Pinterest Reinvents Itself to Prove It’s Really Worth Billions

It’s a week before Pinterest will announce its biggest engineering and design overhaul ever, and Ben Silbermann is really excited. Well, excited in a Ben Silbermann kind of way. He is, after all, an introverted, product-minded CEO. His voice has only one volume: fairly low. “This is me excited, Jessi,” he says. “This is my high energy.”

Nearly three dozen designers and engineers have spent the past several months tearing out the skeleton of Pinterest’s iOS app and rebuilding it. They’ve rethought its design, down to the font. Silbermann helmed the process, bringing his kids by the office on the weekends while he checked in on the team as they raced to transform Pinterest into a sleeker, faster service. He passed up the McDonald’s hash browns that the engineers ordered, instead pitching in by testing out the new build, surfing the app on an antiquated iPad 3 and iPhone 4 to make sure it worked great even on slower devices.

The effort has paid off. Now, as we talk, the bugs are mostly gone. The pins are loading two or three times faster. And the launch is four days ahead of schedule. Who ever heard of an engineering overhaul finishing early?

In many ways, Pinterest is a perfect reflection of Silbermann: It’s a quiet overachiever. The social service offers a clean, efficient interface where people can save images or discover new ones. To the average pinner, the changes Pinterest is announcing today will seem subtle, even unnoticeable. But subtle is its own strategy. Beneath the surface, the company has made significant changes. Silbermann believes they can help transform the digital pinboard he and cofounders Evan Sharp and Paul Sciarra invented into the dominant global visual search engine. He thinks they will drive new people to try Pinterest and spend more time on it. “We’re trying to build a catalog of ideas for the entire world,” he says. “It’s only as good as the diversity of ideas inside it.”

Pinterest is already wildly popular—in America. Its success here happened suddenly. Pinterest was founded six years ago, but no one really noticed it at first. Then in 2012, it exploded into the mainstream, propelled by the mass adoration of mostly women, mostly in the middle of the country. There was Reese Witherspoon gushing about it to Conan O’Brien, and Mitt Romney’s wife organizing her family photos on it. In March alone that year, the number of people using Pinterest jumped 52 percent, and venture capitalists began chasing after Silbermann with their checkbooks open. Four years later, Pinterest continues to grow; 87 million people use the service every month in the United States, according to ComScore, up 61 percent from a year earlier. That enthusiasm compelled US marketers to spend a reported $100 million on advertising in 2015. Though Pinterest President Tim Kendall won’t confirm if that figure is correct, he says Pinterest quintupled its ad sales last year. Kendall says that a lot of people come to Pinterest because they’re thinking about buying something. “That’s not consistently the case with really any other service,” Kendall says. “They’re receptive to brands, and they’re even looking for them.”

But advertising is still novel for Pinterest, which didn’t have any revenue model at all until 2014. For it to fulfill its promise as a global visual search engine—and a great place for brand advertising—it will need to transcend its initial American female audience and find a way to appeal to everyone. In the US, Pinterest hasn’t been able to gain popularity among men at the same rate as women; 75 percent of the people using it are women, according to ComScore. And for all of the enthusiasm trained upon it domestically, Pinterest has had trouble translating itself internationally. Only 45 percent of the people who use it are abroad, according to Kendall, much less than Facebook or Twitter had at this point in their development. “Pinterest was founded for and aimed at Midwestern moms, and that’s been a strong base for them, but it’s not necessarily true that mindset occurs throughout the world,” says eMarketer principal analyst Debra Aho Williamson. “It remains to be seen whether the idea of pinning and sharing is relevant in many markets.”

Now, as the company reaches the social media equivalent of middle age, its success is not at all assured. It’s a growing business, sure, but it has yet to prove that it is, as venture capitalists have anticipated it would be, an $11 billion business. When Pinterest raised money at that valuation last year, the company said it would generate $2.8 billion in ad sales by 2018, according to funding documents that were leaked to TechCrunch last fall; to do this, Pinterest will need to keep up the robust sales growth. This will require more people to sign on to the service and use it more frequently for more things.

Which is why Silbermann has been coming in on the weekend. The changes the company is about to unleash will determine whether Pinterest can earn its deca-corn valuation, paving the way for a successful initial public offering, or whether it will remain forever niche–confined to the female crafters who first fell for it.

Taste and Tech

There are two reasons Pinterest hasn’t grown as fast Internationally as Facebook and Twitter did early on. First, it’s a different type of service that relies on showing people images that resonate with them, rather than connecting them to their friends. “Messaging services are about finding the right people in your phonebook and giving you the tools to communicate with them,” Silbermann says. “Pinterest is about matching you with the right ideas.” Second, Pinterest just flat-out needs to work better, particularly on the low-end devices often found in different countries.

