This article was first published in the December 2015 issue of WIRED magazine. Be the first to read WIRED's articles in print before they're posted online, and get your hands on loads of additional content by subscribing online.
General Electric Appliances used to spend years -- and millions -- developing new products. Today, its spin-off startup crowdsources and creates in months, at a fraction of the cost.
The Louisville Fire Department, clad in boots and blue uniforms, has been summoned to the headquarters of FirstBuild, on the campus of the University of Louisville in Kentucky, where an illegal activity is taking place.
The activity in question is the cooking of pizzas which, every few minutes, come bubbling and steaming from a small wall-mounted oven. The last time FirstBuild, a small startup making novel appliances, tried to make pizza, the ensuing smoke triggered an alarm -- wired to the city network -- scrambling fire engines to the address. "You're not allowed to cook unless it's under a hood," says Natarajan Venkatakrishnan, FirstBuild's director (he goes by "Venkat"). "The fire marshal said: 'The next time this happens, I will cite you.'"
There is no hood because FirstBuild is trying to build something quietly revolutionary: an indoor, ventless and doorless pizza oven for the home that mimics the extreme heat of the wood or coal-fired ovens traditionally found in pizzerias -- at a fraction of the cost. A traditional home oven has two "rods" or heating elements -- this one has 20 to 30. Even as the inside temperature reaches 500°C, the outside of the oven is cool. The company is researching an app that would allow users to input what sort of pizza they would like -- from charred Neapolitan to Chicago-style pan -- triggering the proper heat profile.
Since the first episode, engineers added a catalyst -- not unlike the exhaust-scrubbing device found on cars -- that appears to have solved the smoke problem. And instead of issuing citations, the fire and rescue personnel are busily eating slices topped with mozzarella and prosciutto -- a smart bit of diplomacy by FirstBuild. "What's the cost range?" the fire chief asks between mouthfuls. "We might want to take the display model to the firehouse."
The plucky startup troubling convention is hardly a new story, but what is remarkable is that FirstBuild is actually an offshoot of GE Appliances, the behemoth with sales of $6 billion (£4bn) a year. Its factory, Appliance Park, is so massive that its car park has traffic lights to manage the flow of 7,000 workers coming and going. And Venkatakrishnan is no T-shirt-clad incubator graduate but an Appliance Park executive. He shuttles between the "mother ship", as he calls it, and the 50-employee FirstBuild like a buttoned-up headmaster sneaking downtown at night.
FirstBuild is an experiment by parent company GE that combines the power of so-called "open innovation" -- the idea that new products can come from outside a company's own walls -- with the speed of additive and low-volume manufacturing, topped off with the novel promise of crowdfunding, or getting a passionate, self-selecting market of consumers so early to adopt that they are willing to buy something before it has even been built.
GE Appliances, Venkatakrishnan says, is optimised for large-scale efficiency: squeezing out waste from huge production runs. Generally, the company will not even look at a product unless it's making 50,000 units a year. This is good, he says, for churning out products in an established market. But in terms of innovation, he says: "It's like a battleship," taking a long time to start, a long time to stop and a long time to change direction. "Launching a refrigerator takes two and a half years from mind to market," Venkatakrishnan says. Or imagine a new product, like a slushie maker. "It takes six months to a year to do the engineering feasibility," he says. "Then it takes a year and a half for us to go into the system to make a decision. Then you get a team together and scale that up, which takes another year and a half." Four years later, you have your slushie maker. And so do your competitors.
By contrast, consider FirstBuild's most successful product to date, a small machine called Opal that produces "nugget ice". These are small, soft pellets of ice popular at fast-food restaurants and convenience stores in the south of the US. Prized for their cooling ability (the ice has more surface area) they can, curiously, also be comfortably chewed. It was a niche market, with an uncertain consumer demand for making nugget ice at home. "There was a lot of risk," says Kevin Nolan, vice president of technology at GE Appliances and FirstBuild's other key driver. To put out a new product introduction, or NPI, at Appliance Park, he says, is a bet of many millions of dollars. "How much have we spent on Opal so far? Maybe $50,000."
