17
Failure Isn’t
Facebook doesn’t always hit it out of the park, and that’s a good thing
An appetite for wildly irrational levels of risk and celebration of failure are two of the most central aspects of Silicon Valley and its lore. This is not a place for the timid. It is, instead, a crucible of creative destruction. New dreams are built—literally—on top of old ones that failed to keep up. Just see Google’s campus at the site of the now defunct graphics workstation maker SGI and Facebook’s constantly expanding campus on the site of acquired-and-forgotten-former “dot in dotcom” SUN.
A lack of fear of failure is what has changed Silicon Valley itself to no longer revolve primarily around its namesake “Silicon” and practically ensures that it will continue to evolve to meet the needs and opportunities of the future the same way it has since its origin among fruit orchards in the 1950s.
Striking out forward with no certainty of outcome is Silicon Valley’s reason for being.
Risk is required and failure inherent, but the two have to be managed differently by the valley’s two primary types of consumer businesses.
Hardware makers like Apple and even Tesla have to tread more cautiously. Cycle times to build new product categories—and/or recover from mistakes in previous products—are long, investments in the large teams and complex manufacturing required are high and expectations from paying consumers unforgiving. You inherently get fewer at-bats and are forced to swing in a considered way to clear the fences as often as possible.
Software, on the other hand, now almost exclusively in the form of free-to-use Internet-based services like Google and Facebook, cannot afford to act with anything but wild abandon. Risks are comparatively negligible—prototypes can be built quickly with small teams and little variable costs—making competition, even from smaller players, rampant and disruption a reality around every corner. As Zuckerberg said in 2011, “In a world that is changing really quickly, the only strategy that is guaranteed to fail is not taking risks.”
That’s why it feels like Google and Facebook are constantly throwing things against the wall to see what sticks, while Apple takes three to five years to build their first devices in a new category, and Tesla will arrive at their $35,000 Model 3 after a 10-year journey.
In this competitive soup, Facebook is certainly not the first to embrace failure but has become one of the best.
Hacking Failure
At Facebook, failure doesn’t just happen. Risk and failure are institutionalized as opposed to merely allowed or encouraged.
Risk and failure live in its culture, made real through the hundreds of posters that dot nearly every available vertical space on campus—“What would you do if you weren’t afraid?” “Move fast and break things” “Fail harder” “Think wrong”—and the actions of all the people around you, not just the culture-carrying veterans.
They live in one of Facebook’s most celebrated practices: Hackathons. Multiple times a year, the company—especially engineering but increasingly other functions as well—comes to a stop and, for a 24-hour period, works on building prototypes in very small groups formed around passion projects that otherwise would not be able to receive time and attention. Important to closing the loop on this cultural touchstone is that Zuckerberg himself reviews the best projects, and a few are subsequently greenlighted to be developed into real products. Some Hackathon projects—like tagging people in photos, the ability to post videos and the game-changing performance benefits to Facebook infrastructure of HipHop for PHP—succeed wildly, while others—like Snapchat “homage” Slingshot—do not, but all efforts and outcomes are celebrated.
And they live, most importantly, in Zuckerberg himself—who shows it in big ways, such as the acquisitions of Instagram, WhatsApp and Oculus VR and moonshot projects like the Connectivity Lab and Artificial Intelligence Research, and in small ways like his personal project in 2016 to build a Messenger bot to control his own home and ownership of his leadership failures—that, like everything else at Facebook and to a greater extent than any other public company CEO, an appetite for risk and failure starts with him.
Good thing, because there’s been a lot of both. Whether additions to Facebook’s main service, products for advertisers, acquisitions or entirely new apps, many a Facebook swing has been a miss.
“Failures”: The Big Blue App
Although they become harder to remember with every headline of another hundred million people using Facebook’s main service on a daily basis, the company’s central effort has not escaped its share of unfulfilled dreams.