The first issue concerns taste—a quality that is, as Silbermann says, as much art as science. Amassing and conveying ideas, it turns out, requires a nuanced understanding, and that understanding of themes and concepts is different in nearly every country. What’s more, even when you know this, making the right choices about which images to show pinners isn’t always intuitive. For example: If you’re a foodie living in France, you really want to see French foods on Pinterest, Silbermann says. But if you’re a French fashionista, then you want to see global styles, not only French ones. “It’s not as easy as just showing you a bunch of things in French,” he says.

'We’re trying to build a catalog of ideas for the entire world. It’s only as good as the diversity of ideas inside it.' Pinterest CEO Ben Silbermann

These subtle distinctions extend to the design of the interface as well. For example, typefaces that are popular in the US may not work as well other places. Last year when an employee from Pinterest’s Tokyo office relocated to its San Francisco headquarters, she explained that for Japanese audiences, the Pinterest homepage inspired that feeling you get when someone tries to overcome a language barrier by talking louder.

Of course none of this matters if Pinterest can’t address its second issue: it must work reliably and quickly for everyone. “Speed is one of those features people don’t ask for but really appreciate,” Silbermann says. “It makes everything else you build a lot better.” And it can’t crash or lag. Pinterest was initially engineered to work best on new devices like the latest iPhone, but if it’s going to catch on other places, it also has to work just as well on that old iPad 3 or even a Motorola Moto G.

Last summer, Silbermann decided it was time to tackle these problems. He dispatched researchers to better understand the cultural challenges. He sent them to the five markets in which Pinterest has concentrated its initial efforts by setting up offices: the United Kingdom, France, Germany, Brazil and Japan. As a result of their work, Pinterest has begun tweaking its algorithm so that pinners are shown culturally relevant images in the categories that matter to them. He also sent senior executives to visit regularly. And he began bringing people from other offices to headquarters more frequently.

As important, Silbermann realized he needed to refashion Pinterest so it was irresistible to those audiences—faster and sleeker, with a new visual interface. He felt Pinterest needed a design approach that could appeal to a more diverse audience, an approach so minimalist it would train the people who use it to focus on the images they pin rather than on the icons on the page. That way, Pinterest could use the same page design to appeal to Korean men or British woman—they’d be attracted by the images they discovered and saved, not turned off by the stuff around them.

Additionally, the company needed a new, more standardized approach to engineering—one that would allow the service to run faster but would also help the engineers working on it in the future to do their jobs faster so the company could release updates more quickly. It had, by this point, grown just enough that it no longer felt small. Now there were 300 engineers, and if things continued to go well, that number would continue to rise. As the company scaled upward, engineers and designers had to work harder to communicate with each other about the products they were building. Pinterest had started to look slightly different in the Web, Android, and iOS versions. A new engineering framework could help Pinterest work and look its best, no matter where a Pinner was using it.

Solving Pinners’ Problems

If Pinterest’s new release had spirit animals, they’d take the form of Scott Goodson and Tom Watson. They are very different: Watson is a tall (6’9”) designer who presides over a new design systems group from a minuscule Portland office that used to be a warehouse for Model Ts. Goodson is a Bay Area coder who graduated high school at 14, college at 18, and launched his career at Apple after he met Steve Jobs through a chance encounter (“My mom met a guy who knew him, and asked him to call me. I was at my grandma’s house when he called out of the blue.”). Six months before Pinterest would need to pluck every last iOS engineer in the company off their projects to work on the new update, Goodson and Watson reconsidered how Pinterest should approach engineering and design.

Goodson took on the engineering challenge. He had come from Facebook, where he built Paper, an app that flopped among Facebookers but impressed engineers. Now, as head of core experience at Pinterest, Goodson’s job is to provide an engineering framework for Android, iOS and mobile web developers. He’s in charge of providing engineers the tools and directions they need to improve Pinterest. He also makes sure that the way people used Pinterest is consistent, whether they’re accessing it from a PC in Belgium or an iPhone 5 in Brazil. Along with a team of four other engineers, he spent the summer and early fall last year evaluating software options.

By early November, they had landed on a strategy for the first platform they planned to tackle, iOS. They chose to embrace the open source AsyncDisplayKit framework, which Goodson had helped to write while at Facebook. This software allows engineers to execute code across multiple “cores”—the individual processing units inside iPhones. “iOS was designed a long time ago for single-core devices and placed a high priority on the simplicity of developing software,” Goodson explains.

With AsyncDislplayKit, developers would solve some big problems for pinners, who often swipe through multiple images in every direction on their screens in quick succession. By making use of multiple processors at the same time, Pinterest can prep upcoming photos at the same time as it renders the one you are browsing. You no longer have to wait even a split second for a new photo to materialize. Second, the software enables intelligent preloading— Pinterest can sense which images you are most likely to scroll to next, and load only those. Goodson says this will increase speed by a factor of at least two and increase the rate at which the app responds when winners press buttons by a factor of ten, while also putting less strain on a device. That means Pinterest will also work better on less powerful devices (In the process, Goodson’s team developed the software’s first official new version of the open source developer’s kit, AsyncDisplayKit 2.0, which will launch within the next few months.)