Five months before WIRED's visit, FirstBuild had announced on its website the Nugget Ice Challenge. Community members were invited to present a 2D design rendering for an affordable counter-top machine that could service a family of four. The entries were voted on and, less than a month later, the winning submission -- from designer Ismael Ramos in Guadalajara, Mexico -- was chosen. FirstBuild then began to work on the product.
Meanwhile, community members batted around ideas online about the machine's form and function. In July 2015, Opal went live on Indiegogo -- not as a finished product but a piece of artifice, a rendered dream of nugget ice. "We made fake videos," notes Nolan, taking ice from an industrial machine and adding it to the Opal mock-up. By October, more than $2.6 million had been raised by 6,281 backers. "We asked Indiegogo if we could we collect the money after the product is delivered," Venkatakrishnan says. "They said, 'No, that is now how you test people. You have to take the money up front. That means they believe in it.'" "We had a big a-ha moment," Nolan says. "We didn't have a channel to talk to early adopters." GE's channel was big-box stores such as Home Depot or the mass builders who buy standardised appliances in bulk. If Opal had failed to generate enthusiasm on Indiegogo, says Nolan, "we would have killed it". But it is a success, and a curious one: a product designed over the internet by someone in Mexico (and refined by an ethereal network of enthusiasts); being built not in Appliance Park, but in a former shipping-and-receiving warehouse; marketed and sold not in GE Appliance's traditional retail outlets but on Facebook and Indiegogo. FirstBuild is questioning the very model that has held sway for decades across town. "We are saying we can hit the market if I can make 15 units a week," says Venkatakrishnan says. "Not only that, but with an investment profile that is much lower. If a product takes off, we can transition it to the mother ship," he says. "If it doesn't, it's just a bad idea."
Nugget ice, and FirstBuild's other early projects, began with an off-road vehicle costing $100,000 and modelled on the P-51 Mustang fighter-bomber.
The Rally Fighter -- based on tens of thousands of designs from thousands of participants, all protected by a Creative Commons licence -- is the car that made Local Motors, the company launched in 2009 by ex-marine (and Princeton and Harvard alumnus) Jay Rogers, famous. It was Wikipedia on four wheels: rather than closed teams of designers working in labs and on private test tracks, everything was out in the open, evolving before enthusiasts' eyes.
A few years ago, Rogers was invited to speak at a GE summit by Beth Comstock, the company's CEO of innovation, about the new world of manufacturing augured by companies such as Local Motors. It was just one in a series of initiatives, rippling through the corporate giant that looked to breach the razor-wired, security-badged, NDA-shrouded world of traditional R&D and expose any number of design challenges -- from jet engines to air conditioners -- to crowdsourcing. Innovation was one of CEO Jeffrey Immelt's highest priorities; on his appointment in 2001 he ploughed a billion dollars into R&D. As GE chief learning officer Raghu Krishnamoorthy had described it, the operational excellence prized by legendary former-CEO Jack Welch was less viable in a turbulent new world of small-scale, global disrupters. With Immelt, GE was seeking to overhaul its entire culture, with mantras like "stay lean to go fast" and an emphasis on agile, "horizontal" approaches.
But innovation was not open innovation, at least not at first. In 2010, five years after it launched its Ecomagination green-tech initiative, which featured internal crowdsourcing, GE shifted gears. It opened the project to outside innovation (with Brightidea), putting hundreds of millions of dollars into harvesting ideas -- such as the energy-monitoring startup PlotWatt. Suddenly, you could see Immelt sharing the stage with the founders of startups whose revenues were beneath even rounding errors in GE's $157 billion portfolio. "GE has to be good at ways to take ideas and make them big, reliably and quickly," Immelt noted at one conference. Where every app developer wanted to become the next GE, GE wanted to become like an app developer. "We're saying it's OK to try things earlier, and bring customers in earlier," Comstock said. "You're giving people the freedom to move faster to make more small mistakes."
