Fy Agency - Logo

Growth Hacking: Hype or Holy Grail For Start-ups

Growth HackingChamath Palihapitiya had screwed up. It was November 2007, and the executive in charge of Facebook’s platform and monetisation had a full-blown crisis on his hands. Days earlier, he’d overseen the launch of what Facebook called “a completely new way of advertising online”: an ad network known as Beacon that would post the commercial activities of Facebook users to their news feeds when on 44 third-party websites, unless they actively opted out. Palihapitiya, a rising star who’d run AOL’s instant-messaging division and founded MyMusicChannel, had underestimated the resistance to Beacon: from furious users such as Sean Lane, whose wife learned of a surprise gift through a wall post that stated, “Sean Lane bought 14k White Gold 1/5ct Diamond Eternity Flower Ring from overstock.com”, and who subsequently led a class action against Facebook; and from the 50,000 signatories in ten days to a MoveOn.org petition against the programme headed: “Facebook, stop invading my privacy!”

“We were sitting in the third floor of 156 University Avenue [in Palo Alto], and TV trucks with cranes were bearing down on us, and people outside protesting,” Palihapitiya, now 37, recalls. “I was so naïve with the press — trying to explain cross-site Javascript toThe New York Timesis generally a bad idea. The episode had a really chilling effect on people’s perceptions of Facebook.” Mark Zuckerberg finally agreed to roll Beacon back. “We’ve made a lot of mistakes building this feature, but we’ve made even more with how we’ve handled them,” Zuckerberg wrote on the Facebook blog. “We did a bad job with this release, and I apologise.”

Palihapitiya had been at the company for less than a year. “I had a lot of guilt,” he says now. “I felt I had to undo a bunch of damage that I’d caused. So in 2008, I said, OK, we need to grow again. I remember sitting in my bedroom, talking [on the phone] to Zuck and Sheryl [Sandberg], saying, ‘Let’s create a growth team.’ The six-month plan would be to build credibility by proving and disproving the things that people think are anecdotally right, and make it fact-based. And that would probably give us another six months before we all got fired.”

“In Japan, we list your blood type in your profile. We knew they used blood type to create trust with people; we’d asked a Japanese person.”

His growth team, initially six people but now around 200 excluding analytics, was an unusual hybrid of product and marketing specialists, who sought to boost user recruitment and engagement through a combination of A/B testing, data analytics, behavioural psychology, design and user experience. Palihapitiya’s team soon found unexpected ways to kickstart Facebook’s growth: “In Japan, we list your blood type in your profile. We knew they used blood type to create trust with people; we’d asked a Japanese person. So we think about it; then we test it; if it works we use it; then if it stops working we get rid of it. But in Russia, we bought keywords for every single person’s name and ran ads. As Russians were ramping on the internet, the first thing they’d do was search for themselves, and we captured all that traffic.”

The growth team’s mission was to discover “these really important psychological trends around how people really feel about friendship, how they really think about people connecting to them, how these things drive subtle behaviours around how you engage with the product,” he says. “We’d figure out that we could persuade a woman to upload a photo to Facebook if she was shown 11 photos; from that we found we could show that [photo] to seven men, and four of those people would invite five other people… We could link this multi-hop effect of the emotional decisions people were making — and it all started with one photo.”

Palihapitiya, a keen poker player, bet more than his professional reputation on his growth machine’s success. “I made a deal with Zuck,” he says, “saying, ‘You could pay me X now, or pay me multiples of X if we get to a billion users.'” Good call: by August 2008, Facebook’s user base had doubled to 100 million; the following September it was at 300 million; and it hit a billion in October 2012. This success soon inspired other Bay Area startups to build growth teams with a mission to test, track, measure and redesign their products to success. In place of traditional marketing, advertising and PR, fast-growth businesses such as Zynga, LinkedIn, Groupon and Dropbox became known for obsessing with A/B tests, landing pages, virality factors and email deliverability rates. Thus began a wider “growth hacking” movement, which today incorporates a booming sector of specialist agencies such as Growth Devil, “possessed by metrics and experimentation”; vendors of “growth-hacking tools” such as Colibri; video channels such as Growth Hacker TV, offering “specific actionable tactics”; and events such as the London Growth Hacking Conference this October (£600 to learn “the strategies, mindset and tactics you need to go from marketer to growth hacker”).

Which raises a question: for all its current buzz, is growth hacking really a new idea, accessible through an almost scientific growth “toolkit” that any business can adapt? Or is it just conventional marketing repackaged in hype?