Around the time that Goodson was getting his ID badge and learning where the coffee was at in Pinterest headquarters, the company’s creative director Andreas Pihlström and cofounder Sharp flew up to Portland to see Watson, who was the lead product designer. The trio had two objectives: They wanted the design to be more subtle so that people would notice the images being pinned instead of Pinterest’s logo and icons. They hoped this would broaden its appeal—to more men as well as to people outside of the US. Also, they wanted a flexible page design that could work and look the same on any device, no matter the screen size. “The more we standardized, the more efficient we could be as a group,” Watson says.

Watson and his crew drafted a system that relied on a series of columns stretched grid-like across the screen to adapt the design across devices. An iPhone 6, for example, might have two columns of images while an iPad might have four, but the images would look consistent—pinners would immediately recognize they were on Pinterest. These changes aren’t likely to be especially noticeable, even to frequent users—more of a subtle touch-up that will make Pinterest look and feel the same on an Android or an iPhone, and will draw pinners to look first at the pins, instead of the icons. The new method will enable future teams to draw up plans for all devices at the same time, rather than having to design separately for every different screen. Meanwhile, graphic artist Susan Kare, who had made her name conceiving fonts and icons for Apple early on and joined Pinterest last summer, refreshed several of Pinterest’s icons and designed new ones.

Watson also developed a new approach to typefaces. In the West, we see a lot of Helvetica. But not every language looks good in the bold classic font. In particular, Japanese characters are better rendered in Hiragino, which has a lighter imprint, more similar to a brushstroke. Watson had to choose fonts that would allow the languages to complement each other visually on a page. “If we don’t choose the right font, and get the scaling to match exactly right, it will feel like you’re not culturally appropriate to the country,” he says. In Roman languages, Pinterest will use Helvetica Neue; when that doesn’t render well, it’ll rely on other fonts, like Hiragino for Japanese.

By late fall, Watson had assembled Pinterest’s design protocols on a Google doc available to the entire company. Goodson had built out the framework that would allow Pinterest engineers to rebuild the entire app in this new framework.  It was crunch time.

A Better Reflection

In January, Pinterest moved most of its engineers to a new, larger office. The new digs, composed mostly of concrete and plywood, were built to look a lot like the service’s app: clean and spare. And in one central location, a product manager named Adam Barton, who had been tasked with running the code and design overhaul, roped off a line of desks. He pulled all 20 of the company’s iOS engineers from their projects. They joined Goodson, Watson, and a few other designers and engineers to embark on the overhaul.

Each morning, the team met for 30 minutes, dialing in remote members from Portland and New York. “I worked for 42 days straight,” says Barton, whose early experience as a contract negotiator for Google had given him the skills to keep everyone on task. One of his most important jobs: He was the guy who went to McDonald’s every Saturday to pick up the hash browns. The fast-food outlet requires a manager’s approval for anyone buying more than $50 worth, according to Barton, so he’d have to ask the honcho on duty to sign off on the order. Meanwhile, everyone at Pinterest had access to the new software and was encouraged to test it. “I reported bugs multiple times each day,” Silbermann says.

By the time I sat down with Silbermann and the other engineers and designers, the feat was largely done. Goodson was excited about a team trip to Vegas to see Calvin Harris. There was also talk of a camping trip, but someone still had to plan it. Silbermann was already looking beyond; the team will take on the Android platform next. In anticipation of this, he’s swapped out his iPhone for a Nexus and a lower end Moto G. And in June, he’ll spend most of the month on the road, visiting international offices and a few other places.

Of course, if Pinterest’s overhaul is a success, pinners won’t really notice it. The ones using it on an iPad 3 may see that unlike most of their other newer apps, Pinterest isn’t buggy and won’t spontaneously shut down. Those of us who open it on our iPhones during lengthy conference calls to surf through images of tree house resorts or kitchen renovations (guilty, yes) may find it’s more responsive and quicker.

But really, the point is to make Pinterest much bigger by making it less of a big deal—less flashy, easier to use, and more of a utility. If more people spend more time browsing pins, it stands to reason advertisers will be drawn to spend more money on the site, too. Earlier this month, Pinterest announced plans to start selling ads in the UK.

Companies often reflect the personalities of their founders. Engineers and product managers alike told me that as a boss, Silbermann is the kind of guy who listens more than he talks, and then reflects your best ideas and lets you be in charge of them. He wants Pinterest to be the kind of service that does the same thing.

• gallery
• Pinterest
• Social Media
• Startups
• Unicorns