GE was by no means on the leading edge of open innovation (Procter & Gamble reported that, by 2006, more than 30 per cent of its intellectual property had input from the crowd). But it now seemed to be going all in. Its FastWorks programme, inspired by Eric Ries's concept of the "lean startup", looked to bring the agile processes of software -- start fast, fail fast, learn fast -- to a company whose roots stretch back to Thomas Edison.
It paired with Kaggle, the platform for data-science-prediction contests, on a competition to improve aircraft efficiency. It spent $30 million on Quirky, the website that paired inventors with manufacturers, to probe the burgeoning internet of things market and opened up 20,000 pieces of IP in its patent library to Quirky's community. It put thorny engineering challenges on platforms such as GrabCAD, and found solutions from people who were not only not GE employees, but often came from other disciplines. With FirstBuild, GE was further refining its approach. Where Quirky often sent inventors to Chinese factories, Rogers imagined a platform integrated within GE's vast range of manufacturing assets. "We came in with a very specific methodology," Rogers says. "We wanted them to do open hardware development, and specifically we wanted to co-create with a community all the way through the development of a complex device."
Rogers, in essence, was asked to import Local Motors' approach -- down to the layout of the factory -- to a division at GE. The first thought was oil and gas, but, as he notes, "If you're blowing and going and making a lot of money, there's less desire to try new methods." He eventually came to GE Appliance, where he found "very willing partners". The early success, he says, is striking. "When GE launched its ultra-efficient water heater -- a three-year development effort, and a multi-hundred-million-dollar campaign, it sold 20,000 units the first year," he says. "It looks like Opal is going to push 6,000 units in 30 days, with a three-month development cycle." "It had to be proven that complex systems, even if they were just software, could be made by a crowd of contributors," Rogers says. At FirstBuild, Venkatakrishnan points to new micro-manufacturing technologies. "For $4,000 I have a desktop thing where I can edge circuit boards that didn't exist five years ago. I've got laser cutters where I can cut metal. I've got a pick-and-place machine [used in the assembly of circuit boards]. Five years ago, if I had to make an entity to design and build and sell in low volume, where would I find the equipment?"
The rise of the maker/hacker-space movement, meanwhile, provided the community that FirstBuild taps into. In trying to bring this spirit into the world of corporate appliance development, it had to climb a steep curve of unlearning. Before embarking on a new project, says Venkatakrishnan, a big manufacturer would undertake projections of the costs of tooling and setting up assembly lines. But what if a startup wanted to create, say, a USB-powered refrigerator? "They would go to the market, buy a refrigerator, cut the case and liner and put USB ports in it," he says. "Then they would put Arduino and Raspberry Pi in the back, and you have enabled a USB-powered refrigerator. That's exactly what we did." Indeed, a week earlier, FirstBuild had issued its first royalty cheque for that very same USB refrigerator, called the ChillHub, to a local schoolteacher who, Venkatakrishnan notes, "used to come here and work on Arduino". Everything happens faster. As Ted Lee, who heads Louisville's Office of Civic Innovation, notes, "you come in, you join, you submit an idea, you're into the IP licensing right away."
It doesn't hurt, of course, when the startup has a benefactor with deep pockets, scores of engineers and decades of experience. "It's like having a rich parent," Venkatakrishnan laughs. "Like the kid who has a car and says, 'I am self-sufficient, but my dad helps out with insurance and he pays for petrol once in a while.'" Danielle Blank, who heads LVL1, Louisville's noted maker space, says FirstBuild is "sort of like our hot sister", with glitzier tools.
