When I Wore Google And Levi’s “Smart” Jacket

One evening last week, I got on my bike as I normally would. But this was no ordinary ride. Instead of pulling out my phone to get directions from Google Maps or figuring out what music was playing by glancing at my screen, I was hearing directions and music in my ear, and controlling it all by simply brushing or tapping on the cuff on my connected jean jacket–the product of a yearlong collaboration between Levi’s and Google’s Jacquard, an experimental project to design smart garments that act as interfaces. I felt like I was living a future that interaction designers–not to mention the movie Her–have prophesied for years. A world where I didn’t need to rely solely on my phone to access information.

But suddenly, Jacquard told me to turn the wrong way down a one-way street with no clue as to why, or how to correct the error. When I stopped, the app didn’t have a map to help me orient myself. So I had to open Google Maps and figure out where to go, the “old fashioned” way. As I tested the jacket, there were moments like this one where I felt like I was at my wit’s end without a phone screen. There were also times I preferred using the cuff, for instance as a way to answer calls or switch songs.

Jacquard embodies almost a decade of hype around the future of interfaces: Finally, a voice-powered, screen-free interface that liberates us from our phones and lets us focus on what’s around us! But it also embodies some of the skepticism around this vision of technology. The newfound ability to ditch the screen doesn’t mean we want to be without it all the time. The reality of how we use our devices is more nuanced.

To learn more about the jacket, I spoke with designers and Levi’s, Google, and Ideo about the long process of bringing it to consumers–and how experimental garments could change the garment industry and our relationship with tech itself.


At Google I/O 2015, Ivan Poupyrev–one of this generation’s most influential interaction designers and futurists–announced Project Jacquard, a bold project to develop textiles that could replace touch screens as a smartphone interface. This week, after more than two years of collaboration with Levi’s, the first Jacquard product arrived: A $350 jacket (technically called the “Levi’s Commuter Trucker Jacket with Jacquard by Google”) that hit levis.com and is now available for sale in select Levi’s stores.

“We didn’t seek out tech,” Paul Dillinger, vice president of global product innovation, says about the Google collaboration. To Levi’s, Project Jacquard was always about delivering a better product to its consumers, rather than being the first to market with an innovation. “We wanted to make a better ride for the cyclist, not make a technical object. It just so happened that this facilitates a better ride by moving some of your interface away from your phone and onto a jacket.”

Silicon Valley has a reputation designing solutions for problems that aren’t really problems. Google and Levi’s wanted to make sure the jacket would truly be useful, and for Dillinger and Levi’s commuter-collection consumers, that meant making mobile phone use more convenient–thereby making bike rides safer. “When we’re talking about being on a bike, safety and convenience are pretty much the same thing,” Dillinger says, pointing out that answering calls, reading texts, and managing navigation are routine activities for bikers. “I’m not pegging this as a safety garment, I’m not an advocate of even having your phone on when you’re on a bike. But this is taking behaviors that are already happening and making them easier and thereby more safe.”

The Jacquard-enabled jacket looks virtually identical to Levi’s analog commuter version you can find in every store. But much of it had to be redesigned from scratch: From the Jacquard thread itself, to how the threads detect and read motion, to how the yarn connects to the embedded chip and antenna that then communicate with your phone. And that’s just the hardware. It was also important to both Levi’s and Google that the coat didn’t look like a stage prop from a sci-fi movie. “Except for all those ways it’s different, it has to be just a jacket,” Dillinger says.

What’s more, Jacquard–which is both a physical product and a software platform–had to work within existing supply chains. Textile manufacturing is a heavy-duty process. Yarns are passed through heavy machinery and looms and the Jacquard thread needed to be resilient through all of this. Levi’s denim manufacturing process–which involves exposing the textile to water and flame, passing it through heated presses and rollers, and violently torquing the fabric by machine–is especially brutal. “You can’t just take random materials and expect textile companies to use them,” Poupyrev says. “They have to look, feel and behave just like a normal yarn.”

