YOUTUBE’S OFFICES IN San Bruno, California are spectacular. Standing desks, couches, nap pods. Kitchens around every corner stocked with free food. I’m told there’s even a pool. It’s all par for the Silicon Valley course, really.

However, one thing inside The House Cat Videos Built feels out of place: one of the Wi-Fi networks crawls at mind-numbing speed. When John Harding, YouTube’s VP of engineering, takes out a tester phones and taps the YouTube icon, the app takes more than a minute to load. As the thumbnails and icons render, Harding pulls out another phone, connected to the same network. He taps the home screen to open something called “Video App.” It opens in seconds, a scrolling list of video thumbnails on a white background.

This app isn’t called “Video App” anymore. Starting today, it’s called YouTube Go, and it represents more than a year of work to rethink YouTube for a new kind of user. The so-called “next billion” Internet users are coming online, many of them in India, Indonesia, Brazil, and China. They aren’t like the users who came before. They have different devices, different connectivity, different social norms, different ideas about what the Internet is.

Spend enough time talking to Silicon Valley types and you’ll hear them say “we made something we’d want to use.” That’s why laundry startups proliferate. But what do you do when you’re making something for everyone else? If you’re YouTube, you send your engineers, researchers, and designers to India, Indonesia, Singapore, and beyond to find out how the next billion will use YouTube, and how they might change the service forever.

The Human Network
India is one of many countries YouTube is targeting with Go, and a fairly representative one: It has more than a billion people occupying every socio-economic class and living in every imaginable setting. India also is uniquely enticing for any company looking for growth markets, says Forrester research director Ashutosh Sharma. “It’s a country of 1.25 or 1.3 billion, and only 200-250 million people have Internet connectivity,” he says. It’s also more open and accessible than, say, China. “There’s a growing middle class,” Sharma says, “and people crave foreign brands. It’s going to be even more open in coming years.”

Google has offices in India and the YouTube team has partners and engineers there, but the company decided early on to send the people making the new app to meet their new users. Nibha Jain, YouTube’s research lead for the Next Billion Users group, says her team was learning as soon as it got off the plane. “We realized,” she says, “that when we got there, we would look into our phones to see, oh, where do we go? But what you need to do when you’re in India is roll down the window of your car and ask the guy on the street.” The research team found that every experience, even with technology, is inherently social. The next billion team started calling this the Human Information Network. “I have a phone, I have an individual experience,” Jain says. “When we move to these societies, it’s very social, it’s very integrated. How do we respond to that?”

The team spent months talking to individual users, or sitting down with small groups to figure out how they use apps together. There was the guy in a one-room house who couldn’t wait to show Johanna Wright, YouTube’s VP of product management, his favorite WWE wrestling videos but couldn’t get them to load fast enough. Or the guy who got videos from a brother-in-law who worked in a bakery with decent Wi-Fi. He’d download videos at work, then use the Shareit app to send them to his friends and family. The YouTubers saw more languages than they expected, more varied use cases for the Internet, and an entirely different way of sharing content. People didn’t discover videos on Twitter, they found them by swapping SD cards or sharing directly with their friends.

Ultimately, the next billion team came up with a handful of core principles for a new YouTube experience. They focused on making the app work even on even the cheapest phones, enabling sharing between people, localizing the app as much as possible, and maximizing data-friendliness. That last one was the hardest, because it meant the app had to work offline far beyond just letting you download videos (though you can of course download videos). Data is expensive, and connectivity can be hard to come by. Many people keep data off, except when they need it. “I can honestly say, after being in India,” says Arvind Srinivasan, the team’s engineering director, “that the 2G in our office is fantastic.” For YouTube, that meant the app had to feel alive even without a connection. YouTube Go compresses and caches thumbnails for videos so you can poke around and see what’s there. You can see, share, and watch videos without ever pinging a cellphone tower.

Because it’s designed to operate offline, YouTube Go’s sharing also is an almost entirely local experience. Once you’ve downloaded a video, it sits on your phone like any other file. In the YouTube Go app, you go to a sharing menu, and it shows you who’s around waiting to receive a video. Tap their name and the video and the app sends it over a local Wi-Fi network. The app does a light check-in with the YouTube server to credit the creator (and ensure the video’s not deleted), and unlocks the video on the new device. In theory, one person could download a long video and share it with everyone on Earth, one by one, without ever having to download the entire file again.

YouTube Go also includes a preview feature that lets you see a few frames of a video before deciding to spend data to watch it. “You really want to know a lot more about it than the thumbnail,” Akkad says, “before you invest 200MB both on your phone and from a data perspective.” It only uses six-second ads, which don’t cost much data. And they compress the ever-living hell out of every video, to make downloads as small as possible. Before you download anything, the app tells you the size of the file, so you can decide if it’s worth the money.

Coming Online

The relentless expansion of Silicon Valley into the rest of the world’s business has been, and will continue to be, a messy affair. Just ask Facebook. Or Uber. Or, uh, Google. Everyone is still pushing at the borders, and the smart ones realize that the users they want are nothing like the users they have. That is forcing companies to re-think what they make and how they operate, and that will change the Internet for the rest of us.

This might help explain why YouTube goes out of its way to say YouTube Go is the very start of something. The new app doesn’t have subscriptions, or trending modules, or comments, or many other things you’d expect. There are no new features for creators, either. Some of that is intentional. “Giving them something that YouTube has iterated on with its users over ten years, all at once? That’s a little daunting,” Akkad says. “So we want to stage how we introduce people to different functionality.” YouTube Go is launching with a few thousand users, then expanding in stages before YouTube makes it available to everyone early next year.

The phased rollout is also a chance for YouTube to continue learning what works and what doesn’t. The team is keenly aware that it is building a product for people far removed from its comfy San Bruno offices, and that few of its ideas about YouTube are useful there. Before starting work on Go, a group of engineers went through an internal Google bootcamp designed to open their minds to new ideas. Srinivasan calls it “almost a spiritual experience.” Everyone constantly reminds me, and seemingly themselves, that their job is not to build what they think would be cool, but what their user wants and needs.

If YouTube Go is successful, it will change every part of YouTube. It’s just a numbers thing: If a billion-plus people start experiencing the service through local sharing and downloads, they won’t give that up to switch to the “real YouTube.” And whether it’s video previews or video compression or the overall lightness of the app, many of the changes will appeal to everyone. YouTube Offline launched in India two years ago, and is now part of the product around the world.

A billion people are about to tell YouTube what they want, and they’ll do the same with every tech company you know. Those companies would be crazy not to listen.