In the earlier days, the internal enthusiasm around Virtual Gifts (from February 2007 to July 2010) and Questions (July 2010 to October 2012) to drive greater structure and engagement around two popular activities between people on Facebook—birthday wishes and learning from the “hive mind” of your friends and friends of friends—was met with some uptake by the community, but it did not spread to wide and regular use, the bar for which, at Facebook, hovers in the hundreds of millions of monthly users. Interestingly, the two suffered from opposite issues as Virtual Gifts was simply not used by a broad enough set of people to “role-model” the behavior—we do what our friends do—while Questions suffered from too many seeming nonexperts weighing in on the queries posted by friends, making the corresponding outcomes seem less curated than the smaller but more successful community at Quora, a service dedicated to the premise and cofounded by former Facebooker Adam D’Angelo.
Two misses more central to the business followed with the all-hands-on-deck construction and almost equally fast shutdown of Deals (April 2011 to August 2011), a way to buy and share coupons for goods and services from businesses directly in News Feed, and the awkwardly named Graph Search (March 2013 to December 2014), a vision of search constructed specifically for Facebook’s giant collection of people, photos, connections, Pages, Places and content. Deals, an attempt to respond to the seemingly meteoric success of coupons-reinvented-for-the-Internet purveyor Groupon, suffered from issues inherent to the category, as special group deals proved faddish and unsustainable by the entire sector but most especially by the small and medium-sized businesses they were supposed to grow. They have since been replaced by more effective and consistent ways to raise awareness and purchase for businesses and products over the longer term, including Offers, in-line Buy buttons and the ability to relate Facebook marketing activity to in-store footfall. Graph Search, a highly structured semantic search of Facebook (e.g., “friends of Jane Doe who like Beyoncé and Vin Diesel,” “my friends who work at Apple,” “my friends from Europe who have visited Italy,” “pictures of me and John Doe”) was sophisticated but difficult-to-use and did not substantially grow the use of Facebook search beyond looking for people. Since that project, a focus on searching status updates and comments for keywords has been more pragmatic and successful, but more is possible, and necessary, for Facebook.
“Failures”: Advertising
Facebook has built the preeminent mobile advertising product in the world, but the road to that product has on occasion been bumpy.
In its early days, Facebook misstepped with a product called Beacon (November 2007 to September 2009), launched on the overreaching premise that “Once every hundred years media changes . . . ,” that posted selected activity elsewhere on the Internet back to Facebook. The product was initially only an opt-out function—meaning you had to proactively turn the feature off—causing a privacy furor over fears that information both harmless (that surprise birthday gift purchase) and not so harmless (that DVD of questionable content) would be broadcast to your five hundred closest friends. Zuckerberg had to apologize for the product and shutter it in the face of a class-action lawsuit. It would, however, be the predecessor of the opt-in and more successful Connect product that allows you to log into other sites and apps on the Internet and share actions and content back to Facebook if you so chose.
The development of Sponsored Stories (January 2011 to April 2014) followed. A sophisticated product that allowed advertisers to amplify actions people had taken with a business on Facebook—such as Like, Check-in and Offer redemption—to that person’s friends. While effective—the “social context” of seeing your friend engage with a business proved significant in the awareness and effectiveness of the advertising—the product was complex and unpredictable to use and triggered a class-action lawsuit that Facebook had to settle relating to the use of people’s likenesses in these ads. A crucial simplification of Facebook’s advertising products followed. Focused on ease of use, sophisticated targeting and effective content types, it led to the renaissance of its advertising business.
“Failures”: Acquisitions
Although Instagram, WhatsApp and Oculus VR make the headlines, you cannot expect your acquisition engine to have a perfect record.
Facebook acquired physical gifting service Karma (May 2012 to August 2014) to power its Gifts feature—especially around birthdays—to much enthusiasm in the market, which saw an opportunity for Facebook to participate directly in commerce—more than a billion users means more than a billion birthdays a year—rather than just encouraging it through advertising. The service, however, was not able to consistently grow engagement. Even if it had, the possible future complexity of Facebook having to be increasingly responsible for physical warehousing and fulfillment was a difficult premise against best-in-class providers like Amazon. Instead, Facebook focused on enabling retailers’ key asset (the logistics of consumer purchase) with its own key asset (knowledge of and access to those consumers).
