers, Rebble was ready the next day.
More than 177,000 people have connected their devices to Rebble’s services. Not everything can be free, because the APIs for voice dictation and weather are not cheap — $750,000 per year, if 100,000 people used them, Berry said. And yet, nearly 9,000 people pay yearly subscriptions.
Keeping Pebbles Ticking
Pebble watches were built to hit low price points, with most models selling below $150. They are small devices, made by a company without a lot of manufacturing leverage. Most started out with week-long battery life, but the oldest devices are nearly seven years old now. Pebbles are prone to certain mechanical failures, and despite the success of the Rebble community, some are quite tricky to repair. But Rebble is working on it. The iFixit community has provided some repair guides, and is sourcing batteries and spare parts.
Two Rebblers, Astosia and Tation, have opened a Shapeways store full of 3D-printed Pebble 2 cases and buttons for sale, because the silicone buttons on the original Pebble 2 are breaking down over time. “Removing the screen is stress-inducing,” Astosia said, but other than that, she claims a Pebble 2 case swap is a fairly straightforward transplant of internals. Pebble itself released the 3D printing files for all its watches, so fans have been experimenting.
All Pebbles will eventually die, though. The goal, then, is to create an entirely self-built, reverse-engineered version of the Pebble’s original firmware: RebbleOS. And then find, or maybe hack together, future watch hardware on which it might run. This idea isn’t as crazy as it sounds — Pebble is based on FreeRTOS, an open-source kernel that supports a vast array of hardware. Already, some proofs of concept are running on Rebblers’ workspaces.
Joshua Wise, a lead developer with infectious enthusiasm, thinks a completely Rebble-built watch is more of a project management challenge than a moonshot. The biggest challenge to keeping Rebble running on existing Pebbles is triaging Pebble’s smartphone apps, which occasionally disappear from their respective iOS and Android stores. Beyond that, it’s about freeing up time to experiment and dream of the future — and getting Bluetooth to work reliably, which is “always a pain in the butt.”
Migicovsky, now a partner at Y Combinator, is actively using Rebble, and is proud of Pebble’s longevity. “Every so often I take out my OG Pebble that I grabbed off the assembly line — one of the first ones,” Migicovsky says. “Its manufacture date was December 26, 2012!”
Not many modern, web-connected devices live on for years after their maker goes out of business and shuts down its servers. Fewer still have not only an active repair and support community, but a forward-looking mission. Rebble is a welcoming, open-source, community-minded effort, with a responsible financial model behind it. It’s hard to believe it exists, and feels like some still-raw chunk of 2013 tech optimism that can’t possibly survive into the future.
Except, it might.
This article originally ran on iFixit’s website.
ADVERTISEMENT