But she says FirstBuild is not merely a corporation's clumsy attempt at bottling the magic of garage innovation. "Venkat used to hang at LVL1 a lot," she says. "You can see at FirstBuild that they're trying to emulate the community aspect." She says the maker-space mantra is: you come for the tools and stay for the people.
FirstBuild opens its doors to anyone: during WIRED's tour, Nolan pointed to a man at a workstation and asked, "Who's that guy on the SolidWorks?" "He's a Louisville master's student," replied Venkatakrishnan (access to Louisville students is key; indeed, the JB Speed School of Engineering is building, next to FirstBuild, an Institute of Product Realization that, as John Usher, the school's acting dean, explains, will comprise a range of similar microfactories).
During the run-up to Christmas, FirstBuild bristles with people making handmade gifts. "Why do we do that? Why would we let someone making a TV cabinet for their home come here?" Venkatakrishnan asks. Nolan jumps in: "The hope is we're going to find someone who is interested in what we're doing, or we learn something from them." He wants it to be the "Chelsea Hotel" of electronic innovation -- random creative people rubbing together in an eclectic space.
You do not find the marginal players if your doors are closed. "When we started Local Motors, Elon Musk said that engineers don't like to share," says Rogers. "Seven years later he wants to open-source Tesla's system." By not opening up its designs, Rogers argues, a company is "hobbling its ability to move fast. Speed is the new black".
So why wasn't such a seemingly viable model launched earlier? The answer: several large structural changes had to happen. At the end of the 20th century, describes Henry Chesbrough, the University of California business professor who first popularised the concept of open innovation, knowledge workers were growing in number and mobility, making it easier for them to strike out on their own. They were also finding more willing funding thanks to the growth in private-venture capital. There were technical innovations, too: not just the internet, but the rise of Linux and open-source platforms.
Learning fast
FirstBuild's first steps -- and the project that inspired it.
Sitting on a high shelf in the corner of the FirstBuild micro factory sits an industrial-looking device. "Someone had the idea you could dry clothes with a vacuum," Nolan says. "Like a Dyson?" WIRED asks. "No, like outer space." The idea came from a community member in Illinois: suck the air out and the vapour will "flash off". FirstBuild took a crack at it. It didn't work, but, as Nolan says, "we learned a lot on the physics of it."
The idea is that innovation can come from anywhere. Venkatakrishnan says he gets a number of letters each month from prisoners. "They write to us with their ideas," he says. One of FirstBuild's products, a simple but elegant oven drawer that opens towards the user when the door is opened (so the cook does not have to reach deep into the oven), came from a retired GE engineer. As Leadbeater notes, even when engineers create technical innovations, it is often the random consumers who determine their use.
At an LVL1 hackathon, coffee enthusiasts came up with an Arduino-powered tray to use in a convection oven for precise home roasting of coffee beans. That idea, says Venkatakrishnan, could percolate back into the 10,000 Wi-Fi-connected ovens he says GE has shipped. An app could create the proper temperature profiles. "That is not a standard algorithm for an oven," he says. "We would never have imagined you could roast coffee beans in an oven, or that you might prefer my oven because it can do that. It's like peeling an onion: you find people. They come out of the woodwork."
The ultimate hope is that the agile, low-volume techniques tested at FirstBuild might start to trickle back to Appliance Park, and to the larger corporate parent (a notion complicated by the pending sale of its Appliance division to Electrolux, which will also get FirstBuild as part of the deal).
In FirstBuild's factory, Nolan hands me a small plastic clip. When the company began shipping the oven with the slide-out drawer, early users complained of a rattle in the drawer. Engineers responded by 3D-printing a mould and making a clip to stop the rattle. Nolan notes that the clip is starting to chip. As it is programmed into the CNC machine, however, he can simply print another. "I can break a whole lot of $20 tools," he says. Start fast, fail fast, learn fast. As Nolan puts it: "Let's stop arguing and let's make it."
Tom Vanderbilt wrote about Planet Labs in WIRED05.15
This article was originally published by WIRED UK