To aid another one of its key constituencies, developers building Internet services and mobile apps, Facebook bought software-as-a-service provider Parse (April 2013 to January 2016) to provide easy-to-use infrastructure services that allowed developers to concentrate on building and growing their services. In a story mirroring that of Karma, it would be Amazon and their Amazon Web Services that proved more adept and popular at the task, and Facebook once again retreated to its unique and lucrative asset (the ability to drive installation and reengagement of the developers’ apps through advertising).
Although the story has not been fully written, Facebook’s acquisitions for advertising beyond Facebook, including targeting and ad-serving platform Atlas in 2013 and video ad supply-side platform LiveRail in 2014, have not fared well. A complete rewrite of the Atlas technology and huge competition from Google’s pervasive DoubleClick competitor has made adoption of Atlas a much longer road than Facebook anticipated, causing Atlas to be relegated to a cross-Internet measurement tool rather than its larger ambitions. Issues with the quality of LiveRail’s content supply (sites beyond Facebook’s own often suffer from click fraud by so-called bots not actually representing a human user and ads that are served but never seen) have caused wide swaths of the service to be shuttered to prevent degrading Facebook’s own effective and high-quality inventory. Its homegrown Facebook Audience Network, an easy-to-use extension of ad formats and sophisticated targeting based on Facebook’s knowledge of people to carefully qualified mobile apps has proven a much greater success on this front.
“Failures”: Mobile Apps
Nothing is more important in today’s Internet than the ability to successfully build and grow mobile apps. Facebook, Instagram, Messenger and WhatsApp are the premiere examples of the species and are all under Zuckerberg’s control. That doesn’t mean Facebook hasn’t had its significant share of struggles on this front. Its aborted efforts number in the double digits.
Perhaps most famous among these is Home (since April 2013), a replacement of the Google Android home screen featuring a full-screen browsing experience of Facebook content and stand-alone messaging. As close as Facebook ever got to building its own mobile operating system or phone, Home managed to show that there is actually such a thing as too much Facebook, with most users who elected to download and install the functionality being overwhelmed by Facebook’s seemingly taking over their phones. Technologies from the effort, however, made their way into more successful efforts, including Messenger and Instant Articles.
In its long competition with Snapchat for person-to-person and small-group sharing that appeals to Millennials, Facebook offered up Poke (December 2012 to May 2014), Slingshot (June 2014 to December 2015) and Riff (April 2015 to December 2015), all meeting with little success. Offering simple messaging, ephemeral messaging requiring return messages and collaborative video editing, respectively, none of the apps showed game-changing differentiation from either Facebook’s own offerings or those from formidable competitor Snapchat.
Launching to plaudits from across the industry for its engaging fullscreen design, the reimagined (and curated) Facebook reading experience of Paper (since January 2014) also found little large-scale traction among users. Its technology, however, lives on in the new reading experience of Instant Articles (and the related advertising product Canvas) in Facebook’s own mobile News Feed.
Similarly, Rooms (October 2014 to December 2015), pseudonymous chat forums for mobile, meant to address people’s interest in connecting without authentic identity during the rise of apps including Whisper, Secret, Ello and Yik Yak, failed to stand out.
The Facebook Creative Lab (January 2014 to December 2015) launched to enable experiments like Paper, Slingshot, Rooms and Riff, was soon shuttered due to its lack of success and a distinct sense that creating an isolated team tasked with creative breakthroughs was not nearly as effective as spreading the opportunity—and responsibility—for these breakthroughs across all teams at Facebook.
The distributed approach, however, is also not a guaranteed success engine and can lead to puzzling stand-alone offerings like Camera (since May 2012), stabilized time-lapse recorder Hyperlapse (since August 2014) and collaging tool Layout (since March 2015) emerging from various fronts, including—and occasionally seeming to compete with—the Instagram team.