The term “growth hacker” was coined in 2010 by Sean Ellis,a former marketer with Dropbox, Eventbrite and LogMeIn, but the underlying strategies around building online viral growth go back far earlier. In 1996, when Sabeer Bhatia and Jack Smith were launching Hotmail, their investor, Timothy Draper at Draper Fisher Jurvetson, asked how they were planning to attract users. Billboards and radio advertising, Bhatia replied. “God,” Draper replied, according to Adam Penenberg’s bookViral Loop. “That’s expensive marketing and we’re giving this away?” Instead, Draper suggested, they should use the emails themselves to promote the service: “Put ‘PS: I love you. Get your free e-mail at Hotmail’ at the bottom.” The founders eventually agreed to test the strategy — minus the “PS: I love you”. Within 18 months, they had gone from zero to 12 million users. In December 1997, Microsoft acquired Hotmail for $395m (£240m).
“What Tim Draper suggested for Hotmail is the greatest growth hack of all time,” according to Alex Schultz, growth VP at Facebook today. “It was brilliant — everyone else was spending money on TV.” Schultz gives Viral Loop, plus Ogilvy on Advertising, to all new joiners on his team — which he calls not “growth” but “internet marketing”. Skills include “data-inspired design”, linguistic understanding (“How do we predict which people in Pakistan should get Hindi, Urdu or English on their feature phones?”) and ensuring that SMS authentication works in countries where email usage is low. “It’s the intersection between engineering, product, marketing and data. Having a combination of marketing mindset and analytical mindset works well. You need a normalising metric, a clear goal that everyone’s aligned on — our goal could have been registered users, but Mark would have been very upset if we’d grown registered users and not active.”

“The most important thing is to have a product that people really love to begin with. Dropbox solves a problem — file synchronisation — that you didn’t realise you had, and it actually works.”

For other businesses, viral growth has come from classic referral strategies. Dropbox — which in May announced 300 million users, up from 200 million last November — initially experimented with paid Google ads when it launched publicly in September 2008. This, founder Drew Houston quickly discovered, was costing $233-388 (£139-228) per acquisition, for a product that cost $99 (£60). Instead, he prioritised a referral programme that offered 2GB of free storage to new users, plus bonus storage to any existing user who shared the link. But even before the public launch, Houston had obsessively tracked what did and didn’t generate interest. An early three-minute video explaining the service, narrated by Houston and packed with geek in-jokes, boosted the beta waiting list from 5,000 to 75,000 people overnight when posted on Hacker Newsin April 2007, he later said.

“We did lots of tactical things, such as our referral programme, shared folders, and other aspects of the product that helped it spread,” Houston tells Wired. “Those tactics can amplify something — but the most important thing is to have a product that people really love to begin with. Dropbox solves a problem — file synchronisation — that you didn’t realise you had, and it actually works. That may not sound that high a bar, but you only hear a couple of times a year, ‘Oh my god, I’ve only had this thing for a week and now I can’t live without it.’ That sort of epiphany is something you want to share. And we had this currency called space, like points, so we inadvertently gamified the whole process of spreading Dropbox.” You can earn extra storage by logging in through Facebook or Twitter; likewise, by giving product feedback.

Constant iteration of the landing page through testing has also optimised the sign-ups. “We focus on two categories,” says Houston. “First, split testing and optimisation; we have a business operations team, a monetisation team, an analytics team, a data science team. That’s one bucket. The one we focus more on, the bigger lever, is the new ways of sharing. People are tempted to bolt on a referral programme like ours and just expect a product to take off. But they underestimate the degree to which it goes back to building something people love.”

In the early days at LinkedIn, testing, measurement and product iteration became an almost religious obsession. Josh Elman, now a partner at Greylock Partners but formerly a product manager at LinkedIn, Facebook and Twitter, recalls repeated testing of the invitation wording: “When we had the phrase in the default invitation that joining ‘will make both of our networks bigger…’, the likelihood that someone who’d signed up would invite more people to sign up was higher.” Growth was boosted by making it easy for members to import their email contacts; engagement was boosted by installing a progress bar that encouraged fuller completion of profiles. “It’s not just looking at the data, but getting to the core principles of why [you want to engage],” Elman says. “The real trick is using data to let you get to these deep psychological insights, and building features or products around that. In the early days on LinkedIn, you signed in, got a ‘welcome to this product, go and invite more people’. But where was the value? So we changed it to, ‘Hey, welcome to LinkedIn, here’s a bunch of people from the same company who are already here. Now who else will you invite?’ I call it the double viral loop — by causing ten colleagues to get a notification that said, ‘Hi, Wired colleagues, David has just joined LinkedIn and asked you to connect with him,’ they now come back to connect. At that moment, LinkedIn says, ‘Hi, David’s colleague John, who else do you know that you’d like to invite?’ At that moment it’s more likely that John will invite more people than simply getting David to invite people in the first place.”
And the launch of the LinkedIn jobs site, in 2004-5? “Everyone thought LinkedIn was trying to do revenue. You know the real reason? People were afraid to sign up and upload their profile, as they felt we were a jobs site. But by creating a separate jobs site on LinkedIn, we could convince you to go to LinkedIn and it wouldn’t look like you were looking for a job. It’s all psychology.”