So Google and Levi’s had to develop Jacquard’s “hardware” at the smallest scale possible: Individual connected threads that can be woven into any textile. The final Jacquard thread–which was developed by a group of textile designers based in Japan and took more than two years to finalize–is composed of a super-durable metal alloy over-braided with traditional fibers, like nylon, cotton, or polyester. The idea is that the thread itself can be as invisible or visible as the end manufacturer wants it to be.

Once Levi’s and Google were confident in their yarn, they had to figure out how to weave it into a jacket. “We thought, what would be the right way to express the functionality in the textile, but staying within the design language of textiles?” Poupyrev recalls.

First, the Google team made the yarns white so that users could see where the interactive section the garment was located–that way, user wouldn’t need instructions about where to swipe and brush the interface. The white threads proved to be too stark a contrast with the denim, and Levi’s didn’t like the aesthetic. Google went back to the drawing board, and made a version that blended seamlessly with the jacket’s normal fabric. The only problem? Users were confused about where to interact with it. In the end, they landed on a solution that raised the Jacquard threads slightly. Inspired by a common weaving mistake called a “missed pick” that gives fabric a slightly bumpy surface, the effect is subtle. The Jacquard fabric is the exact same color as the jacket but looks like faint lines, and feels almost like tiny ridges.

Despite all the complex technical challenges of figuring out the jacket’s hardware, its usability was even more complex.

Ideo’s hardware prototypes for the jacket’s snap tag, from a breadboard to test light indicator patterns to studies of the tag’s form and material. [Photo: courtesy Ideo]


Levi’s and Google turned to the design consultancy Ideo to imagine the jacket’s user experience and its language of touch-based interactions–a challenge, since nonverbal communication is nuanced and Levi’s and Google wanted to make sure the touch gestures used to “talk” to Jacquard wouldn’t carry different meanings to different people across cultures. The Jacquard Jacket had to be universal.

“If you brush your shoulder, it could mean, ‘I’m getting real with this guy,'” Poupyrev says. “If you’re rolling up your cuffs, it could mean you’ve come up with a great invention, like a scientist.”

Gender was a significant part of the equation, too, since men and women interact with their clothing differently. For example, research by Google and Levi’s showed that women don’t put their hands in jacket pockets as much as men. Then there was the element of accessibility for riders. The interface needed to be easy to reach from a bike. Lastly, it had to be easy to stitch into the jacket itself from a manufacturing perspective.

For all of these reasons, the team landed on the cuff as the spot to ground the jacket’s interface. Users interact with Jacquard by touching the textile interface on one cuff of their jacket, where a microprocessor reads these gestures and relays them to a mobile app via Bluetooth.

Initially, Levi’s and Google wanted the jacket’s electronics to be no larger than a button, and early on Ideo explored solutions like a fabric strap that riffed on trench coat and anorak cuffs. Those designs weren’t feasible given all the components involved–namely a battery, an antenna, and microprocessor–so instead, all of that technology is embedded in a component called a “snap tag” that affixes to the cuff and encloses the electronics in flexible, soft-touch plastic. It’s a thermometer-shaped band with a USB connector on one side and what looks like a typical Levi’s button on the other (it’s actually another interface that vibrates and glows when you receive calls or texts). It securely snaps to the jacket using pogo pins.

“[The snap tag] is stealthy, quite humble, and it gets out of the way and enables the jacket to do what it does best, which is a being good piece of fashion consumers love,” Will Carey, the former executive design director at Ideo and now head of design at Logitech, says. “But it also enables this new behavior and new interaction paradigm to exist, which is the gesture touch zone on the cuff.”

It’s the most overtly tech component of the jacket, and the designers at Ideo made over 150 prototypes before arriving at the final form. “Consumer electronics are typically rigid and atomically firm–batteries, antennae, etc.–and generally sits in slabs,” Carey says. “When you want to put that onto a garment, it’s an interesting technical challenge.”

Users communicate with the jacket through four gestures: Brushing the interactive patch downward, brushing it upward, double tapping it, and covering it. “We wanted to create a ritual as you go about your daily commute, whether that’s walking or cycling,” says Carey.