The Future of “Failure”
Along with all the failure has come enough success to feed a cycle of increasing scale of risk, more than occasional new failure and larger successes for Facebook. It’s no longer just small coding projects. Instead, it’s ever larger acquisitions, consumer hardware, and expansive fail-faster factories like the Connectivity Lab and the latest organized risk-taking effort launched in 2016, dubbed Building 8. Reporting to CTO Mike Schroepfer and headed by no less an organized failure celebrity than former head of both the U.S. government’s Defense Advanced Research Project Agency (DARPA, which originated the Internet) and Google’s Advanced Technology and Projects (ATAP) group Regina Dugan, Zuckerberg intends to invest hundreds of millions of dollars over the next few years on the team’s efforts to build software and hardware systems beyond the mainstream of Facebook efforts. A full-fledged incubator with a separate COO, head of creative and business and technology staff, its seriousness cannot be questioned. After all, it was Dugan who labeled her group at Google—working on projects like 3D indoor mapping systems, modular mobile phones and 3D animation—a “small band of pirates trying to do epic sh*t.”
While this upward cycle is the exact opposite of the toilet bowl of stagnation that occurs when some large companies constrict in the face of risk or fail to recognize that the worldview shaped by their prior success is being eclipsed by new forces, it does not last forever. Even Alphabet (Google’s parent company)—who still needs to be considered the trail-blazer when it comes to the heady mix of risk and failure now commonly referred to simply as moonshots in Silicon Valley—has had to curtail, focus and rationalize its speculative activities, including self-driving cars, healthcare devices, high-altitude balloons for Internet delivery in developing countries, wind farms and the reconceptualization of entire city infrastructures. And they enjoy as their launch pad the second-best profit machine in the world behind Apple.
Flying ever closer to the sun means some questions that Zuckerberg could shove to the back of his mind in the past now crawl into the light. How much risk is too much risk? Is there enough success among all the risk? When does trying to out-innovate Snapchat and not succeeding time and again become alarming and not just the natural order of competition (with The Wall Street Journal having projected in October 2016 an IPO for Snapchat parent Snap Inc. in 2017 possibly valuing the company at $25 billion or more this threat has never been more real)? What happens to all this upward momentum if the advertising business that funds all of it begins to show cracks?
When, in other words, does all that harder and faster “Failure” become just . . . failure?
The two Facebook “failures” closest to actually being failures in the sense of diminishing Facebook’s current and future trajectory are its lack of progress in presenting a next-generation search experience across their own data as well as that on the open Internet and its apparent inability to grow successful new mobile products internally.
In search, Facebook has the two assets most necessary to do well: a search box at the top of its user experience and a leading slice of the time and attention of a billion users a day (search algorithms themselves are relatively commoditized at this point). Its display advertising business is a perfect adjacency to a search advertising business in competition with Google. It will need this asset as the opposite sides of the digital advertising chessboard we covered back in Chapter 8 will eventually grow to overlap, and maintaining revenue growth will have to come at least partially from taking share in digital rather than counting merely on money moving from television to digital and feeding Google and Facebook evenly. After more than five years of trying, however, Facebook is not appreciably closer to a unique and monetizable search offering that would cut into, much less displace Google’s.
In mobile apps, Facebook’s success beyond Facebook has come either from spin-offs of functionality in the main app (Messenger, Groups and Lite) or from acquisitions (Instagram and WhatsApp). Their original efforts have not been very . . . original, especially in their battle against Snapchat, the next most engaging mobile app and fastest growing. Their inability to put forth a differentiated, engaging product has forced Zuckerberg to play defense via acquisition. He has done this well with Instagram and WhatsApp, but it will leave Facebook exposed when the next acquisition refuses to sell or gets snapped up by Google instead.
Zuckerberg remains undaunted in the face of these issues. On the heels of announcing a breakout quarterly earnings performance in April 2016—at a time when stalwarts Apple and Alphabet had failed to meet Wall Street expectations—he made his intentions clear: “When I look out at the future, I see more bold moves ahead of us than behind us.” (Admittedly, this rhetorical flourish also served as enticing context for the introduction of a new class of Facebook stock meant to keep Zuckerberg firmly in control of the company as he divests his command position to fund his and his wife’s philanthropic effort, the Chan Zuckerberg Initiative.)
After a dozen years of successfully pursuing his mission, Zuckerberg—not unlike fellow large-scale risk-takers Larry Page, Jeff Bezos and Elon Musk—has given missions back their good name. To the extent that he “competes,” he does so only with his mission and the time he has to execute it.
In that context, almost any risk is worth taking.