“Our technique was to find a name for the problem (‘grey charges’) and put a number on it ($14.3bn leaking from American’s wallets every year).”

Airbnb famously boosted growth by integrating an automated way to cross-post accommodation listings on Craigslist. Twitter found that return rates shot up once it had persuaded new users to follow ten people. BillGuard, which monitors customers’ credit-card bills for unauthorised spending, boosted app downloads by inventing a new term: “We had the nearly impossible task of creating awareness for a problem on a startup budget,” says founder Raphael Ouzan. “Our technique was to find a name for the problem (‘grey charges’) and put a number on it ($14.3bn leaking from American’s wallets every year) from a third-party firm analysing our data. The result: thousands of news outlets, morning shows, articles talking about grey charges, just as we launched our iPhone app to solve the problem.”

Typically, online businesses that find their way to sharp user growth tap a combination of data analytics and traditional marketing tools. PlayBuzz, which makes it easy to create and share social quizzes, had 13,000 active users in January. Then founder Shaul Olmert began experimenting: A/B testing and measuring all aspects of the product, from format to design, using fractions of the audience; search engine optimisation; experiments to promote engagement on Facebook, Google+, WordPress, Twitter and Pinterest. “We were up to three million in February, 12 million in March, 30 million in May, and by June we were getting five million daily uniques,” says Olmert. “From widgets to ’embed’ buttons and plugins, we’re making it easy for our users and partners to share our content.”

To Jim Goetz, the Sequoia Capital investor who in 2011 put $8m into a company called WhatsApp — which, after growing to half a billion active users, sold to Facebook for $19bn in February — sees this approach as “an enormous opportunity” for any consumer-facing business. “It’s probably your biggest lever,” he says. “You’ve got A/B testing; experimentation with the App Store — you see the best gaming companies launch in Canada, Australia or Spain first and then iterate; then people such as Amazon do live ops where they’re changing things every day. It’s a blend of engineering and a go-to-market team member that’s part of this growth-hack, live-ops process. The companies that do it best have built it in house.”

“Growth hacking is spending time engineering how your product gets out into the world.”

However, a growth-hacking backlash is emerging. Earlier this year, digital marketing consultant Muhammad Saleem published a much-discussed article titled “Growth Hacking Is Bull”. Dropbox’s strategy was simply traditional referral marketing, Saleem argued; Airbnb’s Craigslist strategy was simple cross-posting. “Not only is growth hacking a meaningless phrase used to rebrand online marketing by the people who’ve spent the better part of their careers maligning online marketing,” he argued, “but it is harmful to your company. Growth hacking perpetuates this myth that you can achieve hockey-stick growth by using short-term ‘hacks’.”

What’s new, Josh Elman argues, is the application of an engineering mindset to product marketing. “Growth hacking is spending time engineering how your product gets out into the world. Nest, SmartThings and WhatsApp are building incredible technology and products. The big shift is what happens when we take the creative brilliance of design and engineering and apply it to how we get that product into people’s hands.”

Chamath Palihapitiya is more hesitant. “I hate the term ‘growth hacker’. There are a lot of snake-oil salesmen in this field,” he says. “Let’s not create some wizard-behind-the-curtains thing about this concept called growth hacking. It existed well before me. It’s called product and marketing.”

What does have scalable value, Palihapitiya says, is “the deeper psychological understanding of product and behaviour”. Since leaving Facebook in summer 2011, he has taken his approach to The Social+Capital Partnership, a Palo Alto-based fund that defines itself as “a growth practice”. After making an investment, he will send in a growth team — “abstracted Navy Seals”, ranging from physicists to machine-learning experts — to sit with the company for a month, extracting data to frame how they should think about their product. “The framework works for any business,” he says. “I own a chain of restaurants and I’ve deployed one of my guys to go in there and fix it. They’re measuring costs, how long customers stay, changing the menu. Per-ticket price is up 35 per cent in one restaurant.”

What advice does Palihapitiya have for entrepreneurs reading Wired? “Users are only ever in three states — they’ve never heard about it; they’ve tried it; and they use it. What you’re managing is state change. So the framework is, what causes these changes? The answer should be rooted more in preference, choice and psychology than in some quantitative thing.

“So with Wired, there are people who’ve never heard of it, people who may read it, people who read it and may stop. Look at the psychological things they thought they were going to get from you — validation, community, intellectual stimulation, that stuff. When you figure out why people stay with the magazine — the community, the association that’s happening — you should be saying, shit, we should be organising this stuff offline too.”

This methodology could become truly powerful, he adds, in re-assessing large traditional businesses. “If I ever chose to raise a much larger private-equity-style fund and buy businesses, this framework could be really valuable,” he says. “The problem is, those businesses are so deathly boring. I have no interest in them. I’d rather build the future than fix the past.”


Shared from WIRED blog written by editor David Rowan about Tony Fadell in 07.14