You assign functionality to these gestures using the Jacquard app. For instance, you can program controls that include playing or pausing music, skipping songs, or finding out what song is playing. The app’s voice navigation can alert you of your ETA at your final destination, and tell you your next direction. You can decide if you want the snap tag’s light to alert you of incoming call or texts, or you can program it to only illuminate when specific contacts get in touch. Meanwhile, a “keep track” feature lets you keep count of anything you choose, recording a specific gesture you make repeatedly during a ride.

The idea is that you assign gestures to actions based on what is personally most important to you. “If the user can be given the choice as to how to determine the meanings of each gesture, we have something that will provide emotional durability and resonance with consumers versus the imposition of a new interface modality,” Dillinger says.


For my first bike ride wearing Jacquard, I journeyed from my apartment in Crown Heights on a half-hour ride to the east side of Greenpoint. It’s an area I don’t visit frequently–the perfect test drive, since I would need to rely on Jacquard’s navigation.

I opened the app and programmed the gestures for what I thought I would need for the evening. I told the app to skip to the next song when I double-tapped the cuff, and asked it to give me my next turn when I brushed out. For “brush in,” I decided to test out the “keep track” feature, asking Jacquard to keep track of how many drinks I ordered (I was really looking for an excuse to try this feature). Next, I typed in my destination’s address and hit “save.” I was hoping to see my route mapped out for me–it’s helpful to have a basic mental map, even if a voice is helping you navigate–but the app doesn’t offer any visual cues. I decided to quickly pull up the address on Google Maps so I could have a rough idea of where I needed to go. I told myself I wouldn’t look at Google Maps again all night, since it wasn’t in the spirit of moving beyond a screen interface.

I didn’t last very long. With an ear bud in one ear I pedaled off and brushed my sleeve. “Getting next direction,” the Jacquard app said in my ear. “Turn left on Bedford Avenue.” I turned left, then brushed my sleeve to get the next direction: “Something’s wrong,” I heard. “I can’t get directions.”


I brushed my sleeve again, and the app successfully told me the next turn. The next few times I retrieved directions, it worked fine–since they were standard intersections. But as I quickly discovered, the jacket’s navigation is buggy. It runs into trouble when streets aren’t straight and where they don’t meet at 90-degree angles–a frustrating problem in any city that’s not a perfect grid. Once, it instructed me to make a left on a one-way street heading in the opposite direction. This is not correct, or safe. I brushed two more times and it told me the same thing. Luckily, I had crossed that intersection before using Google maps, which usually tells me “slight left to stay on Willoughby Avenue.”

Jacquard’s navigation didn’t distinguish soft turns when I used it. Good thing I knew when not to listen to it. After every brush, Jacquard told me, “Getting next direction,” then lagged slightly before offering the actual direction. Instead of information coming to me at regular, intuitive intervals–like Google or Apple Maps’ user experience–I had to constantly go to the information. The jacket’s navigation requires constant attention at every turn, unlike Google Maps’ set-it-and-forget-it approach. At first this was fun—Whoa! My jacket is talking to my phone! My phone is talking to me!—but after the novelty wore off, it was annoying.

The more attention I paid to navigation, the less attention I gave to my surroundings. In a city where many pedestrians cross the street mid-block, where drivers cut bike riders off, where car passengers open their doors without looking for oncoming bikers, where cars obstruct bike lanes, where other cyclists ride the wrong way down a bike lane, and where wheelchairs use the bike lane–all of which happened on my ride–continually retrieving directions was another thing to worry about, increasing my cognitive load.

Another side effect of a voice interface that doesn’t provide enough detail? I definitely rode slower than I normally do. Jacquard doesn’t tell you if your turn is 1 block away or 10 blocks away. To make sure I didn’t miss some turns, I was probably riding at about 65% of my normal speed (which is probably safer, come to think of it, though likely not the designers’ intent). Most navigation apps reroute you if you make a wrong turn. Jacquard does, but not in an intuitive way. At one juncture, the app told me to, “turn right towards Lorimer,” but it didn’t tell me on which street to turn. I was already on a designated bike route (and had ridden on this street before, so knew I was heading in the right direction) and didn’t trust the app’s suggestion based on my earlier difficulties.

It was 9:45 p.m. I decided I needed a smooth ride home. I pulled up Google Maps and charted the rest of my way back. A few minutes later, I got a call. When you receive calls, Jacquard’s “brush in” gesture automatically becomes “pick up,” while the “brush out” gesture silences it–no matter what you pre-programmed those gestures to do. The same thing occurs for the first 10 seconds after you receive a text; brush in and Jacquard reads it to you. The metaphor here is either bringing information in, or pushing it away. The interaction worked perfectly and was more convenient, for me, than reaching for the remote on my ear buds.

The gestural interface itself was easy to master. I had to make sure I rubbed the sleeve with enough pressure–grazes won’t register–and tapped it with purpose–not the typical light touches I’m used to using on my iPhone’s screen or laptop’s track pad. Once, the jacket registered a double tap as a “brush in” when my second tap actually had a slight upward lilt. Brushing across the full two-inch length of connected Jacquard threads yielded the most consistency. A few times, my gestures were short, and the app didn’t pick up the motion–all part of the learning curve. Ten minutes into my ride, the gestures were second nature. The brushing, tapping, and holding felt intuitive and comfortable. The snap tag attached securely and tucked into place.

My conclusion? I prefer Google Maps’ bike navigation experience over Jacquard’s. I missed the level of detail in Maps’ directions. Perhaps later iterations of Jacquard will integrate with Google’s other, more advanced apps. The jacket as a textile interface worked, which is a huge technical accomplishment. It’s also worth noting that Jacquard offers pedestrian navigation; at walking speed, it was easier for me to keep an eye on street signs. I liked walking with Jacquard, and welcomed the change from keeping my nose glued to my phone screen. But as a cyclist, the navigation wasn’t useful enough for me to justify using it over what I already had.

Surely some people will like being early adopters of a product that’s the first of its kind and already part of design history. Some might prefer to retrieve directions step-by-step while others might prefer to use the navigation apps they already have. Some might decide that not having to look at a screen is worth an extra $200. But if I was looking for a denim jacket to use on my commutes, I would spend my money on an analog version and simply pull over if there was a call I needed to answer or a text I needed to reply to.

Levi’s and Google say that this is just the beginning of the Project Jacquard Trucker–and Project Jacquard itself. Depending on the consumer response, they might update the software to add more functionality. My biggest ask? Refine the navigation to be closer to what cyclists already receive through existing systems. There were moments I preferred to use Jacquard and others that had me running back to my screen. What the jacket gave me above all was choice, and the agency to figure out how best to use the technology in my pocket.

Therein lies the challenge–and potential–for designers working with voice, screen, and gestural interfaces: figuring out which interactions are most useful for specific situations and applying them in a nuanced way, rather than forcing us to adopt one single paradigm.

“Interfaces matter when they fulfill user need,” Poupyrev tells me in an email follow-up. “Voice interfaces are resolving a very important friction point: meaningful interaction with technology when direct physical contact is inconvenient or impossible, like when a device is far away or when our hands are occupied, such as when we are cooking or driving. Touch interaction and tactile feedback, similarly, fulfill user needs when using voice is inconvenient or impossible, like when you are at a meeting or in a public space and do not want to attract attention, or when you are in a noisy environment. In situations when the device is immediately accessible, like you wear it, an act of a simple touch to your jacket’s cuff could be easier, faster, and more satisfactory for customers then pronouncing a voice command. I believe both voice and touch will have important places in our future devices.”

Poupyrev reiterates that Jacquard is really a computing platform for smart apparel. He envisions it supporting multiple modes of interaction in the future, like being able to talk to your jacket or your gesture in midair (the snap tag is already equipped with an accelerometer to detect motion). That future is still far off, however.

Perhaps the biggest knock against the Project Jacquard Trucker–and the dream of clothes as technology–is the simplest. While the jacket itself will likely last years, the connectivity might not. According to Google’s care instructions, it’s only guaranteed to withstand 10 washes–and can’t go into the dryer or be dry cleaned. Someone biking intermittently might be able to hold off washing the jacket for months at a time, but that, paired with the challenging interface, reveals the connected jacket for what it is: an admirable proof of concept, a major step in the history of design, a fun garment that’s sure to offer a party trick or two, but not quite a daily-go to that we can treat as just another piece in our wardrobe.